AN INTRODUCTION TO

POSITIVE ECONOMICS
Some E.L.B.S. Economics Titles

Amtey an introouctios ecomouic& sor

and Martin students in india and Pakistan Allen & Unwin
Crowther an outline of money Melson

Hanson monetary THEORy and practice Macdonald and Evans

Hicks VALUE AND CAPITAL Oxford University Press

Lipsey AN INTRODUCTION TO POSITIVE Wetdettfeld and

ECONOMICS Ntcolson

Marshall principles of economics Maemillan

Paish benham’s economics .

Pstman

Pigou THE ECONOMICS OF WELFARE Macmillan

Prest PUBLIC FINANCE Weidenfold and

}ltCBlson

Reddaway the development op the Indian

ECONOMY Allen & Unwin
Stonier
and Hague a textbook of economic theory Longmans
An Introduction to

POSITIVE
ECONOMICS
RICHARD G. LIPSEY
Professor of Economics at the University of Essex

ENGLISH LANGUAGE BOOK SOCIETY

and
WEIDENFELD AND NICOLSON
 © 1963 by Richard G Liptey
First printed October 1963
Second impression July 1964
Third impression January 1965
Fourth impression June 1963
E LBS edition first published 1963
Fifth impression February 1966
Second edition October 1966
Second impression August 1967
Second ELBS edition 1968

StI in Menep/iaU BeskmnlU and

Qfi mnife and printed ij affset in Great Bnh
n ip
Clowes and Sons Ltnttled London and Beccles
CONTENTS

The Use of this Book xiii

Notes on the Second Edition xix

Part 1 Scope and Method

1 Introduction 3
2 The Tools of Theoretical Analysis 19
—Some Common Techniques
Appendix 33
3 The Tools of Analysis
Statistical 43
Appendix— Graphing Economic Obser\'ations 55
4 The Problems of Economic Theory 58

Part 2 The Elementary Theory of Price

5 A General View of the Price System 71

6 The Theory of Market Behaviour: Some Preliminar)' Con-
siderations 76
7 The Elementary' Theory of Demand 80
8 The Elementary Theory of Supply 94
9 The Elementary Theory of Market Price 99
10 Elasticity of Demand and Supply 108
Appendix —A Formal Analysis of Elasticity 122
1 1 Some Predictions of the Theory' of Price 126
12 The Elementary' Dynamic Theory of Price 152
13 A Postscript and a Preview 160

Part 3 The Intermediate Theory of Demand

14 Some Basic Theorems and Predictions 167
15 Theories of Household Behaviour 177
16 The Theory of Demand: Measurements and Tests 204

Part 4 The Intermediate Theory of Supply

1 7 Background to the Theory of Supply 225

18 The Organisation of Production 235
 Vlll CONTENTS
19 The Measurement of Costs 247
Appendix —
Balance Sheets, Income Statements and Costs of
Production The Views ofthe Economist and the Accountant 255
20 The Vanation of Cost with Output 260
—
Appendix Long-run Vanations in Output 279
21 The Very Long Run 287
22 The Theory of Perfect Competition 295
—
Appendix Mathematical Proof of the Rules for the Profit-
Maximising Behaviour of a Firm 313
23 The Theory of Monopoly 315
Discrimination 328
25 Competition among the Many Monopolistic Competition 335
26 Competition among the Few The Theory of Oligopoly 343
27 Some Predictions of the Theories of Competition and Monopoly 352
28 Monopoly Versus Competition Predictions about Performance 372
29 Criticisms and Tests of the Theory of Supply 391

Part 5 Distribution

30 A General View of the Theory of Distribution 405

/• 31 The Demand for Factors Marginal Productivity Theory 414
^32 The Supply of Factors 422
33 The Pricing of Factors of Production m Competitive Markets 434
34 Labour Unions, Collective Bargaining and the Determination
of Wages 452
Interest and the Return on Capital 470
Criticisms and Tests of the Theory of Distribution 482

Part 6 The Economy as a Whole

37 The Interaction among Markets General and Partial Analysis 497
38 Micro-Economic Policy
39 From Micro- to Macro-Economics 537

Part 7 The Circular Flow of Income

40 The Model of the Circular Flow of Income 545
41 Equilibrium in the Circular Flow of Income 564
42 Some Predictions of the Simple Theory of National Income 583
43 Household Consumption 594
Investment and Saving W
45 Fluctuations m the Level of Economic Activity
46 Foreign Trade and National Income 631
•^Government and the Circular Flow of Income 641
 CONTENTS IX

Part 8 Money, Banking and the Price Level

48 The Nature and History of Money 665
49 The Banking System and the Supply of Money 677
50 The Demand for Money 694
51 The Determination of Price Level 700
Appendix A—Micro-Economic Implications of Various
Aggregate Relations BeUveen Expenditure and the Price Level 717
—
Appendix B Cost-Push Versus Demand-Pull; A Case Study 720

Part 9 The International Economy

52 The Gains from Trade 733
—
Appendix Does the Price System Lead to a Pattern of Trade
which Accords wath the Balance of Comparative Advantage? 741
53 Exchange Rates 747
Appendix—^More about Exchange Rates 764
54 Tariffs and the -Gains from Trade 768
55 International Economic Experience 778

Part 10 Economic Growth and De\^lopment

56 Economic Growth 799

57 Underdeveloped Economies 816

Part 1 1 Macro Econootc Policy

58 Macro-Economic Policy 829

Index ,
862
FACT AND THEORY IN
ECONOMICS

. . —
Einstein started from facts the Morley Michelson measurements of
light, the movements of the planet Mercury, the unexplained aberrancies
of the moon from its predicted place. Einstein went back to facts or told
others where they should go, to confirm or to reject his theory by observa- —
tion of stellar positions during a total eclipse.
‘. . . It is not necessary, of course, for the verification of a new theory' to be
done personally by propounder. Theoretical reasoning from
its facts is as
essential a part of economic science as of other sciences, and in a wise
division of labour there is room, in economics, as elsewhere, for the theoreti-
cian pure and simple, for one who leaves the technical business of verification
to those who have acquired a special technique of observation. No one
demanded of Einstein that he should visit the South Seas in person, and look
through a telescope; but he told others what he expected them to see, if they

looked, and he was prepared to stand or fall by the result. It is the duty of
the propounder of every new theory, if he has not himself the equipment for
observation, to indicate where verification of his theory is to be sought in
facts —
^what may be expected to happen or to have happened if his theory is
true, what will not happen if it is false.
‘[Now consider by way of contrast the behaviour of the participants in a
current controversy in economics.] None of them takes the point that . . .

the truth or falsehood of . .

.
[a] . . . theory cannot be established except by
an appeal to the facts; by facts himself. The distinguish-
none of them tests it

ing mark of economic by this debate is that it is a

science as illustrated
science in which verification of generalisations by reference to facts is neg-
lected as irrelevant. ... I do not see how [membere of the public who . . .

survey the controversy] can avoid the conclusion

. . .that economics is not a
science concerned with phenomena, but a survival of medieval logic, and
that economists are persons who earn their livings by taking in one another’s
definitions for mangling.
‘. I know that in speaking thus I make enemies. I challenge a tradition
. .

of a hundred years of political economy, in which facts have been treated

not as controls of theory, but as illustrations. I shall be told that in the Social
Sciences verification can never be clean enough to be decisive. I may be
Xll FACT AND THEORY IN ECONOMICS
told that, m these sciences, observation has been tried and has failed, has led
accumulations of facts which themselves lead nowhere I do not
to shapeless
moment that this charge of barrenness, of past enquiries can be
believe for a
sustained, tomake it is to ignore many achievements of the past and to
decry much solid work that is being done at this School and elsewhere But
if the charge of barrenness of realistic economics in the past were justified
completely, that would not be a reason for giving up observation and
verification It would only be a reason for making our observations more
exact and more numerous If, in the Soaal Sciences, we cannot yet run or

fly,we ought to be content to walk, or to creep on all fours as infants

For economic and political theonsing not based on facts and not controlled
by facts assuredly does lead nowhere
‘There can be no science of society till the facts about society arc available
Till 130 years ago we had no census, no knowledge even of the numbers and
growth of the people till fifteen years ago we had no comprehensive records
,

about unemployment even m this country, and other countnes are sull
where we were a generation or more ago, social statistics of every kind
—
about trade, wages, consumption are everywhere m iheir infancy
‘
From Copernicus to Newton is 150 years Today, 150 years from the
Wealth of Nations, we have not found, and should not expect to find, the
Newton of economics If we have travelled as far as Tycho Brahe we may be
content Tycho was both a theorist and an observer As a theorist, he
believed to his last day in the year 1601 that the planets went round the sun
and that the sun and the stars went round the earth as the fixed centre of
the univene As an observer, he made with infinite patience and mtegnty
thousands of records of the stars and planets, upon these records Kepler, in
due course, based his laws and brought the truth to light If we will take
Tycho Brahe for our example, wc may find encouragement also It matters
little how wrong wc are with our exisUng theories, if wc are honest and
’
careful with our observations

Extracts from Lord Bevendge’s farewell address as Director of

the London School of Economics, 24th June, 1937 Published in
Politico, September 1937
THE USE OF THIS BOOK

This an introductory text book in economics, starting at an elementary

is

stage and progressing, in some places, to an intermediate level. It is designed

to be read as a first book in economics. I hope, on the other hand, that the
book will not be ivithout interest for someone who has already studied one
of the many existing basic text-books written at a first year university
standard.
The book had its beginnings when I was asked to give the basic economic
theory lectures for the revised B.Sc.(Econ.) degree introduced at the London
School of Economics in October 1961. 1 started to write my first few lectures
and, almost before I knew it, a book was well under way. Had I appreciated

at the outset what was involved in carrying such a project through to

completion, I would never have begun it, but that is probably true of a
great many enterprises in all fields.
There are three major themes of this text-book which should be men-
tioned here: first, an attempt to explain what economic theory is about and

how one can go about criticising it effectively and hence improving it;
second, an attempt to elaborate, in so far as is possible within the confines
of an introduction to economic theory', the relation between theory and
real-world obser\'ations and third, a consideration of the relation between
;

economic theory and economic policy.

The first major theme of this book is how one can go about being in-
telligently and constructively critical of the existing body of economic
theory. I have tried to address myself throughout to the intelligent student
of honours-school quality. I have assumed that the student was interested in
his subject and that he wished to know, at every stage, just what was going
on and why. There is a tradition of trying to sneak quite complex bits of
analysis past the student without telling him what is happening. This may
be the best thing to ’do if the object is to get through an examination a large
mass of people who have neither interest nor ability in economics, and who
are hostile to the basic idea of a Social Science. I am not interested in
reaching such a public. I have assumed that I am addressing an intelligent
set of students, who may or may not be honours students and intending
specialists, but who want to learn and who do not have closed minds. One

of the troubles with the traditional approach of sneaking analysis past the
XIV THE Use of this book
student is that, when the intelligent reader feels that there is something
wrong with what he has been taught, he does not know how to go about
being in an effective way I have made a point of telling the
critical of it

student what going on, to say ‘now we are doing comparative static
is

equihbnum whatever it might be, and I have devoted con

analysis’ or
siderable space to an analysis of both sensible and silly criticisms of the
thcones described I do not accept the idea that the possibility of cnticising
what he has learned should not be mentioned because if it is the student will
be lead to make hasty and confused criticisms A good student will always
attempt criticisms and evaluations of what he has been taught It seems to
me that his criticisms are much more likely to be informed and relevant
ones if he is given both practice and instruction m how to set about effec-
tively challenging what he has been taught than if he meets a conspiracy of
silence on this topic
The second major theme is that of the relation between economic theory
and observation One of the most unfortunate tendencies m the teaching of
economics, particularly in Britain, is that of making a clear split between
economic theory and applied economics All too often economic theory is
taught merely as logical analysis, and is, at best, only vaguely related to the
world, while applied economics becomes description unenlightened by any
theoretical framework Economic theory is meant to be about the real
world We by the use of theory, to explain, understand and predict
seek,
phenomena in the real world, and our theory must therefore be related to,
and tested by, empincal observation of the world around us The student of
economic theory needs to ask at every stage what are the relevant magni-
tudes and quantities in the real world This is the theme set by the quotation
from Beveridge that opens this book '
The third major theme, which should be noted here, is that of the relation
between economic theory and economic policy The distinction between
positive and normative statements is well knosvn to professional economists,
but all we fail to communicate its significance to our students
too often
Even the best Amencan text books often manage to convey the idea that
economic theoryjustifies the pnvale-enterpnsc, market economics, found in
most Western countries The expenence of interviewing, for admission to
the London School of Economics and later to the University of Essex,
students who have done A-Level economics at school makes it painfully
obvious that from somewhere - I am never sure from where - students get
such ideas as the ones that the ‘La^v’ of Comparative Advantage proves that

1 I (alee It that in the last sentence quoted, Bevendge is saying that it does not matter how
wrong present theories may be as long as we make careful observations of the facts on which
these theories stand or fall and then discard or amend these theones when they are found to
be inconsistent with the facts
 THE USE OF THIS BOOK XV

nations ought to specialise in the production of certain goods, that economics

has proved that rent and price controls are wcked and ought not to be
used. The student who canthink of good reasons why, under some circum-
stances, price and rent controls and tariffs might be desirable, reacts by
dismissing economics as Medieval Scholasticism, or as a fraud perpetrated
by whichever political party he happens not to support; and so he should
do if economics did purport to prove such propositions. Economic theory
cannot, of course, ever show us what we ought to do, but only ivhat will
happen if we do do certain things. The uses and the limitations of economic
theory in dealing with matters of public policy is a theme which recurs
throughout the book.

The study of economics can be both interesting and rewarding. To the

student not intending to specialise in economics it can givT some under-
standing of the functioning of the economy, and some appreciation of the
issues involved in current controversies over economic policy. It may also
give him some idea of the methods which have been applied -with some
modest success in one Social Science. To the would-be economics specialist
the study of an introductory book, such as this one, can be the beginning of a
real adventure. The scope of his chosen science opens up before him. At
first he encounters theories which add to his understanding of the world,
but ver>' soon he begins to encounter problems: observations for which
there are no satisfactory explanations, and theories which are generally
agreed to be but which have not been adequately tested.
unsatisfactor)',
Both of these constitute a challenge, in the first case for the development
of ne%v theories and in the second for the making of a careful set of observa-
tions to test an existing theory". One of the interesting things about
economics in the 1960’s is that the frontiers of knowledge in terms of
unsolved problems can be reached very' quickly, even though, as the reader
of this book will soon learn, it may take a very long time to reach the
frontiers of knowledge in terms of techniques available for the handling of
problems.
Economics is a subject quite unlike some other subjects studied at school.
Economic theory has a logical structure; it tends to build on itself from stage
to stage. Thus the student who only imperfectly understands some concept
or theory will run into increasing difficulty when, in subsequent develop-
ments, this concept or theory is taken for granted and built upon. Because of
its logical structure, quite long chains of reasoning are encountered: if A
then B, if B then C, if C then D, and if D then E. Each step in the argument
may seem simple enough, but the cumulative effect of several steps, one on
top of the other, may be bewildering on first encounter. Thus when, having
followed the argument step by step, the reader encounters the statement
XVI THE USE OF THIS BOOK
‘now obviously if A then E’, it may not seem so obvious at all This is a
problem which almost everyone encounters with chains of reasoning The
only way to deal with it is to follow the argument through several times
Eventually, as one becomes familiar with the argument, it will become
obvious that, tf A then E
Another problem is posed by the fact that economics has a large technical
language or jargon At first the student feels that all he is being asked to do
IS to put complicated names to common sense ideas To some extent this
is

true At the beginning, economics consists largely of making explicit ideas

which appeal strongly to common sense and which one has already held in
a vague sort of way This is an absolutely necessary step, because loose
thinking about vaguely formed ideas is one of the quickest routes to error in
economics Furthermore, the jai^jon the single word or phrase given to the
common sense idea, becomes necessary m
the interests of brevity of ex
pression as the subject is built up The student should himself try the exercise
of removing every technical term from the argument of one of the later
chapters m
this book and replacing these terms with the full verbal descrip*

tion of the ideas expressed The ai^ument would then become too cumber-
some If we are going to put several ideas together to see what follows from
them, then the single clearly defined word or phrase to refer to these ideas
becomes a necessary part of our equipment
It follows from all of this that the student should use this book in quite a
different way than he would use a book on many other subjects A book
on economics is be worked at and understood step by step, and not
to
to be read like a novel It is usually a good idea to read a chapter quickly in
order to sec the general run of the argument and then to re-read it carefully,
making sure that the argument is understood step by step The student
should be prepared at this stage to spend a very long time on difficult
sections He should not be discouraged if, occasionally, he finds himself
spending an hour on only two or three pages A paper and penal is a
necessary piece of equipment in his reading Difficult arguments should be
followed by building up one s own diagram while the argument unfolds,
rather than by relying on the pnnied diagram which is, perforce, complete

from the beginning Numerical examples should be invented to illustrate

general propositions At various stages the student is asked to put the book
down and think out the answer to some problem for himself before he reads
on the student should never read on wtthout attempting to do what is asked He
should also make his own glossary of technical terms, committing the
definitions to memory The first time a technical term is introduced, it is

printed m
capital letters so that it may easily be recognised as such
After the book has been read in this detailed manner, it should be re-read
fairlyquickly from cover to cover, it is often difficult to understand why
 THE USE OF THIS BOOK X\T1

certain things are done until one knows the end product, and, on a second
reading, much that seemed strange and incomprehensible, will be seen to
have an obvious place in the analysis.
In short, one must seek to understand economics, not to memorise it (the
technical vocabulary must, of course, be committed to memory) Memorisa- .

tion is the royal road to disaster in economics; theories, principles and

concepts are always turning up in slightly unfamiliar guises. To one who has
understood his economics this poses no problem; to one who has merely
memorised it, tliis spells disaster. The required approach is not more difficult than,
but it is different from, that encountered in many other subjects.

The first chapter of this book is a general essay on scientific method,

particularly as it applies to the Social Sciences. None of the ideas are
unduly difficult but they may seem rather abstract and ‘up in the air’ to
someone unable to relate them to a detailed knowledge of some social
science. The ideas are, however, of critical importance. The student,
particularly the one who is just beginning his study of economics, should
read the chapter carefully to get the general drift of the argument. He
should refer back to it at times when the issues raised turn up in the con-
texts of particular bits of economics and, finally, when he has finished the
whole book, he should re-read Chapter 1, making sure that he follows the
argument fully. This last piece of advice is so important that, were it not for
the costs of printing, I should be inclined to print Chapter 1 as both the
first and the last chapter of the book.

Chapter 2 deals with the tools of analysis used by the economist. The
student does not need to be a mathematician to learn economics, but he
does need to know the equivalent of O-Level mathematics. There is not
room in an economics text-book to teach the elementary mathematics that
is assumed in economics. Chapter 2 does, however, outline some of the most

important bits of elementary mathematics commonly used. The student

who is unsure of any of the points briefly outlined must go back to an ele-
mentary text-book and review these bits of necessary technique.
There are occasional footnotes that are particularly difficult or that are
addressed primarily to advanced students or to teachers. Footnotes that
can be omitted for any of these reasons are flagged by an asterisk (e.g. ).
Students who are finding the going difficult will omit these starred footnotes
as a matter of course. Those who like challenges will attempt to under-
stand them (and will succeed most of the time!) but they should never be
discouraged if a starred note appears too difficult: it can always be omitted

without in any way affecting the reader' s ability to follow any subsequent argument in
the text.

The student should seek to master completely one text-book, but.

XVlIl THE USE OF THIS BOOK
generally, he should not confine himself exclusively to this one book,
he
should read sections m
other texts, parucularly those sections dealing with
ideas that he finds difficult Tlierc arc many first-class books to which he
can make reference and only a few can be mentioned here, and unfortu-
nately there is an even larger number of not-very-good books and the
student should seek expert advice before adopting a book for major study
Finally I should like to say a word of thanks to all those people who have
made this book possible In so far as the ideas and viewpoints expressed
here are novel, they are the common property of all my colleagues who are
members of the L S E Sta^ Semtnor on Mtthodology, MeasuTtment and Testing
in Fconomics /Ml tint I did m the 1st edition w.is to give a slightly persona!
expression to this general viewpoint Mr K Klappholz read the manuscript
of the removing countless blemishes, and contnbuted greatly to
1st edition,

some of the novel ideas expressed herein Mr R Cassen read the proofs and
contnbuted many last-minute improvements Mr G C Archibald gave de-
tailed scrutiny of the proofs of Fconomics and his penetrating cnticisms led
to many improvements The executor of Lord Beveridge’s estate has kindly
granted permission to quote extracts from Devendge’s farewell address at
the London School of Economics
Thanks are also due to Macmillan and McGraw-Hill for permission to
quote from Lord Robhins' Fssay on the Nature and Significance of Economic
Seunce and Sir Roy Harrod's The British Economy respectively Harper and
Row has also been generous m giving their permission to quote at length
from material fint prepared for Economics I am also extremely grateful to
the many and tcachen, who have taken the trouble to write
users, students
to me pointing out errors, making comments and suggestions Economics
IS a subject m which one never stops learning and it is always gratifying to

realise that one can still Icam from one’s students I hope that the readers
of this book will continue to teach me with as many further comments and
criticisms as they have in the past Those who have done so are too numerous
to mention, but I must at least single out Mr John Knapp of the University

of Manchester who has, together with his students, contributed many con-

structive criticisms and suggestions for improvement The largest single

debt of all I owe to Professor P O Steiner His contribution to the present

book IS too large to be done justice to by a one-sentence
acknowledgement
here, and I hav c made it the subject ofa separate note on page xxii Mr V M
Blandon and Mrs Wickens gave valuable research help on the 1st edition,
J
while Mr P Geary and Mr J Stilwell give extensive research assistance to
the 2nd edition Mrs S Craig, Mrs Evelyn Dean and Miss Tina Brown
have
shown unlimited patience in dcabng with the manuscript The usual dis-
claimer of course holds here for all shortcomings and mistakes remaining
I am solely to blame
NOTES ON THE SECOND EDITION

The second edition of An Introduction to Positive Economics is based on a

thorough rewrite of the text book Economics which was written for the
American market by P.O. Steiner and myself, which book was in turn
based on a complete rewrite of the first edition of An Introduction to Positive
Economics.A little more detail concerning the relation between the present
book and the American text is given at the end of this note, in the meantime
my concern is to explain to users of Positive Economics the most significant of
the changes made between the and the present edition.
first

The present edition, although still short by the standards of American

text books, is much longer than the first edition. The increased length is

mainly accounted for by increased applications of the theory, rather than

by extensions of the theoretical content over the first edition. In this version
I have tried to come substantially closer to the target that I originally set

myself of eroding some of the barriers between theoretical and applied

economics. I have tried to remove every bit of theory that I could not show
operating in real situations, and as the other side of the same coin I have
tried to give applications of every bit of theory that I did include. This latter
task has been very consuming of space and accounts for the bulk of the
addition to length between the two editions.
In some places, however, I have added to the coverage of theoretical

topics particularly where there were important real-world problems that

could not be handled given only the original coverage. A great deal, for
example, has been added on the theory of costs which allows the student to
deal more adequately with some of the aspects of the theory of the firm and
with those problems of public policy that are based on a distinction between
private and social costs. The theory of the firm has been extended to cover
both price discrimination and the modern theory" of oligopoly. The latter
allows one to put problems of public control more in perspective than can
be done when only the theories of competition and monopoly are available.
In the theory of demand, while still avoiding indifference curves, I have
added a long section on marginal utility theory'. I have done this because
the mileage one can get in terms of policy applications at an elementary
level seems to me to be much higher with utility theory than with either of
XX NOTES ON THE SECOND EDITION

the other two theories of demand Indeed one of the insights of economics
that IS m public debates seems to me to be the one based
of most assistance
on the between total and marginal utility I hope that m
distinction
Chapters 15 and 38 I have gone some way towards justifying this statement
and towards showing that this distinction is one that investigators m other
social sciences would do well to keep mind m
In the first Chapter I have abandoned the Poppenan notion of refuta-
tion and have gone over to a statistical view of testing that accepts that
neither refutation nor confirmaDon can ever be final, and that all we can
hope to do IS to discover on the basis of finite amounts of imperfect know-
ledge what IS the balance of probabilities between competing hypotheses
Although I was not far from this view by the time I completed the first
edition, several years of experience with applied empincal work have con-
firmed me to It When I came to rewnte the chapters
on Testing that con-
cludes each part I was nevertheless amazed by how much this change
helped to remove some of the difficulties that presented themselves to some-
one proceeding in the Poppenan mode The chapter on measurement and
tests of the theory of demand is a particular case m point Some problems

that seem intractable to a believer in single-observation refutations yielded

quickly to the person approaching them with the concept in mind of
probabilistic judgments between competing hypotheses
A chapter has been added on statistical analysis This chapter makes no
pretense at teaching the theory of statistics but it does try to provide a
bridge between the economics the student will learn m this book and the
he will learn elsewhere The very least I want the student to go
statistics

away with is the knowledge that some of the best brains of the twentieth
century have been applied to the problem of how to be ‘scientific’ when all
the available observations come from non-laboratory situations For the
student to dismiss non-laboratory subjects as irretrievably unscientific is
Ignorant if nothing else One of the purposes of the chapter is to persuade
the student that he should at least suspend judgment until he has acquired
further knowledge of the staDstics and econometrics
In the first have given, largely under the influence of
half of Part 4 I
a
Professor Steiner my co author of the American book Economics, quite
large dose of opportunity cost theory This has led to the
abandonment of
the concept of ‘normal profits’ and the substitution of ‘the opportunity cost
of capital’ in its place ‘Profiu’ now refers to what is
sometimes called ‘pure
profits’ the returns over all opportumtj costs which, in competili\c situa-
associated with the dynamic forces of change I have continued to
tions, IS
different
emphasise the distinction between relative returns to one factor in
return lo
industries and the aggregate returns to one factor rcHtive to the
problem of the theory of micro allocation, the
other factors The first is part
 NOTES ON THE SECOND EDITION XXI,

theor)' of which is well developed, while the second problem is part of the
theor}-of macro distribution which is not well developed. Thus if someone
asks me: ‘how are profits determined?’ my answer is: ‘if you mean total
profits (as a share of total national income), I don’t know and neither does
any other economist, while if you mean relative profits in different indus-
tries I can -have a shot at telling you (although I will first have to ascertain
if when you talk about ‘profits’ you mean pure profits or the opportunity
cost of capital, or both)’.

There a new chapter on Labour Unions and wages which attempts to

is

enlighten some real problems with a systematic use of elementary theory'.

The chapter on tests of the theory of distribution has been expanded to go
beyond simple criticisms to some discussion of the empirical relevance of
the theory. The micro half of the book is concluded with a long and com-
pletely new chapter on micro economic policy.’
The whole section on macro economics has been drastically revised and
lengthened. Part 7 which contains the guts of the simple macro model now
contains eight chapters. The chapter on the model of the circular flow has
been simplified greatly. While not retracting anything in the original
version, experience has shown that much of the detailed treatment was
more suitable to an intermediate than to an introductory course.
Chapter 41 on equilibrium in the circular flow is unlike the treatment in
most other text books. In my opinion most introductions contain crude
fallacies. I have, however, tried to avoid criticisms of other treatments as

inappropriate for a text book and I hope that (since important divergencies
appear only in the treatment of dynamics and not in statics) the rest of the
book is still available to someone who disagrees with my approach here. I
have removed the appendix on savings and investment and have argued
the whole case of errors in the standard treatment where it is more suitable:
in a forthcoming journal article.^

1 I find it quite impossible to accept Professor Hick’s view that it is impossible to talk about
policy issues without the (in my opinion largely sterile) apparatus of the new welfare economics.
If the amount of policy in the present book is not a refutation, surely Christopher Dows’ book
on The Management of the British Economy, which is all about policy without a mention of welfare
economics, is.
2 Anyone who is still worried about savings and investment being actually unequal should
contemplate the following one-injection, one-withdrawal model;

(1) Y = C+T
(2) 0 = C+G
(3) y= 0
, From which we deduce that T= G
This is formally identical with the /=5 modek. Do we really believe that although planned
XXII NOTFS ON THE SECOND EDITION

A great deal lias been added on the

demand for and the supply of money
and the chapter on the price lc\cl has been completely rewritten and I hope
that the argument is now much clearer than it was A chapter
on inter-
national economic experience Ins been added in the hope that with
4
chapters of trade theory under hts belt the student will be able to
make
sense of some of the key issues of twentieth century trade policy
The book concludes with a new chapter on macro economic policy The
student who gets this far m the book should get real saluc for his efforts by
realising just how many of current policy issues he is now able to under-
stand I base tried not to advocate my own view that the Bntish people
are being saenficed quite unnecessarily and quite uselessly on the cross of
'the sanctity 0
/ the pound' yust as much m
19G6 as they were 1926 Tlie m
advocacy of such persona! views is an abuse of a text book but, if they have
shown through in this chapter or elsewhere, I can only repeat what I say
in my lectures ‘The object of .a good introductory course m economics is
tb leach you some economic theory that is generally accepted to teach you
enough about current problems so you sec the issues that are involved m
them, and to teach you enough about what 1 $ not known to make you realise
that any strongly held view on a complex matter of policy will contain a
sufficient element of guess, hunch and judgment that no matter how
great the prestige of the holder the validity of his view can never be taken
as certain Positne economics seelis to, and does succeed in, namning the range of
uncertainty on issues that concern us, hut it can never reduce the uncertainty to zero'

4 note on the delation Delueen 'Economics’ and VJn Introduction to Positne

Economics' Shortly after the publication of Positne Economics the question
arose of publishing an Amencan version TTic fint idea was to change
pounds shillings and pence to dollars and cents in the theoretical sections
and to add matcrnl on American institutions in the descriptive sections I
was insistent from the outset that institution'll material could not just be
tacked on but had to be integrated into the theoretical treatment with the
question ‘Docs the institution affect the behaviour we arc considering’
kept to the forefront I was lucky to be able to enhst the serv ices of my friend
Professor P Steiner of the Dnivcrsily of Wisconsin for this task As soon
O
as we began ourtask it became apparent that wc would need to do a com-
plete rcwnic We would use maienal from Positive Economics but we would
produce a more or less new book In the event the task took two years and
the result was published in 1966 under the title Economics by R
G Lipsey

goiemmftital expendilures may diverge ptannwi lax revenues actual government ex

’
>s always balanced
pend.tures arc always equal to actual la* revenues ex post the budget
Yet, since ihis modelis formally idenlicat to the usual savings
and investment model every
and investment you must also be prepared to
thing you say about ex ante and ex post savings
say about C and Tl
 NOTES ON THE SECOND EDITION Xxiii

and P.O. Steiner. Both of us worked and reworked all the material until
the contribution of each author became quite unidentifiable. Both of us
had a lot to lose in the effort since on the one hand Steiner might become
known as the person who had changed the pound signs into dollars in
Lipsey’s book while on the other hand Steiner might be thought to be
responsible for all of the changes between the first edition of Positive
Economics and Economics many of which I had already planned before the
collaboration took place. In the event the work became a true joint effort
and for anyone who knows both books it is both authors’ opinion that they
are exactly 50/50 responsible for all the changes between Positive Economics
and Economics.
No sooner was a manuscript of Economics completed than I began to
prepare a second edition of Positive Economics. Although I am the sole author
of the changes between Economics and the second edition Positive Economics

it goes without saying that I owe an enormous debt to Professor Steiner for

his part in creating the present version of this book. I have already men-
tioned that it became almost impossible to sort out our separate contri-
butions to Economics-, thus Professor Steiner’s influence persists throughout
the whole of the present book. It is nonetheless possible to identify his
influence as being particularly strong in Part 4.

Although this short historical note deals with a ver)' personal matter I

have felt that I could not do justice to Professor Steiner’s contribution to

the present book without it. To my collaborator, and in many places my
teacher, I and the substance of the present book owe a very great deal.

The University of Essex R.G.L.

 PART 1

SCOPE AND METHOD

CHAPTER 1

INTRODUCTION

Why is the histor>' of most capitalist countries one of several years of boom
and plenty followed by several years of depression and unemployment with
consequent poverty for a great many of its citizens? Why, during the 1930’s
in most capitalist countries, was there a ten-year period in which up to one
person in four was unemployed while factories lay idle and raw materials
went unused, when, in short, everything was available to produce goods
which were urgently required and yet nothing happened? Are the Marxists
correct in arguing that a return to such a situation of mass unemployment
is something from which the capitalist countries are saved only by their high

levels of arms expenditure ? Or is it true that, due to the work of an English

academic economist tvridng 30 years ago from King’s College, Cambridge,
we now have it in our power to prevent such a situation of mass unemploy-
ment from ever occurring again?
Why do many countries have balance of payments crises ? What is the
point of international trade anyway, and would we not be better off if we
made ourselves self-sufficient? Were the majority of academic economists
right in saying that Britain should have devalued the pound in 1965, or were
the majority of the international banking community right in arguing that
devaluation would have been a disaster? Both of these groups could not
have been right!
Why, in the history of the world, have there been periods of rapidly rising
prices alternating with periods of stable and sometimes falling prices ? Why
do the prices of some commodities fluctuate tvidely while the prices of others
are relatively stable? Why is it that, as tvith many agricultural products,
price fluctuations give rise to large variations in the incomes of those who
produce them, while, with some other products, price fluctuations hardly
cause any variation in producers’ incomes ?
What determines the level of wages and what influences do unions have
on the share of income going to labour? What functions do unions fulfil in
today’s world, and have they not long ago fully fulfilled their his
 INTRODUCTION 5

isgood and what is bad. We say that normative statements depend upon
our VALUE JUDGMENTS.
Disagreements may arise over normative statements because different
individuals have different ideas of what is good and bad and thus of what

constitutes the good life. Disagreements over normative statements cannot be settled
merely by an appeal to facts.
The between positive and normative may be clarified by con-
distinction
sidering some assertions, questions and hypotheses that can be classified as
positive and normative. The statement ‘it is impossible to break up atoms’ is
a positive statement which can quite definitely be (and of course has been)
refuted by empirical observations, while the statement ‘scientists ought not
to break up atoms’ is a normative statement that involves ethical j'udgments.
The questions ‘What government policies will reduce unemployment?’ and
‘What policies will prevent inflation ?’ are positive ones, while the question
‘Ought we to be more concerned about unemployment than about infla-
tion?’ is a normative one. The statement ‘a government deficit will reduce
unemployment and cause an increase in prices’ is a very simple hypothesis
in positive economics, an hypothesis that could be tested by an appeal to
empirical observation, while the statement ‘in setting policy, unemployment
ought to matter more than inflation’ is a normative hypothesis which cannot
be settled solely by an appeal to observation.^

1 Having grasped this distinction, the student must beware that he does not turn it into an
enquir>'-stopping, dogmatic rule. From the fact that positive economics does not include
normative questions (because its tools arc inappropriate to them) it docs not follow that the
student of positive economics must stop his enquiry as soon as someone says the word ought.
Consider the statement; ‘It is my value judgment that we ought to have rent control because
controls are good.' Now it is quite in order for you as a practitioner of positive economics to ask
‘Why?’, It may then be argued that controls have certain consequences and it is these conse-
quences which are judged be good. But the statements about the consequences of rent control
to

will be positive testable statements. Thus the pursuit of what appears to be a normative state-
ment will often turn up positive hypotheses on which our ought conclusion depends. There are,
for example, probably few people who believe that government control of industry is in itself
good or bad. Their advocacy or opposition will be based on certain beliefs about relations
which can be stated as ppsitive rather than normative hypotheses. For example: ‘Government
control reduces (increases) efficiency, changes (docs not change) the distribution of income,
leads (does not lead) to an increase of state control in other spheres.’ A careful study of this

very emotive subject will reveal an agenda for positive economic enquiry which could keep a
research team of economists occupied for the next ten or 20 years.
Philosopher friends have persuaded me that, when pushed to its limits, the distinetion
between positive and normative becomes blurred, or else breaks down completely. The reason
for this is that when examined carefully most apparently normative statements reveal some
positive underpinning (e.g., ‘Unemployment is worse than inflation because the (measurable)
effects of unemployment on human beings arc judged by the majority of adult citizens to be

more serious than the (measurable) effects of inflation’). I remain convinced, however, that, at
this stage of the development of economics, the distinction is a necessary working rule the
6 SCOPE AND METHOD
The disunction between
positive and normative follows from the fact that
It ISlogically impossible to deduce normative statements from positive
as
sumptions and vice vtrsa Thus if I thmk something ought to be done, I can
deduce other things which, if I wish to be consistent, ought to be done, but
I can deduce nothing about what is done (i c is true) On the other hand,
,

if I know two things are true, I can deduce other things which must be
that
true, butcan deduce nothing about what is desirable (i c , ought to be)
I

Consider a simple example Suppose I believe (1) that it is a moral pnn-

ciple that one ought to be charitable to all human beings Then if I am told

(2) that the inhabitants of China are not Christians but are human beings
then It follows (3) that one ought to be charitable towards Chinese We have
thus deduced from (1) and (2) a normative principle about how we ought
to behave, no positive statement about how we do behave can be deduced
from (1) and (2) Now, suppose someone else comes along and says, ‘You
ought not to be charitable toward the Chinese because moral pnnciples
dictate that you should only be charitable toward Chnstians Ifwcnowgct ’

into an argument about whether or not we should be chantable toward the

Chinese this argument will turn on our value judgments about how one
ought to behave These are questions on which reasonable people sometimes
just have to agree to disagree If both sides insist on holding to their views on
chanty, and even if both are perfectly reasonable men, there is no civilized
way of forcing one to admit he is wrong
Now, assume 1 say (1) that capital punishment is a strong disincentive to
murder and (2) that the Chinese abolished capital punishment after the
Revolution so that (3) the number of murders must have nsen in China
since the Revolution The two factual statements, (1) and (2), and the de
duction that follows from them arc all positive statements We can deduce
nothing about the moral desirability of abolishing capital punishment from
statements (1) and (2), even if we were certain they were factually correct
Now let us say someone else comes along and say's, ‘The number of murders ’

has not nsen in China since the Revolution, in fact the number has fallen
If he holds to this view he must deny one or other of my first two positive
statements He might deny statement (1) by saying, for example, that capital
punishment is actually an incentive to commit murder,* or he might deny
although the Chinese pretended to abolish
statement (2) by saying that,
after the
capital punishment as a propaganda move, in fact they retained it

to confusion rather than to clanty The

present abandonment of which would contnbutc
juiuficauon for this vnew is that although sec are not sure what to
make of an apparently
know a purely
normauve statement (because it may have a ponuve underpinning) we do
positive statement when we see one
I Strange though it may seem at fim oght this
wew has been senously advocated and a
strong pfxma Jaeu case may be made out m iB favour See Sidney Silverman Hangtdhtl Iwtxmt
 INTRODUCTION 7

Revolution. In both cases the disagreement is over factual statements. If

we gathered enough facts and if both parties were reasonable, one side could
be forced to admit it was wrong.
Economics, in common with other sciences, is concerned with questions,
statements and hypotheses that could conceivably be shown to be wrong
(i.e., falsified) by actual observations of the world. We do not have to show

such statements to be consistent or inconsistent with the facts tomorrow or

the next day we only have to conceive of the existence of factual evidence
;

that could show them to be wrong. Thus an appeal to the facts is an appro-
priate way in which to deal with them.
Normative questions cannot be settled by a mere appeal to empirical
observation. This does not, of course, mean that they are unimportant. We
must decide such questions as ‘Should we permit the existence of a private
educational system alongside of the State one?’ and ‘Should people who
cannot obtain jobs because of their own personal failings be shielded from
the full consequences of their failure or should they be left to starve?’, but
we must decide them by means other than a simple appeal to facts. In
practice, we on them.
usually tend to settle such questions by voting
So far then, we have said that the separation of the positive from the nor-
mative is one of the foundation stones of science and that scientific enquiry,
as we normally understand it, is confined to positive questions. We must
now consider in more detail just what the scientific approach is and how
scientific theories are developed and used.

THE SCIENTIFIC APPROACH

Very roughly speaking, the scientific approach consists in relating questions
to evidence. When presented with a controversial issue, the scientist will ask
what is the evidence both for and against. He may then take a stand on the
issue, the stand being taken with more or less conviction depending on the
weight of the evidence. If there is little or no evidence, the scientist will say
that, at present, it is impossible to take a stand. He will then set about
searching for relevant evidence. If he finds that the issue is framed in terms
that make it impossible to gather evidence for or against it, he will then

usually try to recast the question so that it can be answered by an appeal to

the evidence.^ This approach to a problem is what sets scientific enquiries

off from other enquiries.^

1 One of the really challenging problems to the scientist is to find out how to pose a question
in the general spirit of the problem which people are interested and in a form capable of
in

being answered by reference to evidence. There is no formula for this ; it is a real art and one
of the most difficult of all problems.
2 Other approaches might be to appeal to authority, for example, to Aristotle or the
 SCOPE AND METHOD

In some the scientist, having reframed the question is then able to

fields,
generate observations that will provide evidence for or against the hypo
thesis Fxperimenlal sciences, such as chemistry and some branches
of psycho-
logy, have an advantage because it is possible for them to produce relevant
evidence through controlled laboratory experiments Other sciences such
as astronomy and economics, cannot do this They must wait for time to
throw up observations that may be used as evidence in testing their theories
The ease or difficulty with which one can collect or even manufacture
evidence docs not determine whether a subject is scientific or nonscientific
although many people believe that it does, it is merely one of the factors
determining the degree of ease with which the scientific enquiries of \ arious
fields can be pursued ‘ The way in which scientific enquiry proceeds does,
however, differ radically between fields in which laboratory experiment is
possible and those in which it is not In this chapter we consider general
problems more or less common to all sciences In Chapter 3 we shall deal
with problems peculiar to the nonexpenmental sciences, which must accept
observations in the forms in which they are thrown up by the world of actual
experience

The Scientific Attitude in Everyday Life

It IS often said that we live m a scientific age Over the last several hundred
years the citizens of most Western countries have enjoyed the fruits of
innumerable scientific discovenes But the scientific advances that have so
profoundly affected the average citizen have been made by an extremely
small minoniy of the population These advances have gencrall) been ac-
cepted without even the slightest idea either of the technical nature of the
discoveries involved, or of the attitude of mind that made them possible
If wc take as a measure of the influence of science the degree of dissemina-
tion of the fruits of science, then we live in a profoundly scientific age but
if we take as our measure the degree to which the general public under-

Senptures to appeal by introspection to some inner eapenence (to start oir all' reasonable men
nature of the problem or
will surely agree ) or to proceed by way of definitions to the true
concepts under consideration
1 It 18 often thought that scientific procedure consists of gnndmg out
answen with reference
to blind rules of calculation and that it is «Jy m the arO that the exercise of real imagination

is misguided for there arc no set

rules for the framing of quesUons It is
IS required This view
a step that often requires great imagination Also the collection
of relevant evidence often

requires ingenuity (e g the Michtlson Morley expenmenl) hat the scientific method gives

IS an impersonal set answenng some questions but what questions to ask and
ofcniena for
exactly how to ask them and exactly how to obuin the evidence are dilTerent problems

requiring upon occasion great feats of imaginauon

 :

INTRODUCTION 9

stands and we are definitely in a pre-

practises the scientific approach, then
scientific era. Indeed, the scientific method of answering questions by
appealing to a carefully collected and coordinated body of facts is a method
that is seldom adopted by the public.
Consider, for example, the current argument about capital punishment.
It is possible toadvocate capital punishment as an act of pure vengeance or
because we believe that a good moral code requires per se that a person who
kills ought himself to be killed. If we argue about capital punishment on

these grounds, we are involved in normative questions depending upon

value judgments. But the great majority of arguments for capital punish-
ment are not of this type, but rather are predictions about observable be-
haviour, and thus belong to the field of science. These are usually variants
of the general argument that capital punishment is a deterrent to murder. In this
general form it is probably very' difficult to test the proposition and it will
be necessary' to state a number of more specific propositions which fall under
this general hypothesis. Consider one such example

If there is capital punishment for murder involving robbery,

then the robber will be less inclined to take a lethal weapon
with him on his mission. If he does not take a lethal
weapon \vith him, he will in fact commit fewer murders
when surprised in the course of his robbery.

It is truly amazing how people can become committed to answers to such

questions without a consideration of the available evidence. A survey of the
press whenever the issue arises will show that most of the reasons given for
saying yes or no to such questions are profoundly unscientific. How many
of the participants know, for example, what proportion of murders in the
course of robbery’ are made wth lethal weapons brought by the criminal to
the scene of the crime and what proportion are made %vith anything found
at hand after the criminal has been discovered? Yet it would seem to be
impossible to have an informed discussion on the issue without this element-
ary piece of factual knowledge.
Indeed a study of most of the arguments and against capital punish-
for
ment svill usually reveal a maximum of empirical questions and a minimum
of empirical evidence used to arrive at the answers given. If we really be-
lieved in a scientific enquiry into human behaviour we would try to state
the arguments about capital punishment in terms of a specific set of pro-
positions and would then set out systematically' to gather evidence relating
to each of these propositions. We
may' conclude that many hotly debated
issues of public policy are positive and not normative issues, but that the
scientific approach to such positive questions is very' often ignored.

1*
10 SCOPE AND METHOD

A SCIENCE OF HUMAN BEHAVIOUR?

The preceding discussion raises the question of whether or not it is possible
tohave a scientific study in the field of human behaviour When considering
whether or not a scientific study of such subjects as the causes of unemploy-
ment and the consequences of capital punishment can be made, it is often
argued that natural sciences deal with inanimate matter that is subject to
natural laws’, while the social saences deal with man who has free will and
cannot, therefore, be made the subject of (inexorable) laws

Is Human Behaviour Predictable?

Stated carefully the above view implies that inanimate matter will show
stable responses to certain stimuli, while animate matter will not For
example, if you put a match to a dry piece of paper the paper will bum,

while you stibject human beings to torture some will break down and do
if

what you want them to do and others will not, and, more confusing the
same individual may react differently to torture at different times Whether
human behaviour does or does not show sufficiently stable responses to
factors influencing u as to be predictable within an acceptable margin of
error is a positive question which can only be settled by an appeal to
evidence and not by a priori speculation ‘

Deterministic and statistical hypotheses Before we consider the

state of the evidence on this matter we must distinguish between two kinds
of hypotheses, deterministic and statistical ones Deterministic hypotheses
admit of no exceptions An example of such a hypothesis would be the state-
ment ‘If you torture any man over this penod of time with these methods he
will alwaysbreak down Statistical hypothecs admit of exceptions and
’

purport to predict the probability of a certain occurrence An example

would be ‘If you torture a man over this period of time with these methods
he will very probably break down — m fact if you torture a large number of
men under the stated circumstances about 95 per cent of them will
break down ’
In such an hypothesis wc do not purport to predict what
an individual will certainly do but only what he will probably do This
of error
does allow us however, to predict within a determinable margin
what a large group of individuals will do

Stability in human behaviour In feet it is a matter of simple observa

tion that when we consider a group of individuals they do not behave in a
to vanous stimuli
totally capncious way but do display stable responses
prior
It may be defined as that which
is
I A pTton IS a phrase commonly used hf economists
to actual experience or as that which is innate or based on innate ideas
 INTRODUCTION II

which act on them. The warmer the weather for example, the higher the
number of people visiting the beach and the higher the sales of ice-cream. It
may be hard to say when or why one individual will buy an ice-cream but
we can observe a stable response pattern from a large group of individuals
the higher the temperature the greater the sales of ice-cream.
'

Many other examples will come to mind where, because we can say what
the individual will probably do - without being certain of what he will do -
we can say -with quite remarkable accuracy what a large group of individuals
will do.
No social scientist could predict, for example, when an apparently
healthy individual was going to die, but death rates for large groups are
stable enough to make life insurance a profitable business. It could not be
so if group behaviour were capricious. Also, no social scientist can predict
what particular individuals will be killed in auto accidents next holiday,
but he can come very close to knowing how many in total will die, and the
more objectively measurable data he is given concerning, for example, the
state of the weather on the day, and the increase in auto sales over the last
year, the closer he will be able to predict the total of deaths. If group human
behaviour were in fact random and capricious there would be no point in
trying to predict anything on the basis of sample surveys. The fact that
80 per cent of the people sampled said they intended to vote for a certain
candidate would give no information about the probable outcome of the
election. Today’s information might be totally reversed tomorrow. That
there are discernible trends in election polls and by-elections is proof of the
fact that in politics people do not act at random.
An example that has been important recently is that economists can
predict with fair accuracy what households as a group will do when they
find their income increased. Of course, some individuals may do surprising,
and as far as we can see, unpredictable things, but the total response of all
households to a change in tax rates which leaves more money in their hands
is predictable within quite a narrow margin of error. This stability in the

response of the public’s expenditure to a change in their available income

was the basis of American economists’ ability to predict successfully the out-
come of the major tax reduction first proposed by the late President
Kennedy and finally passed under the Johnson administration.*
1 Of course, it does not follow from anything that has been said so far that people never
change their minds, or that future events can be foretold by a casual study of the past. Students
sometimes think in terms of a simple dichotomy: either there are historical laws apparent to the
casual observer or there is random behaviour. They observe a prophet extrapolating a trend
(i.e., predicting that some change will take place in the future merely because it took place in
the past) and, seeing him make an utterly mistaken prophecy, conclude that, because the
prophet cannot prophesy, human behaviour is random and thus unamenable to scientific

study. The stability we are discussing is a stable response to causal factors (e.g., next time it
12 SCOPE AND METHOD

The 'Law' of Large Numbers

We may now ask how it is that we can predict group behaviour while
never
being certain what any single individual will do Successful predictions
about the behaviour of Urge groups arc made possible by the so-called
statistical ‘law’ of large numbers Very roughly this ‘law’ asserts
that random
movements of a large number of individual items tend to offset one another
This law IS based on one of the most beautiful constants of behaviour in the
whole of science, natural and social, and yet it can be derived from the fact
that human beings make errors’ The ‘law’ u based on the normal curie of
errorwhich the student will encounter in elementary statistics
Let us connder what is implied by this ‘law’ Ask any one person to
measure the length of a room and it will be almost impossible to predict in
advance what sort of error of measurement he will make Thousands of
things will affect the accuracy of his measurements and, furthermore, be
may make one error today and quite a different one tomorrow But ask
one thousand people to measure the length of the same room and we can
predict vnthin a very small mai^n of error how this group will make its
errors' We can assert with confidence that more people will make small
errors than will make large errors, that the larger the error the fewer will
be the number of people making it, that the same number of people will

overestimate as will underestimate the distance, and that the average error
of all the individuals will be zero Here then is a truly remarkable constant
'

pattern of human behaviour, a constant on which much of the theory of

statistical inference is based
common cause should act on all members of the group wc can predict
If a
how the average behaviour of the group will react even though
successfully
any one member of the group may act in a surprising fashion If, for ex-
ample, we give all our thousand individuals a tape measure which under-
states ‘actual’ distances, wc can predict that, on the average, the group will
now understate the length of the room It is, of course, quite possible that
one member who had in the past been consistently undermeasunng distance

gets warm,ice cream safes wiff nsej aad not a mere staftrfcfy of trerrJ cfiangw g .
ertam
sales willgo on nsing in the future because they have risen in the pail)
measured by
1 For purposes of measuring the error we define the true distance to be that
will be very small
the most precise instruments of scientific measorenienl (whose range of error
the of error of our one thousand bymen all wielding tape measures)
Those
relative to range
familiar with statistical theory will reahse that the predictions in the text
assume that all the
factors causing
necessary conditions such as the existence rf a Urge number of independent
individuals to make errors are fulfilled The purpose of the discussion m
the text is not to give

the student full appreciauon of the subtleties of statistical theory but to persuade
him that the
commonly held view that free will and the absence of deterministic certainty about human
behaviour makes a scientific study of human beha«our impossible, is misguided
 INTRODUCTION 13

because he was depressed psychologically, will now overestimate the dist-

ance because the state of his health has changed; but something else may
happen to some other individual that will turn him from an overmeasurer
into an undermeasurer. Individuals may do peculiar things for reasons,
which, as far as we can see, are inexplicable, but the group’s behaviour,
when the inaccurate tape is substituted for the accurate one, will nonetheless
be predictable, precisely because the odd things that one individual does will tend to
cancel out the odd things some other individual does.

THE NATURE OF SCIENTIFIC THEORIES

So far we have seen that there is real evidence that human behaviour does
show' stable response patterns. We must now consider how we can set about
explaining this behaviour so that we can predict the outcome of certain
events and so that we can intervene to change certain occurrences that we
do not like. Such explanations are prowded by theories.
Theories grow' up in answer to the question ‘Why?’. Some sequence of
events, some regularity between two or more things is observed in the real
w’orld and someone asks why this should be so. A theor>' attempts to explain
w'hy. Whether or not a theory takes us any nearer to an understanding of
‘ultimate reality’ (whatever we may understand by this concept) is a very
difficult philosophical question. Whatever may be the answer to this
question, one of the main practical consequences of a theory is that it en-
ables us to predict as yet unobser\'ed events. Thus, for example, national
income theory predicts that a government budget deficit^ will reduce the
volume of unemployment.^ The simple theory of market behaviour predicts
that, under certain specified conditions, the introduction of a sales tax will
be accompanied by an increase in the price of the commodity concerned
and that the price increase will be less than the amount of the tax. It also
allows us to predict that, if there a partial failure of the potato crop, the
is

total income earned by potato farmers rvill increase Thus a theory arises
!

in an attempt to explain, or to account for, certain observed phenomena,

and a successful theory has the major practical result that it enables us to
predict in advance the consequences of various occurrences.

The Pervasiveness of Theories

All we actually observe in the world is a sequence of events. Any explana-
tion w'hatsoever of how these events are linked together is a theoretical con-
struct. Theories are what w'e use to impose order on our observations, to

1 A budget deficit arises when the government spends more than it raises by way of taxes.

2 This is not so simple a prediction as might appear at first sight. What it means and how it
can be tested is considered in Chapter 58,
 14 SCOPE AND METHOD
explain how what we see is
linked together Without theories we would only
have a shapeless mass of meaningless observations If we arc to make any
sense at all of what we see, the choice is not one between theory and
observation but between better orworsc theories to explain our observations
In a particular case we might see a change m
company taxation followed
by some change m the behaviour of companies The practical man may
think the link obvious and indeed in some sense it may be but nonetheless
is

It requires a theoretical construction Before we can link these two

events
together we need a theory of what the managers are trying to do, and of
how they try to do it, plus the assumption that the managers know what
behaviour will achieve their goals Such theories of the behaviour of firms
arc considered m Part V of the present book

True in theory but not in practice Misunderstanding about the

place of theories m scientific explanation gi\es nse lo many misconceptions,
not the least of which is the common prejudice against theories and the
belief that they can be successfully dispensed with, as evidenced by the
statement in the quoted heading The next time you hear someone say,
‘That may be true in theory but not in practice (or, indeed, the next time
you say it yourself) you should immediately reply, 'All nght then, tel! me
what docs happen in practice ’ It is almost certain that vou will not be told
mere facts, but that you will be given an alternative theory - a different
explanation of the facts The speaker should have said, ‘The theory m ques-
tion provides a poor explanation of the facts {i e ,
it is contradicted by some
factual observations) and that his alternative theory is a better one

The Construction of Theories

A theory consists of a set of definitions, stating clearly what we mean by

and a set of assumptions about the way m which the world
various terms,
behaves Having defined terms and made assumptions about behaviour, the
is im-
next step is to follow a process of logical deduction to discover what

plied by these assumptions For example, if we assume that businessmen

always try to make as much profit as is possible and if we make assumptions

about how taxes affect their profits, wc can denvc implications about
how
the predictions
they will behave when taxes change These implications are
next changed,
of our theory, we predict for example that when taxes are

businessmen will react in the way implied by our theory

concern Consider a theory

1* Assumptions about behaviour often cause ihe student real
this assumption
that surts out Assume that there is no government Surely says the reader
takesenouslyanythmg that comes out of the theory ut u
IStotally unrealistic and I cannot
whatever the government oes
assumption may merely be the econoroma way ofsaying that
even whether or not itexisu uirrilaanlS0rih*pinp<>stsefhup<irUcula,thto,)
Now put this way
 INTRODUCTION 15

The Nature of Scientific Predictions

We have seen above that a successful theor}' enables us to predict as yet

unobserved events. We must now consider with a little more care just what
is the nature of a scientific prediction, and in particular if it is the same thing

as being able to prophesy the future course of events.

The critical thing to notice about a scientific prediction is that it is a
conditional statement of the form '’if you do this then such and such will

follow’. If you mix hydrogen and oxygen under specified conditions, then
water will be the result. If the government has a large budget deficit, then
the volume of employment will be increased. It is most important to notice
that this prediction is v'er)' different from the statement: T prophesy that in
two years’ time there udll be a large increase in employment because I
believe the government will decide to have a large budget deficit.’ The
government’s decision to have a budget deficit or surplus in Uvo years’ time
will be the outcome of many complex factors, emotions, objective circum-
stances, and chance occurrences, which cannot be predicted by the econo-
mist. If the economist’s prophecy about the level of employment turns out
to be wrong because in two years’ time the government does not have a
large deficit, then all we have learned is that the economist is not a good
guesser about the behaviour of the government; we will not have refuted
any economic theory. However, if the government does have a large deficit
(in two years’ time or at any other time) and then the volume of employment

does not rise, we have refuted a (conditional) scientific prediction in the

the statement becomes an empirical assertion, and the only way to test it is to see if the pre-

dictions which follow from the theory do or do not fit the facts that the theory is trying to
explain. If they do, then the theorist was correct in his assumption that the government could
be ignored; the criticism that the theory is unrealistic because we know that there really is a
government is completely beside the point. Assumptions, horvever, are used in economics for
other purposes, particularly to outline the set of conditiorrs under which a theory is meant to

hold. Consider a theory that assumes that the government has a balanced budget. This may
mean that the theorist intends his theory to apply onlywhen there is a balanced budget; it
may not mean that the size of the government’s budget surplus or deficit is irrelevant to the
theory. The student may find it confusing that an assumption may mean many different things
in economics. he encounters an assumption in economic theory he should, therefore, do
When
tsvo things; ask what information the assumption is intended to convey, and remember that it
is not always appropriate to criticise the simplifying assumptions of a theory on the grounds

that they are unrealistic. It is important to remember that all theory' is an abstraction from
reality. If we did not abstract we would merely duplicate the world and would add nothing to

our understanding of it. A good theory abstracts in a useful and significant way; a bad theory
does not. If the student believes that the theorist has assumed away something that is important
for theproblem at hand, then he must beh'eve, and try to show, that the conclusions of the theory
are contradicted by the facts.
16 SCOPE AND METHOD
field of economic theory It is the condiuonal nature of scientific predictions
which causes so much difficulty when wc try to test the theones ‘

The Testing of Scientific Predictions

If we wish to test any theory we confront its predictions with evidence

We
seek to discover if certain events have the consequences predicted by
the
theory Such a task is never lightly accomplished and some of the problems
involved are the subject of Chapter 3 In the meantime we should notice
that as with most other sciences it is never possible to prove or to refute any
theory in economics with 100 per cent certainty
Consider the simple economic theory that predicts ‘if a sales tix is levied
on the product of a competitive industry, then the price of the product will
rise but by less than the amount of the tax’ It is not claimed that this pre-
diction holds only for the years 1945-70, or only m odd-numbered years,
nor IS It claimed to hold only in the USA and Germany but not in France
and Paraguay The prediction says that this result will hold whenever a sales
tax u levied tn an industry that ts competitive We may say that the theory is
unbounded both m time and in space But since we can only make a limited
number of observations wc can never prove conclusively that the theory is
true Even if we have made a thousand observations which agree with the
prediction, it is always possible that in the future we will begin to make
observations which conflict with the theory Since this possibility can never
be ruled out completely (no matter how unlikely we might think it to be),
we can never regard any theory as conclusively proved
It IS also impossible to refute any theory conclusively This matter is con-
sidered in some detail m Chapter 3 and suffice it to say now that, since
human beings make the tests, and since human beings are fallible, it is
always possible that a piece of apparently conflicting evidence arose because
we made a mistake in our observations One conflicting observation does
not worry us very much but as a mass of them accumulates i\e become more
and more worried about our theory and wiU regard it as less and less likely
to be true until eventually we shall abandon it even though wc can
never

be 100 per cent certain we were not making an error in doing so

1 It u important not to treat economic forecasting as bring synonymous with economic

predict the future
prediction Forecasting a type of conditional prediction which attempts to
is
some
by discovenng between economic variables of the sort that the value of F at
relations
future date depends on the value of X today, m which case future Y can be predicted by
the
observing present Many conditional prediciions are not of this form those which reUte
which allow us to
Y today to the value of X today provide significant and useful relations
us to forecast the future
predict ‘if you do this to V you wiH do that to T, without allowing
suggests forecasting
The analogy often drawn between ectmomics and wreather forecasting
rather than the wider class of scientific predictions
 INTRODUCTION 17

When is a theory abandoned? As a generalisation we can say that our

theories tend to be abandoned when they are no longer useful, and that they
cease to be useful when they cannot predict the consequences of actions in
which we are interested better than the next best alternative.* When this
happens the theory is abandoned and replaced by the superior alternative.
We should not be upset by this because we learn new, surprising facts
through the process of upsetting existing theories.
Any developing science will continually be having some of its theories
rejected it will also be cataloguing observations that cannot be fitted into
;

(explained by) any existing theory. These observations indicate the direc-
tion required for the development of new theories or for the extension of
existing ones.^ On the other hand, there will be many implications of exist-
ing theories that have not yet been tested, either because no one has yet
figured out how to test them, or merely because no one has got around to
testing them. These untested hypotheses provide agenda for required new
empirical observations.

The State of Economics

Economics is no exception to this general rule. There are many observations
of the world, for example the distribution of the national product between
wages, profits and the rest, for moment no satisfactory
which there are at the
theoretical explanations. On the other hand, there are many predictions
(for example, that free international trade will make the earnings of labour
lessunequal as between countries) which no one has yet satisfactorily tested.
Thus the serious student of economics must not expect to find a set of
answers to all possible questions as he progresses in his study. He must
expect very often to encounter nothing more than a set of problems which
provide agenda for further theoretical and empirical research. Even when
he does find answers to problems, he should accept these answers as tentative
and ask, even of the most time-honoured theory ‘What observations might
:

I make that would be in conflict with this theory?.’ Economics is still a very
young science and many problems in it are almost untouched. The student
who ventures further in this book may well find himself, only a few years

1 In an advanced science the alternative will be another competing theory. If there is no

competing theory we can try the alternatives of comparing the theory with predictions based
on a naive view such as ‘this year will just be like last year’, ‘any change observed in the past
will go on in the future’, and so on.
2 The development of a new theory to account for existing observations is often the result of

real creative genius of an almost inspired nature. This step in the development of science is the

exact opposite of the popular conception of the scientist as an automatic rule-follower. One
could argue for a long time whether there was more original creative genius embodied in a
first-class symphony or a new theory of astronomy. For a fascinating study of the creative

process the student should read A. Koestler, The Sleep Walkers, especially the section on Kepler.
18 SCOPE AND METHOD

from now, publishing a theory to account for some of the problems men-
tioned herein, or else he may find himself making a set of observations which
will refute some time-honoured theory dcscnbed within these pages
One final word of warning having counselled disrespect for the authonty
of accepted theory, it is necessary to warn against adopting an approach
that IS too cavalier No respect attaches to the person who merely says ‘This
theory is for the birds, it is obviously wrong ’ This is too cheapTo criticise a
theory efiectively on empirical grounds one must demonstrate, by a care-
fully made set of observations, that some aspect of the theory is contradicted
by the facts This is a task which is seldom easily or lightly accomplished

Fig 1 1 provides a summary of the discussion of thconcs and it should be

CHAPTER 2

THE TOOLS OF
THEORETICAL ANALYSIS

If you look at the figure on page 18 you ,

will see two triangles indicating
something must be done in order to move in the direction indicated by the
arrows. In order to accomplish each of these moves the economist needs
to have a set of tools at his command. The first set is composed of all
the apparatus of logical deduction used to discover the implications of
theories. Logical deduction allows the economist to discover the implica-
tions of his assumptions, and thus to deduce from his theories predictions
about observable events.
The set of tools required to accomplish the second move is composed of all
the techniques of statistical analysis that the economist requires when he
comes to test his theories against empirical observations. At that stage, he
seeks to discover how well his theories stand up when confronted by the
facts. The present chapter is devoted to the tools of logical, or theoretical,
analysis, while the tools of statistical analysis are considered in the following
chapter.

THE METHODS OF THEORETICAL ANALYSIS

We have already noted that an economic theoiy' consists of a set of defini-
tions, and one or more hypotheses about the way in which the world
behaves. The economic theorist has the task of discovering what is implied
by these hypotheses. He seeks to make statements such as, costs vary in a

certain way with output and ^businessmen seek to make as much profit as
they can, then a tax on the businessman’s sales will have effects both on the
level of output and on the price at which the product is sold.’ The hypo-
theses of economic theory are hypotheses about the relations between two
or more things (e.g., the relation between the price of a commodity and the
amount of it that people wish to buy). These relations may be described in
20 SCOPr ANP MFTIIOD
%sords, formulated symbolically, or, ifthcrcarc no more ilian three \anabla
invoked, illustrater! graphically by using coordinate geometry Once formu-
lated in a precise way, implications of the hypotheses nny also bedenved
by verbal argument, mathematical analysis or geometry Geometry is, of
coiine, a branch of mathematics Since wish to distinguish geometneal ue
from other mathematical techniques and to avoid the cumbersome expres-
sjon 'mathematical techniques other than geometry', wc shall hereafter
distinguish between 'geometnear and ‘mathcmaticar methods
Winch mctliod is best’ Thisis analogous to asking whether a razor is a

better tool than an axe Hie answer depends upon wliat the task is, and on
the capabilities of the penon using (he too!
Verbal reasoning has the advantage of appealing to common sense at
every step and, if the theory is simple, of being the easiest way of making
deductions It has the disadvantage of being long-winded, even m slightly
complex theories it becomes cumbersome, and, as theories get complex, it

breaks down completely In many cases, it is variually impossible by using

verbal methods to say whether or not 4a given conclusion is implied by the
assumptions of a theory
Geometneal analysis has the advantage of appeal to the eye Most people
find It easier to comprehend a relation between two things when a 'picture'
ofii isdravvn, than wlien u is stated in a mathematical equation The major
drawback of geometry is that it is neccssanly limited to three dimensions
and, for most pnctical purposes, to two dimensions W'e cannot on a simple
graph show a relationship betv\een four things such that each one depends
on the other three Since such rclatiom are common m economics this u a
senous drawback Great ingenuity has been shown by some economic
thconsls in inventing dodges so that the relationship between more than
two things can be shown on a tv>o-<!imcnsional graph Beyond a certain
stage, however, such graphs become extremely cumbersome and difficult to
follow, also there a definite limit beyond v>!uch wc cannot go, after that,
is

more complex relations cannot be dealt with by the tools of simple

coonhnate geometry
At a certain stage of complexity, mathematical analysis becomes neces-
is much
sary and, somewhat before this stage, a mathematical treatment
of mathe-
simpler than a geometneal or verbal treatment TTic advantages
are generality and power, it is possible to handle very complex
maucs
thconcs with relative ease if mathematics is used The
disadvantages are
lacks a certain intuitive
that, for many people, mathematical treatment
verbal and geometneal treatment,
appeal which seems to be obtained from
mathematics
and that for simple problems the high-powered techniques of
appears be a result
t Thcrf II little r\idftife that lh»s»» m the nature of ihtngs rather it to

of our present rduraiional ij-siem

 THE TOOLS OF THEORETICAL ANALYSIS 21

can often be more cumbersome than the lower powered techniques of

geometrical or verbal analysis.
To a great extent these methods are interchangeable; any piece of logical
reasoning that can be done verbally or geometrically can also be done
mathematically, most of what can be done in words can be done in geo-
metry and vice versa. Some things that are done in mathematics, however,
cannot be done rigorously in verbal or in geometrical analysis. Where
various methods can be used, the choice between them will be dictated by
considerations of convenience, economy, and the techniques at the com-
mand of the practitioner and the audience at which he is aiming.
Since a great deal of simple economic theory, both traditional and
modern, is based on the assumpdon of simple relationships between two or
three variables, geometrical techniques can be used extensively. However,
in intermediate and advanced theory the relationships necessarily become
more complex. For this reason most introductor)' textbooks rely almost
exclusively on verbal and geometrical analysis, while an increasing number
of advanced works rely almost exclusively on mathematical analysis, using
only a few geometrical illustrations to give readers a ‘feel' for the more
general analysis. We shall follow the usual practice and rely on verbal and
geometrical argument. In some ways this is to be regretted since the trend
in economics is undoubtedly toward dealing with more complex hypotheses
that can usually be manipulated more easily with mathematical than with
verbal or geometrical tools. In the next section of this chapter, however, the
basic mathematical language that is used to express hypotheses mathe-
matically is introduced. Every student is urged to invest one hour (if it

takes that long) in learning this language. It will pay him large dividends.

EXPRESSING HYPOTHESES: THE CONCEPT

OF A FUNCTIONAL RELATIONSHIP
The idea that one thing depends on another one of the basic notions
is

behind all of science. The gravitational attraction of two bodies depends on

their total mass and on the distance separating them, attraction increasing
with size and diminishing with distance; the number of murders in a
country is thought to depend on, among other things, the severity of the
penalties for murder; the amount of a commodity that people will buy is
observed to depend on, among other things, the price of the commodity;
the higher the price, the less will people buy. When mathematicians wish to
say that one thing depends on another, they say that one thing is vl function
of the other; thus we say that gravitational attraction is a function of the
mass of the two bodies concerned and the distance between them; that
murder is a funcdon of the severity of punishment for it; and that the
 ,

22 SCOPE AND METHOD

quantity of a product demanded

is a function of the pnce of
the product
One of the virtues of mathematical notation is that it allows us to
express
very compactly ideas that require a long-drawn-out verbal expression
There are two steps in giving compact symbolic expression to the relations
we have just described First, we give each concept a symbol, and, second,
we designate a symbol to express the idea of one factor’s dependence on
another Thus, if we let ^ stand for gravitational attraction A/ stand for the
mass of two bodies, and d stand for the distance between two bodies, we
may wnte
g = f(A/. d),

where f is read 'is oV and means 'depends upon’ The whole

a function
equation defines an hypothesis and is read ‘Gravitational attraction is a

function of the mass of the two bodies and the distance between them’ This
IS the same as the verbal statement with which we began

The second hypothesis, that the number of murders depends on the

seventy of punishment for murder, may be expressed as

K= r{S),

where A a measure of the frequency of murders and 5 is a measure of the

is

seventy of punishment for being convicted of murder The functional

expression is read 'The frequency of murders is a function of the seventy
of the punishment for being convicted of murder’ Notice that this is not
necessanly a statement of something known to be true, it is, however, a
positive statement that may or may not be borne out by the facts once we
look at them
Consider the hypothesis

o= r(p),

where D equals the quantity demanded of some commodity, and p equals

the price of the commodity By now the reader should, himself, be able to
state in words what this expression says
The expression
y=^f(X)

says that } is a function of A means that Y depends upon -V Unul we

It

have defined what A is and what F is it says nothing at all about the world
but as soon as we define m
Y and AT terms of measurable magnitudes, we
are making a statement about a relationship
between these magnitudes
The quanuties A and Y in this funcDonal rdauon arc
called variables

Vanations m one of these quantities are associated with variations m the

especi
other quantity The notation often looks fnghtening to the student,
ally to someone who did not get on
weU with his school mathematics How-
 THE TOOLS OF THEORETICAL ANALYSIS 23

ever, once one becomes familiar with it, this notation is extremely helpful,
and since the functional concept is basic to all science, the notation is worth
mastering.
The expression F=f(Z) merely states that Y is related to X; it says
nothing about the form that this relation takes. Does Y increase as X in-
creases ?Does Y decrease as X increases ? Or is the relation more compli-
cated Take a ver)' simple example where Y is the length of a board in feet,
?

and is the length of the same board in yards. Quite clearly, Y=f{X}.

Further, in this case we know the exact form of the function, for length in
feet (7)is merely 3 times the length in yards (A), so we may write 7=3A.

This equation specifies the exact form of the functional relation between 7
and X and provides a rule whereby, if we have the value of one, we can
calculate the value of the other. If, for example, we should want to know
the value of 7 when X is 6, we replace X with 6 and the equation tells us
to multiply 6 by 3. We obtain 18, which is the value of 7. Clearly, this
operation could be repeated for any conceivable value of X and the corre-
sponding value of 7 calculated.
This example is not typical of all functional relationships because it is
true by definition. It is not an hypothesis because it could not be refuted.
It merely states in functional form the relation between the definitions of a
foot and a yard. It is nonetheless useful to have a way of writing down
relationships that are definitionally true.
Now consider a second example. Let C equal the total spending of a
nation on all consumption goods in one year, and 7 equal the total income
of all persons in the nation in the same year. Now state the hypothesis

C= f(7), (1)

or, more specifically,

C= -757. (2)

Equation (1) says that we hypothesize that national consumption depends

upon national income. Equation (2) says, more specifically, that expendi-
ture on consumption will be three-quarters {-75 times) as large as national
income. The more specific equation, equation (2), expresses an hypothesis
about the relation between two observable magnitudes. There is no reason
why equation (2) must be true ; indeed, might not be consistent with the
it

facts.But those are matters for testing. What we do have in the equation is
a concise statement of a particular hypothesis.
Thus the general fact that there is a relation between 7 and is denoted X
by 7=f(Z), whereas any precise relation may be expressed by a particular
equation such as 7=2A'^, 7=4A^, or Y=X+2X^ + -5X^.
If 7 increases as X increases (e.g., 7=10-|-2A), we say that 7 is an
 24 SCOPF AND MFTMOD
iNrRFAsiNC FUNCTION of A OF that I and A varn dirfctuy with each
Ollier ir r decreases as A increases (e g , l’= 10-2A ), ssc say ihat I' is a
DEOREASINO FUNCTION ofA Or ihit 1' and A’ VARA INVERSELY WITH each
Other
Fconomic theory u bnsed on rclatioiif bchsccn \inous magnitudes
(e g ,
tlieqinntity demanded of some commodity is rehted to
the pnee of that
commotlity the amount spent on consumption is related to national
,
in-
come) All such rchtions can lie expressed m the form of mathematical
equations It is this fici that gises mathematical analysis importance
in
economics since once our hy poiheses are untlcn down in terms of algebraic
expressions we can use mathematical manipulation to discover what impli-
cations they have about behaviour

THE ERROR TERM

The examples of functional relations considered abov c were all deltmxnisUc
ones in the sense tint tliey were expressed as if they held exactly given the

value of A we knew the value of I exactly The relations considered in

economic theory are seldom of this deterministic son When an economist ‘

says that the vvorld bthavisio that 1 =f(A),hedoes not expect that knowing
A will tell him tx<ictly wlni 1 will be, but only that it vvill tell him what 1
w ill be u Uh some marf^n of error Tins error sg predicting I’ from a knowledge
of \ arises for two quite distinct reasons First there may be other vanables
that also affect for example, we say that the demand for butter
}’ \Mien,
is a function of the pnee of butler, we know tliat other factors
will also demand A change m the pnee of marganne will
mflucncc this

certainly affect the demand for butler, even though the pnee of butter docs
not change Thus we do not expect to find a perfect relaUon between
and Pt that will allow us to predict £)* exactly, from a knowledge of
Second, we can never measure our vanables exactly, so that, even if X is
the only cause of ), our measurements will give vanous l”s corresponding
to the same A In the ease of the demand for butter, our errors of measure-

I Ofeounr an » drfimuoo will hold txactly If for example we

ftjuaijon ihai pxprrsjf*

break up a person infome ()T into tpendinf; (C) and lasinj (S) and define savings as all
t

income not spent we write r*.C+5 This equation which is called a definitional
then
always hold there is
equation is true exactly Ue have defined our tenia so lhat it must
nothing anyone can do to invalidate the equation On the other hand if we believe that
people

three-quarten of their income and sase the other quarter we wnte C=JI,
alwa)s spend
J a J > These two cquaiions are called behavsouni equations because they tell us what we are
exactly indeed they need
assuming about people i behaviour Sodi equations need not hold
tenthofhismcome
notholdaiall Uemight forexample observesomeoneipendmgonlyone
would hold although of
and saving nine tenths in which case neither of the latter equations
coune r=C+5 u still true (by defintCion)
 THE TOOLS OF THEORETICAL ANALYSIS 25

ment might not be large. In other cases, errors might be substantial as, for
example, in the case of a relation between total spending on consumption
goods, C, and total income, Y, earned in the nation; C=f(7). In this case
our measurements of C and Y may be subject to quite wide margins of error,
and we may observe various values of C associated with the same measured
value of 7, not because C is varying independently of 7, but because our
error of measurement is varying from period to period.
If all' the factors that affect the measured value of 7 other than X are
summarised into an error term, e, we write 7=f(7, e). This says that the
observed value of 7 is related to the observed value of X as well as to a lot
of other things, both observational errors and other causal factors, all of
which will be lumped together and called e (the Greek letter epsilon). In
economic theorjq this error term is almost always suppressed, and we pro-
ceed as if our functional relations were deterministic. WTien we come to test
our theories, however, some very' serious problems arise precisely because
we do not expect the functional relations of our theories to hold exactly.

It is extremely important, both when one comes to

interpret a theory in terms of the real world and to
test a theory formally against empirical observations,
to remember that the deterministic formulation is a
simplification, and that the error term is really
present in all our assumed and observed functional
relations.

CAN ECONOMICS REALLY BE EXPRESSED

IN MATHEMATICAL TERMS?

One sometimes hears the argument that although it is all right to use
mathematical equations to express the behaviour of gases, or planets, it is
impossible to do so with human beings. Many reasons have been advanced
in support of this view. We shall consider Tiere two of the most frequently
encountered ones.
The proposition sometimes argued on the grounds that, because human
is

behaviour is subject to ‘free will’ it is too complex to be made the subject of

scientific laws and that mathematics is too rigid to be used to describe any-
thing but deterministic scientific laws. The first point is dealt with in

Chapter 1, wrong, both because terms expressing

and the second point is

random variations are easily accommodated in mathematical analysis, and

because most predictions in natural science also contain error terms.
The second argument often advanced in favour of this proposition is that
human behaviour is affected by far too many variables to make it suitable
26 SCOPE AND METHOD

for expression m
mathematical tcnns Such a cnticism cannot mean
literally that mathematical formulabon cannot handle enough vanables,
for
indeed it can handle any number ‘ What it may mean js that mathematical
notation will not by itself speciiy hypotheses about the way in which the
large number of possibly relevant vanables affect whatever it is we are
trying to predict This is, of course, true mathematical notation cannot
specify anything that the person who uses the notation docs not specify
There may, indeed be many aspects of hts problem that he does not under-
stand or about which he does not have enough hunches to permit him to
formulate hypotheses but this is his limitation, not the limitation of his
technique
^ K^
The implication m this kind of cnticism is often that verbal statement can
somehow overcome ignorance of, or vagueness about, what the person is
considenng Of course it cannot Verbal statement can often mask fuzziness,
but that is hardly an advantage It is an adoantage, not a disability, of mathe-
matical formulation that it exposes precisely what is being said as well as
what IS being left unsaid If we accept the view that somehow, verbal
analysis (or ‘judgment’) can solve problems while we arc unable to state
clearly the considerations that lead to these solutions, then economics is no
longer a science but has been turned instead into a medieval mystery, in
which the mam problem is to be able to distinguish between the true and
the false prophet 2 i
^ “7

THE GRAPHICAL REPRESENTATION OF FUNCTIONAL

RELATIONSHIPS^
Consider the following simple functional relationship

Y = 150- tOX
Fig 2.1 Alternative representations of the function Y — 150 lOA^.
 28 SCOPE AND METHOD
Once we have plotted this line, which ts the function Y= 150- lOZm the
interval from X=0 to A’=20, wc have no further
need for the coordinate
gnd, and the figure will be less cluMcrcdjfwe suppress in l{iu)
For some purposes wc do not really care about the speciHc
numerical values
of the function, we are content merely to represent it as a downward-sloping
straight line We have so represented it in Figure 2 1 (iv) We have replaced
the specific numerical values of the variables Y and X with the letters a b t
and s to indicate specific points Figure 2 l(iv) tells us, for example, th’at’if
we increase the value of A from Oa to Ob, we expect Y to decrease from Os
to Or
The student may feel he has lost ground at this stage, but it is in this form
that most diagrams appear in economic texts To see why this is so we shall
consider an example from economics Say we believe that there is a relaUon-
ship between the amount of labour a firm will hire and the wage rate that
must be paid to labour, thus or, since we expect that other things
will also affect the demand for labour, D,^=[{w, e) Assume that there are
twenty firms of the same size, each producing the same commodity and
roughly comparable in all other ways, except that they arc located in
dificrent parts of the countryAbo assume that wage rates vary throughout
the country We could now measure the relation between wages and demand
for labourby recording the wage that each firm has to pay and finding the
quantity of labour that it employs ' Assuming that we have made these
measurements we then express the data on the scatter diagram shown as
Figure 2 2 The relation between and w now leaps to the eye It is clear
that knowing the wage rate gives us some idea, but not a perfect one, of the
quantity of labour that wll be employed by and large, the quanuty of
labour employed seems to fall as the wage rate nses The line on the
diagram gives us an idea of the general relation between labour employed
and the wage rate
The relation illustrated in Figure 2 2 can be expressed m terms of an
equation in just the same way that earlier we expressed in equations the
relations between distance and distance in yards, and between con-
in feet
sumption expenditure and household income The equation in this case is
I>= 1,000- 12« (where w is expressed m
pennies) This equation shows,

just as does the line in Figure 2 2, that 1,000 workers will be

employed when
the wage rate is zero and that, for every rise of 1 rf m the wage rate, 2
1 fewer
distance of
workers will be hired Finally, if we let « stand for the vertical
described by the
any point from the line, then each observauon is exactly
equation 1,000— 12u) +f
now go on to buiW a theory ^bout these firms, we might
If we were to

labour
unable to affect the vrtge rate by hinng more or
less
1 We assume that each Hrm is
 THE TOOLS OF THEORETICAL ANALYSIS 29

suppress the actual observations and use the stylised relation, Z)= 1,000
—
I2w, to describe the relation between demand for labour and the wage
rate. We
would then have the relation illustrated in Figure 2.3.

Fig 2.2 The relation between the

quantity of labour demanded
and the wage rate.

0 250 500 750 1000

Number of labourers employed

Fig 2.3 The relation between the

quantity of labour demanded
and the wage rate.

0 250 500 750 1.000

Number of labourers employed

-
are illustrating a functional relation by graphs, we have the
When we
great advantage that we can easily compare different sorts of relation with-
out specifying them in precise equations. To compare relations in general
terms we suppress the numbers of the example shotvo in Figure 2.3 and use
letters instead. We are then able to compare two different functional rela-
tions in quite general terms. If in one industry the quantity of labour used
does not change very much as thewage rate changes, the graph of the
 30 SCOPE AND METHOD
reJatJon ^i[w) will be very steep, curve . a fall m the wage rate from
Os to Or in Figure 2 3 causes
demand to increase from Oa to 06, that is, by
the amount ab
If in a second industry, the quantity of
labour demanded
varies a great deal as the wage rate vanes,
then the graph of i) 2 =f(a») will
be flat, curve Z)l 2 so that the
> fall m the wage rate from Os to Or will cause
demand to go up from Oc
Od^ that a, by an amount td The student
to
should note that when we move from Figure 2 3
to Figure 2 4, not only do
we suppress the numbers, we no longer draw the whole coordinate gnd m

Ftg 2 4 The relation between the

quantity of labour demanded
and the wage rate

This IS merely a consemence to prevent the graph from getting cluttered up

with irrelevant detail The grid is, of course, always understood to be there

and, when required, we draw m the necessary grid lines for instance the
coordinates of the point x m Figure 2 4 are Os and Oe, and the gnd lines xs
and xc arc drawn in because they are needed A student who finds this at all
difficult should redraw all graphs on graph paper until such time as he feels
quite familiar svith graphical analysis

DERIVING IMPLICATIONS FROM FUNCTIONAL

RELATIONSHIPS
So far, we have discussed the various ways in which functional relations can
be described, verbally, geomcincally and mathematically When the econo-
mist has laid out the functional relations of his theory, he must then discover
what IS implied by these relations He wishes to make statements such as ‘If

relations a and b hold, then relation c must necessanly hold as well’ In the
process of making logical deductions from his theones, he may again
employ
verbal, geometncal, or mathemaOcal forms of reasoning
His mam concerns
so that he
will be (1) to ensure that his reasorang processes are correct
correctly discovers what is implied by his theory, and (2) that they are
is implied by his theory
efficient so that he discovers everything that
 THE TOOLS OF THEORETICAL ANALYSIS 31

Examples of economics cannot be given until we have a

this process in
rigorously specified theory of some aspect of the economy. You will en-
counter many examples in Chapter 1 1, at which time you will be referred
back to the present chapter. In the meantime, let us consider the following
problem in elementary algebra which is similar in structure, if not in content,
to many economic problems.
A two and a quarter times his son’s age. The sum of the
father’s age is

father’sand the son’s ages is 52. How old are father and son?
We have two conditions and we have to discover what is implied by them
with respect to the ages of father and son. You should not read on until you
have provided a verbal, a geometrical and an algebraic solution to this
simple problem in logical deduction.

THE QUANTITATIVE RELATION BETWEEN VARIABLES

This idea of the magnitude of the changes in one variable in response to
changes in another variable, is extremely important in economics. We
expect the demand for anything to vary with its own price, D=f{p), and we
are interested in how much the demand changes for a given change in price.
We amount of a commodity produced (5) to vary with its own
expect the
price, S=f{p), and again it is important to know by how much S will change
for a given change in price. We expect the volume of unemployment [U)
to vary with the difference between government revenue (^) and expendi-
ture {E), U—f{R--E), and we want to know by how much unemployment
will change for a given change in the budget deficit or surplus.
There is a precise mathematical method of handling problems arising
from the question of how one variable changes as another variable on which
it depends also changes. The branch of mathematics which deals with these

problems is the differential calculus. There is not space here for a consideration
of the methods of the differential calculus. Since many of the basic tools of
economic analysis are merely applications of derivatives, some study of this
concept will bear dividends as the student proceeds in his study of economic
theory. A
knowledge of the calculus is not necessary in order to read this
book. In fact one can usually obtain a first degree in economics without such
knowledge, but those who do have some idea of the calculus will undoubtedly
find it a great help.^
1 An introduction to the ideas of the differential calculus plus a review of very elementary
arithmetic, algebra and geometry can be found in W. W. Sawyer’s excellent little book
Mathematicians Delight (Penguin) ;
a somewhat more advanced treatment may be found in
J.Parry Lewis, An Introduction to Mathematics for Students of Economics (Macmillan, 1959). The

student who wishes to make a serious study of mathematics on his own is probably best advised
to tackle R.G.D. Allen, Mathematical Analysis for Economists (Macmillan, 1953) which is the
most rigorous and the most difficult of the three books mentioned.
 34 SCOPE A^ D METHOD
On the other hand we might iwsh to who are MPs can be admitted) nor
deal simultaneously with tuo different sufficient (since female graduates of
relations A and I’, for example,
between Oxford are not admitted)
•<^=3} andA^=2 — 6y In this case we
We may
summanse the conditions for admission
could wnte as follows

A = A,(y) To be male is necessary but not

and sufficient
^= ^V2(y). To be a male MP u sufficient but not
necessary
where the numerical subscripts dis-
tinguish the two different relations
To be both a male and an Oxford
graduate is sufficient but not neccs
This choice of letters to indicate the
sary
particular functional relation in ques-
To be an Oxford graduate is neither
tion be illustrated many times
will
necessary nor sufficient
throughout this book
To be an MP is neither necessary nor
The fact that vanous notations arc
sufficient
used to indicate fanctioml dependence
To be eilAer a male graduate of
can be confusing, but the intenuon of
Oxford or a male MP is necessary
the author will usually be obvious from
and sufficient
the context
In general, a necessary condition is
something that must be present but by
2 Necessary Conditions and
Itself may not guarantee the result A
sufficient condition is something that if
Sufficient Conditions
present, does guarantee the result but
It is common in popular discussion to that need not be there for the result to
confuse necessary and sufficient con- occur Acondition (orsetofconditions)
ditions Many futile arguments have that IS necessary and sufficient must be
been caused by one person arguing that there and, if there, is enough to
a condition was sufficient for a result guarantee the result
and another arguing that it was not In this club, the necessary and
necessary, each thinking he was contra- sufficient condition for entry is a com-
dicting the other when, in fact, both pound either or condition to be cither
were correct Consider, for example, a a male graduate of Oxford or a male
club that normally admits only males MP If, however, another club were set

who are graduates of Oxford, but that
IS also willing to admit all male MP’s,
whatever their background Bang a
male MP is thus sufficient to admit you
to the club, but it is not necessary to be
one Being a male is a necessary con-
dition for admission (since no females
are admitted on any terms), but it is
not a sufficient condiuon Bang a
graduate of Oxford is b) itself nather
necessary (since non-Oxford graduates
up that was open to all former members
of the House of Commons and to no
one ebe, then to have been an MP
would be a necessary and sufficient
condition for entry into the club
3 Dependent and Independent
Variables
Suppose we say that Y is alway s 3 times
as large as A' Two otherwaysof saying
 SOME COMMON TECHNIQUES 35
the same thing are to say that X is one- of rainfall This allows us to write
third as large as Y and that Y minus
ZX must be zero. We can write C= C{R). (4)

But if knowing the amount of rainfall is

Y= 3X, sufficient to allow us to
deduce the
amount of crop, then knowing the
and amount of crop is sufficient to let us
Y-3X= 0. deduce the amount of rainfall. Thus we
also have
These are three ways of writing the
same functional relation. To express
R= R{C). (5)

the same three forms in general terms, As far as mathematics is concerned, it

we can write does not matter which of these two
ways we choose. But, of course, the
Y=g{X), (1) causal relation is clearly defined in this
X=h(Y), (2) case. The amount of rainfall influences
the crop yield the crop yield does not
;

11 d (3) influence the amount of rainfall.

Equations (1) and (2) are called the
explicit forms of the function. In (1)
is written as an explicit function of X.
Equation (3) is called the implicit form
of the function. All the terms are
gathered onto the left-hand side and
the ^vhole expression is thus equal to
zero. In which of the tliree forms we
choose to write the function is clearly
only a matter of convenience. Any one
can be transformed into the others
merely by transferring terms from one
As a matter of convention, whenever
we think we know the direction of the
Y causal fink between variables we write
the causes as independent variables and
the effects as dependent ones. Thus, as
a matter of convention, we would use
equation (4) instead of equation (5).
Again, if we wanted to say crop yield
(C) depended on fertiliser (F), sun-
shine (.9), and rainfall [R) we would
uTite
C^C{F,S,R). (6)

side of the equation to the other. In equation C is the dependent

(6),
The term on the left-hand side of (1) variable and F, S and R are the inde-
and (2) is called the dependent variable pendent variables.
and the terms on the right-hand side
are called the independent variables (or
variable, if there is only one) As far as
. 4 Exogenous and Endogenous
mathematics is concerned the distinc- Variables
tion between dependent and inde-
In economic theories it is convenient to
pendent variables is quite arbitrary:
distinguish between exogenous and
Y=f{X) necessarily implies A’=g(F). ENDOGENOUS VARIABLES! Endogenous
The convention may be used, however,
to express information we have about
the causal relation between the vari- 1 This is a simplification for purposes of illus-
ables. Assume, for example, that crop tration, but crop yield certainly depends partly
yield (C) depends solely on the amount on the amount of rainfall.
36 SCOPE AND METHOD
vanables are ones that are explained flow, there is so much flow per year or
within a theory, exogenous variables are per month The amount of wheat sold
ones that influence the vanables but IS also a flow - so much per month
or
are themselves determined by lactors year The amount of wheat stored
outside of the theory Assume, for ex- (produced but unsold) in the granancs
ample, that we have a theory of what of the world is a stock, it is just so many
determines the pnce of apples from day millions of tons of wheat The distinc-
to day in London The pnce of apples tion between stocks and flows will arise
in this case is an endogenous variable - many times throughout this book
something determined within the
framework of the theory The state of
the weather, on the other hand, is an 6 Identities and Equations
exogenous variable It will influence The distinction between identities and
apple pnces but will be uninfluenced equations is important and subtle An
by these pnces The state of the weather identity a relation that is true for all
is
will be unexplained by our theory, it is values of the variables, no values can
something that happens from without, be found that would contradict it An
so to speak, but it nonetheless influ example of an identity is

ences our endogenous variable, apple

pnces, because it affects the demand (x+y)* s x^ + 2xy+y^,

for apples Exogenous variables are which expression is true for any numen
sometimes referred to as autonomous and y It should be noted
cal value of x
VARIABLES that identities are usually wntten with
a three bar sign and that the expression
ysx is read y is identical with x
5 Stocks and Flows Identities are statements compatible
Some of the most senous confusions m with any state of the universe
economics have anscn from a failure to Equations arc relations that are true
distinguish between stocks and flows only for some values of the vanables
Imagine a bathtub half full of water but that can be contradicted by other
with the tap turned on and the plug values Thus the expression y=I0+2x
removed, you have in mind a model is an equation wntten with a two-
It is

similar to many simple economic thco-
nes. The level of water in the bath is a
stock -an amount that IS just there VVc
could express it as so many gallons of
water The amount of water entering
through the tapand theamountlcaving
through the dram are both flows Wc
could express them as so many gallons
bar or equals sign and is read y is equal
to ten plus two x This expression is
true, for example, for x=2 and y= 14,
but not for x=2 and y=2 Equations
can be used to stale testable hypotheses,
since they make statements that are
true for some states of the universe but
false for other states , idcntiUcs cannot

per minute or per hour A flow nccessanly be used to state testable hypotheses,
make statements that are
has a time dimension — there is so much since they
of the universe
flow per penod of time A stock docs not true for all states

Identities, therefore, tell us nothing

have a time dimension - it is just so
many tons or gallons or heads
The amount of wheat produced is a
about the world They cannot be the
‘basis’ of any theory (although they
can
 SOME COMMON TECHNIQ^UES 37
be used ver)' helpfully to convey defini- perplexing habits of economists is to
tions of terms) and they can usually be warn the student about the nature of
reduced to the form y=y which, al- identities and then to introduce
though true, is hardly verj' enlighten- national-income theor>' with several
ing. Consider, for example, the pages of identities claimed to be the
statement foundation of the theory.*

y = c-hs, (7)
7 Some Conventions in
where y a man’s income, where c is
is
Functional Notation
his expenditure on goods and services,
and where r is the amount he saves. As Assume we are talking about some
sequence of numbers, say, 1, 2, 3, 4, 5,
it stands we do not yet know if this is an
. .If we wished to talk about one
.

identity or an equation. If y is defined

particular term in this series rvithout
as the amount of money the man
indicating which one, we could talk
earns, c as the amount he actually
about the ith term, which might be the
spends on goods and services, and s as
5th or the 50th. If we now want to in-
the amount he puts in the bank, then
dicate terms adjacent to the ith term,
the equation expresses an hypothesis
whatever it might be, we can talk about
about what people do with their in-
the (i— l)th and the (i+ l)th terms.
comes. This is because there are other
By the same token we can talk about
things the man might do' with his in-
a series of time periods, say, the years
come, such as giving it to his nephew.
1900, 1901 and 1902. If we wish to refer
Now, however, let us keep the above
to three adjacent years in any series
definitions of y and c, but define s in
without indicating which three years,
terms of y and r; r is defined to be all
we can talk about the (/— l)th, the tth
income not spent on consumption.
and the (/+ l)th years in the series.
Thus by definition we have
Now consider some functional rela-
5 = y-c. (8) tion, say, one between the quantity
produced by a factory and the number
Equation becomes an identity, of workers employed. In general, we
(1)
can write Q = where Q is the
y = c+s. (9) amount of production and IV is the
number of workers. If we wished to re-
We fool ourselves if we think we have fer to the value of output where ten
learned anything from (9), for, if we workers were employed, we could then
substitute (8) into it, we get write Qio = Q(^'^io), whereas, if we
= c+y-c, wished to refer to output when some
y
particular, but unspecified, number was
which, of course, reduces to employed, we could write Q/=Q(1T;).
Finally, if we wished to refer to output
J =A
I A criticism of this practice, and references to
which is and s.
true for any values of y, c
places where it is used, is given in K.KIapphoIz
Confusion between equations and and E.J.Mishan, ‘Identities in Economic
identities hasbeen a ready source of Models’, Economka, May 1962. This article is not

error in economics. One of the most suitable, however, for beginning students.
38 SCOPE AND METHOD
when the number of workers is in- can say that F is a function of n vari-
creased by one above the previous level, ables Aj to X„ Now the omission of
wecanwnteQi+i = Q(n'^i) Thisuse intermediate numbers is necessary, for,
of subsenpts to refer to the value of the until we know what number n stands
variables where they take on particular for,we cannot say how many variables
numbers is a most useful notation and there arc In this case we write
one that we shall use at vanous points
in this book y=r(x„ ,.v,)
We may also use time subsenpts to
indicate a lagged relation between \an-
ables A
lagged relation between and X 8 Straight Lines Slopes and
Y IS one in which the value of I' at any Tangents
point of time depends on the value of
X at some previous point of time Let us Consider the following functional rela-
say that the amount produced of a tions
product IS a function of its pnee, so we
write Q = Q{p)i where Q is the amount y= 5A’,
produced and p is the price of the pro- r» X,
duct Production lakes time, and what F= 2.V
IS produced today may not be much

influenced by today’s pnees If we These are graphed in Figure 2 5 You

divide time into montlis and assume a
three months’ lag in output, then we
have Qi=Q{p, 3 ), which says that the
amount produced today depends on
what the price was three time jwnods
ago, which in this case
ago
Another convention used to save
will see that they are all straight lines
through the ongm This is also obvious
from the fact that if we let A=sO in each
of the above relations Y also becomes 0
In the first equation, Y goes up half a
is three months unit every time X goes up by one unit,
in the second cquauon, Y goes up one
is unit every time X goes up one unit and ,

space when there are many inde- in the third equation, Y goes up two
pendent variables in a function Let us units every time X goes up one unit
say that Y depends on six variables A| We now introduce the symbol J to
to Xf, We could wnte this as indicate a change in a variable Thus
means the value of the change in A
Y = YiX„ X2 X3 ^4
, , , As, Xe), and A Y means the value of the change
in F In the first equation if X= 10 then
but this is rather cumbersome, so,
F IS 5 and if X goes up to 16, F goes up
instead, we write to 8 Thus, in this exercise, AX=6 and

Y = r(Ai, Xe),
JF=3
,
Next consider the ratio AY/AX
In

where the dots indicate that the inter- the above example it is equal to 5 In
be general, it will be noted that, for any
vening numbers are understood to

present change we make in X

in the first

a function of
P equation, AYjAX is always 5 In the
Now assume that is

some number of vanables but we do second it is unity and m

the third the

We always 2 In general, if we wnte

not wish to say exactly how many ratio IS
 SOME COMMON TECHNIQ^UES 39

This slope tells us the ratio of a change

in 7 to
a change in X.
In trigonometry the tangent of an
angle is defined as AYjAX; thus the
slope of the line is equal to the tangent
of the angle between the line and any
line parallel to the X axis. In general,
the larger the ratio AYjAX the steeper
the graph of the relation. In Figure 2.5
three lines corresponding to A YjAX—^,
1 and 2 are shown. Clearly, the steeper
the line the larger the change in 7 for
any given change in X.

0 1 234567y Now
tions:
consider the following equa-

Fig 2.5
7 = 2X
Y=bX, then the ratio A YjAX is always 7 = 10 + 2Z
equal to 6.^
7 = -5 + 2X.
We now define the slope of a straight
line be the ratio of the distance
to These are graphed in Figure 2.6. It will

moved along the Y axis to the distance be observed that all three lines are
moved up the X
axis. We start at the 7
point [Xi, 7i) and then move to the
50
point (^2 , Yj). The change in is X
X 2 — Xi or as AX
indicated. The change
in r is 72 ~ Ti or AY. Thus the ratio
40
AYjAX is the slope of the straight line.

1 This is easily proved as follows: Select some

30
arbitrary value of X, called A",, and calculate the
corresponding Tj ;
now take a second value of X,
called X2 ,
and find the corresponding Y2 This
.

gives us 20
y. = bx,
and
¥2 = bX2.
10
If we subtract one equation from the other we
obtain
7^- 7,= bX2-bXi,
Y2-Y,=b[X2-X,), 0

but we have already used the term 4l 7 to describe

the change in 7 and the term /3X to describe the
change in X 4f7=72 — 7, and
(i.e., JX= -10

X2 — Xi). Thus we may write Fig 2.6

JY = bJX, parallel, they have the same .slope.

i.e.,

In all three AYjAX is equal to 2.

Clearly, the addition of a (positive or
 40 SCOPE AND METHOD
negative) constant does not affect the graphically by a curved line and alge-
slope of the line This slope is in-
braically by some expression more
fluenced only by the number attached complex than the one for a straight line
to In general, we may %vnte the Two common examples are as follows
equation of a straight line as
Y=a + bX+cX^
Y= a + bX and
Now, by inserting two values of Jl; say
and ^” and finding
A’i 2, the correspond
ing F s we get
The first example is a parabola that can
Fi = a+A.Y, take up vanous positions and shapes
and depending on the signs and magnitudes
Yi = a + bXi of a, b and c Two examples of para-
and, by subtraction, bolas are given in Figures 2 7 and 2 8

72-Fi = i(^2-.Y,)
or
jy = b^X

The constant a disappears when we

subtract and so does not influence the
slope of the line What the constant
Fii2 7
does IS to shift the line upward or
downward parallel to itself

9 Nonlinear Functions

All of the examples used so far in this

appendix and most of the examples in
the text of Chapter 2 concern linear
relations between two variables A
linear relation is described graphically
by a straight line, and algebraically by
{^<sqii3tK?n It .'S character
istic of a linear relation that the cficct
on y of a given change in X is the same
whatever the values of X and Y from
which we start The graphical ex- The second example becomes a rcct
pression of this is that the slope of a angular hyperbola if we let A=l, and
straight line is constant then the position is determined by the
Many of the relations encountered in value of a Three examples where a = 5,
economics are nonlinear ones In this 2 5 and 5 are shown in Figure 2 9
case the relation will be expressed There are, of course, many other
 ,

SOME COMMON TECHNIQUES 41

Now consider the function

7 = 75-10A-t-•5A^
which is graphed in Figure 2.8. In this
case, the value of 7 falls at first while
X increases, reaches a minimum, and
|
then rises as X goes on increasing. In
this case, 7 reaches a minimum value
of 25 when X is 10. Here we speak of
minimising the value of the function, by
which we mean finding the value of X
for which the value of 7 is at a mini-
mum.

1 1 Functions of More Than Two

examples of nonlinear relations be-
tween variables. In general, whatever
the relation between X and Y, as long
as it can be expressed on a graph it can
also be expressed by means of an
algebraic equation.
Variables
In most of the examples used so far we
have been considering the relation be-
tween only two magnitudes. In most
interesting cases we are concerned with
the relation between more than two
things. The demand for a good might

10 Maximum and Minimum Values depend, for example, on the price of

that good, on the price of a number
Consider the function of competing products, on the price of
products used in conjunction with the
Y = 10X—\X\ product with which we are concerned,
which is plotted in Figure 2.7. It will and on consumers’ incomes. For ex-
be observed that Y at first increases as ample, the demand for butter will
X increases, but after a while Y begins depend on the price of butter, the price
to fall as X goes on rising. We say that of margarine, the price of bread, and
Y rises to a maximum, which is reached how rich or poor consumers happen to
in this case when X=50. Until X=50, be.
Y is rising as X rises, but after X=50, When we wish to denote the depend-
Y is falling as X rises. Thus Y reaches a ence of 7 on several variables, say,
maximum value of 250 when X is 50. V, W
and X, we write 7= 7( F, W, X)
A
great deal of economic theory is which is read 7 is a function of F, IT
based on the idea of finding a maxi- and X.
mum (or a minimum) value. Since Y In mathematics and in economics
is a function of X, we speak of maxi- we are often concerned about what
mising the value of the function, and by this happens to 7 as Avaries on the assump-
we mean that we wish to find the value tion that the other factors that influence
of X (50 in this case) for which the value X are held constant at some stated
of Tis at a maximum (250 in this case). level. There are many ways to denote

2‘
42 SCOPE AND METHOD
this and wc shall use the following frequent use in economics of arguments
notation that depend on the qualification ‘other
thin^ being equal’ (for which
Y= Y{X] I

use the Latin phrase

often
ceteris paribus)
Such arguments are not peculiar to
I IF = IK®
economics They are used successfully
The symbols to the nght of the bar tell in all branches of science and there is

us what being held constant and at

IS an elaborate set of mathematical tech-

what levelThe above example is read niques available to handle them

X
F IS a function solely of with V held When mathematicians wish to know
constant at the level K® and IK held the approximate ratio J } /d A (i e , how
constant at the level IK® In a particular X
Y IS changing as changes) when other
example, we might state the actual factors that influence Y arc held con-
levels of V and IK If, for example, IK stant, they calculate what is called the

were held constant at 10 and K at 5 we partial derivative of Y uith respect to A

would write This IS wntten symbolically as dilBX
We cannot enter here into a discussion
Y= Y(X) I of how this expression is calculated Wc
Ks= 5 only wish to note that finding dYjdX
I
IK = 10 IS a well-recognised and very common

mathematical operation, and the

Partial Derivatives answer tells us approximately how Y is
Students who do not know mathe* aflected by small variations in A when
matics are often disturbed by the all other relevant factors are held constant
CHAPTER 3

THE TOOLS OF
STATISTICAL ANALYSIS

If you look back once again to Figure 1.1 on page 18 you \vill see that the
second of the two processes contained by triangular symbols is the process of
statistical analysis used to test the predictions of theories. Actually the role
of statistical analysis is twofold. First, we \\’ish to use observations from the
real world to test our theories. Second, we nnsh to use such observations to
give us measures of the quantitative relations between economic variables.
Testing and measurement are the purposes for which stadstical tools are
needed. In this chapter we consider them both.
Economic theories are meant to predict the outcome of various changes
in which we are interested they are meant to tell us that if certain things
;

are done then certain things will happen and certain other things will not
happen. When we intervene in the economy to bring about a particular
result, we are relying on some economic theory that links our act of inter-
vention to the result that we desire (i.e., the theory predicts the consequences
of our intervention). But it is always possible that some theory about how
events in the economy are related may be wrong; if tve act on the basis of
an incorrect theor>', we will not obtain the desired results - indeed, we may
bring about results opposite to the ones that -^ve desire. It is critically
important to know whether our economic theories do or do not predict the
outcome of various changes in the economy within some acceptable margin
of error.

TESTING
In order to determine whether or not they do give us acceptable predictions
in the sense described above we must test our theories against the evidence
of what actually happens in the economy. Testing theories against observa-
tions is not a task that is lightly accomplished (or briefly described). In
44 SCOPE AND METHOD
parucular most important to realise that \nc cannot accept a
it is
single
connicting obser^aUon as disprowng anv theory
There are ^^^o major
reasons for this First uc can, and often do, make erron
of obscn-alion
Thus, if we make a single observation that seems to refute our
theory, inc
do not reject the theory until we have niled out the possibility that we
have
made a mistake in what we think we saw ‘ Second, we can nc\ er be certain
that our theory includes all the forces at work Indeed, m economics we are
usually certain that we have left out of our theory some things of potential
importance What we wish to know is whether our theory is able to predict
events in the world within some reasonable margin of error This is some-
thing that cannot be settled by making a single observation It can onl> be
settled by considenng the weight of a mass of evidence
As a first step in seeing how we go about testing theories in economics, w c
must distinguish between laboratory and nonlaboratory methods

Laboratory Sciences
In some sciences, it is possible to obtain all neccssaiy observations from
controlled expenments made under laboratory condiuons In such experi-
ments, we hold constant all the facton that are thought toaffect the outcome
of the process being studied Then we vary these factors one by one while
we observe the influence that each variation appears to have on the out*
come of the experiment
Suppose, for example, we have a theory that predicts that the rate at
which a substance burns is a function of the chemical properues of that
substance and the rate at which oxygen is made available dunng the process

of combustion To test this theory, we can (1) take a number of identical

pieces of some substance and bum
them, varying the amount of oxygen
made available m
each case (Tins allows us to sec how combustion vanes
with the quantity of oxygen used ) We can then (2) take a number of sub-
stances with different chemical composiuons and bum Uiem, making avail-
able identical amounts of oxygen in each case (This allows us to sec how
combustion vanes with chemical composition )
generated
In such an expenment, we never have to use data that arc
when both chemical composition and the quantity of oxygen are varying
con-
simultaneously Laboratory conditions arc used to hold other things
vanetl one at
stant and to produce data for situations in which facton are
a time
theory of phyiia that u not refuted dai^
1 It has been *aid that there u hardJy an accepted
the country Such uoUied
by »ome ichoolboy operating tn a lehtxrf bboraiory aomevrhere in
refutatiom’ do not worry phyiicuu although they would
be womed ifiome day al^t oery
to refute «*me
.choolbo) in the country ihould begin to make obiervation. that appeared
accepted theory
 THE TOOLS OF STATISTICAL ANALYSIS 45

Nonlaboratory Sciences

In some sciences we cannot isolate factors one at a time in laboratory

experiments. In these sciences observations are still used to establish relation-
ships and to test theories, but such observations appear in a relatively
complex form, because several things are usually varying at the same time.
Consider, for example, the hypothesis that one’s health as an adult
depends upon one’s diet as a child. Clearly, all sorts of other factors affect
the health of adults: heredity, conditions of childhood other than nutrition,
and various aspects of adult environment. There is no possible way to
examine this hypothesis in the manner of a controlled experiment, for we
are unlikely to be able to find a group of adults whose diet as children
varied but for whom all other influences affecting health were the same.
Should we conclude that the hypothesis cannot be tested because other
factors cannot be held constant? If we did we would be denying the possi-
bility of many advances in medicine, biology, and other sciences concerned
with humans - and, therefore, with jointly fluctuating factors - that have
actually been made during the last hundred years. To conclude that our
hypothesis cannot be tested is would be im-
tantaifiount to saying that it

possible to discover anything about the relation between chemical com-

position, amount of oxygen, and speed of combustion if these factors could
not be controlled separately. Although the problem is more difficult if both
chemical composition and amount of oxygen vary from one experiment to
another, it should be obvious - and statistical theory provides a formal
proof - that the separate relation between combustion and chemical com-
position and between combustion and amount of oxygen could still be
discovered if enough experiments could be made. Testing is harder when
one cannot use laboratory methods, but, fortunately, it is still possible.
In a situation in which many things are varying at once, we must be
careful in our use of data. If we study only two people and find that the one
with the better nutritional standards during his youth has the poorer adult
health record, this would not disprove the hypothesis that a good diet leads
to better health. It might well be that some other factor had exerted an
overwhelming influence on these two individuals. The less healthy man may
have lived most of his adult life in a disease-ridden area of the tropics,
whereas the more healthy may have lived in a relatively healthy northern
climate. Clearly, a single exception does not disprove the hypothesis of a
relation between two things as long as we admit that other factors can also
influence the outcome.^

1 Note how often in ordinary conversation a person advances a possible relation (e.g.,
between education and some facet of a man’s character) and how someone else will ‘refute’ this
theory by citing a single counterexample (c.g., ‘my friend went to that school and did not turn
 46 SCOPE AND METHOD
How can we testan hypothesis where many things vary at once’
One
method that is sometimes available is to find a large number of people for
whom the other things are very nearlyequal and examine how (toconUnue
our example) their adult health is related to their diet as
children
If we cannot select data chosen m
such a way that other things were held
equal we have to fall back on more formal statistical
techniques that have
been designed to unscramble the separate effects of
several influencing
factors all of which are changing simultaneously The
most common tech
mque used for this task is called regression analysis which the reader will
encounter when he studies elementary statistics The most important
aspect of statistical analysis for present purposesmay be stated as follows
The techniques of statistical analysis show how, given
enough observations, it is generally possible to
identify the relationship, if one exists, between two
variables, even though other things which affect the
outcome are also varying

THE STATISTICAL TESTING OF ECONOMIC THEORIES

AN EXAMPLE
Economics is a nonlaboratorj science ft is rarely if ever possible to conduct
controlled expenments with the economy Millions of uncontrolled experiments
are, however, going on every day housewives arc deciding what to pur
chase in the face of changing pnees and incomes firms are deciding what to
,

produce and how to produce it, and governmental bodies are intervening
in the economy with taxes, subsidies and controls All of these acts can be
observed and recorded Thus a mass of data is produced continually by the
economy One of the basic problems of statistics is how to make sense of
data arising from uncontTolled rather than cxintrollcd experiments Most
things in which economists are interested, such as the volume of unemploy
ment, the level of prices and the distribution of income, are influenced by a
large number of factors, of which vary simultaneously If we are to test
all

our theories about economy, we will have to use those

relations in the
statistical techniques that were designed for situations in which other things
are not held constant

out like that It is a commonplace in everyday convenation to dismiss an hypothesis wth

)
gencrahsa
some such remark as Oh that i just a generalisation All inferesung hypotheses are
lor the two reasons describe
tions and some real or apparent exceptions will always 6e noticed
support the hypothesis as a
on pages 43-4 What we need to know is does the mass of evidence
issue can never be settled one
statement of a general tendency for two things to be related This
of evidence that just happen to be readily
way or the other by the casual quoting ofa few bits
available
 THE TOOLS OF STATISTICAL ANALYSIS 47

Let us consider, by way of an example, a simple hypothesis about the

factors determining the amount of money spent on food by households in
the United States. We shall advance the hypothesis that the larger the
household’s income the larger will be its expenditure on food, or, to use the
language of Chapter 2, that household expenditure on food varies directly
with household income.^
Let us say we start by observing three households. Our data are recorded
in Table 3.1.

Table 3.1

FOOD EXPENDITURE AND TOTAL INCOME

FOR THREE US HOUSEHOLDS
Household Expenditure on food Total household income per year
1 S2,000 S3,000
2 1,900 6,500
3 1,500 8,750

These data may lead us to wonder if our hypothesis is wrong, but, before we
jump to that conclusion, we also note that ‘by chance’ we may have
happened have selected three households that are not typical of all the
to
households in the country. Possibly, we say, the expenditure on food is in-
fluenced by factors other than income and possibly these other factors just
happen to be the dominant forces in these three cases.
To check on this possibility, we select a large number of households in
order to reduce the chances of consistently picking untypical households.
Suppose we do this by selecting 100 households from among our American
friends and acquaintances. A stadstician points out, however, that our new
group is a very biased sample, for it contains households with only a limited
occupational range, and possibly with incomes very similar to one another.
(Since we are especially interested in how expenditure on food varies as in-

come varies, this last point is likely to be a very serious one.) It is unlikely
that this sample of households will be representative of all households in the
US, which is the group in which we are interested.
The statistician suggests that we take a random sample of 1,000 households.
A random sample is chosen according to a rigidly defined set of conditions
that guarantees, among other things, that every household in which we are
interested has an equal chance of being selected. Choosing our sample in a

1 Such a relationship does undoubtedly exist, although it is somewhat more complex than
we allow Chapter 43, we report on the actual studies which
for in this introductory chapter. In
have been made into the determinant of consumption, the majority of which have been made
using US data.
 48 SCOPE AND METHOD
random fashion has two important consequences
first, it makes ,t tety un
likely that our sample will be eatremely
unrepresentative of all households
and, more important, it allows us to calculate
just how likely it is that ouj
sample is unrepresentatn e in any given aspect by
any stated amount The
reason for this is that our sample was chosen by chance, and chance events
are predictable ‘
This second result is very important because it
allows us to
make statements about the probability that all households m the US will
differ by any quantitative amount from the households in our sample
For eachmemberofthe sample we collect certain information,
including,
m tins case, the income of the household and the expenditures on food VVe

Income of household in dollars

Fjg 3 I Household income and expenditures on food

scatter diagram This is done m Figure 3 1

represent these data on a
when we measure income along the horizontal axis and expenditure on food
on the vertical axis Each dot represents one household and its position on
the graph tells us the income and food expenditure of that household (If
you are not already familiar with this basic kind of graph you should now
read the appendix to this chapter )
The scatter diagram in Figure 3 I suggests that there is some tendency for
expenditure on food to rise as income nses The relationship is not
perfect,

well
I This may sound paradoxical at fim But constder this If you pick a card from a
how hkely ml that you will pick a heart’ an ace’ an
shuffled deck of ordinary playing cards
questions and most serious card players do
ace of hearu’ If you know the answers to these

>ou must believe that chance events are in some sense predictable
 THE TOOLS OF STATISTICAL ANALYSIS 49

however, for there is considerable variation in food purchases that cannot

be associated with variations in households’ incomes. There are. for ex-
ample, 1 1 households in our sample with incomes of S6,000 and the chart
reveals that their expenditures on food var^^ from S500 to S2.500.
Using standard tools of statistical analysis we can now do the follounng
three things. First, we can ‘fit’ to these data a line that represents our best
estimate of the actual relation between household expenditure on food and
household income. Second, we can obtain a measure of the percentage of
the variations in household e.xpenditure on food that can be accounted for
by variations in household income. Third, we can apply a ‘significance test’
to discover the chances that the relation we have discovered in our sample
does not exist for the whole population and has arisen by chance because
we just happen to have selected for our sample households that are not
representative of all households in the US.
It is clear from the scatter diagram that we cannot account for all of the
variations in households’ expenditure on food by the variations in household
income. If we could, all the dots would be on a line. We may wish to look
for some other factor that might also exert a systematic influence on food
expenditure. We ask ourselves what could make one household with an in-
come of 86,000 spend twice as much on food as another household with the
same income. We look at the data we collected about the households that
appear in Figure 3.1 as extreme points. The observations marked a, b and c
in Figure 3.1 represent households whose expenditure on food is well above
the typical amount associated with the same level of income. Let us say that
we discover that most of the households in this class contain a large number
of children. The households represented by points d, e and f spend less on
food than we would expect, given a knowledge of their income, and we dis-
cover that most of these households contain only one or two persons. This
leads us to put forward the hypothesis that household expenditure on food
is higher the larger the number of persons in the household. We then investi-
gate this hypothesis by classifying households according to the number of
members and looking at the expenditure on food by each class of household.
We would see immediately that some relation between food expenditure
and number of persons in the household does exist: food expenditure tends
to rise as the number of persons in the household rises. The relation is far

from perfect, however, for there is considerable variation in expenditure on

food that cannot be accounted for by the number ofpersons in the household.
But does not surprise us since
this we already know that expenditure on
food associated with the size of household income. This suggests that we
is

should try to look at one of these two factors, holding the other constant.
One way in which we can tr)^ to do this is by cross-classifying our data as
shown in Table 3.2.
 50 SCOPE AND METHOD
Each TOW of this table exhibits the effect of family size
on expenditures on
food for a given level of income Each column shows the
effect of income on
expenditures on food for a given size of household Reading
down column
(2), for example, we see that households containing 3 to
4 persons had an
average expenditure on food of $690 when income was less
than $2,000,
S900 when income was between $2,000 and $4,000, $1,100 when income
was between $4,000 and $6,000, and so on

Table 3 2
AVERAGE HOUSEHOLD EXPENDITURES ON FOOD
CLASSIFIED BY INCOME AND FAMILY SIZE
Number of Persons per Household
Household income in

dollars per year 1-2 3^ 5-6 7 or more

Less than $2,000 600 690 750 805
2,000-3,999 790 900 960 1,100
4,000-5,999 950 1,100 1,200 1,250
6,000-7,999 1,200 1,400 1,580 1,810
8,000-9,999 1,800 2,600 2,900 3,500
10,000 or more 2,100 2,900 4,000 5,230

After seeing these data, we form the hypothesis that variations in house-
hold expenditure are the net result of two causes variations m household
income and vanations m
number of persons per household We can now
apply the statistical techniques of multtple-regression analysts to determine
three important things First, we can determine the most likely esumate of
the separate effects on household food expenditures of vanations m house-
hold income, number of persons constant, and of vanations number of m
persons per household, household income constant Second, we can calcu-
late a measure of the proportion of the total vanation m food expenditure

that can be explained by associating it with variations both in income and

in number of persons per household Third, we can discover how likely it is
that the relations we have found between food expenditure and income
and between food expenditure and number of persons do not exist for all
US households and have occurred in our data because (by bad luck) we
house-
chose a sample of households that was not representative of all the
holds m the US
STATISTICAL HYPOTHESES*
of
Statistical techniques can help us to measure the nature and strength
certain
economic relationships, and can tell us how probable it is that a
page discuwon of the distinction between statistical and deter
1 See Chapter 1. 10. for a

ministic hypotheses
 THE TOOLS OF STATISTICAL ANALYSIS 51

result has occurred merely by chance. What they cannot do is to prove that
an hypothesis is either true or false. Nor can they tell us when we should
accept or reject the hypothesis. We have already discussed this matter in
Chapter 1, but we now summarise the earlier discussion and then take it a
step further.

Can we Prove an Hypothesis is True?

Most hypotheses in economics are what may be called universal hypotheses.
They say that, whenever certain specified conditions are fulfilled, cause X
will always produce effect Y. We have already pointed out that universal
hypotheses cannot be proved to be correct because we can make only a
finite number of actual observations and we can never rule out the possi-

bility that we shall in the future make a large number of obsert^ations which
conflict with the theory.

Can we Prove an Hypothesis is False?

By the same token, we cannot get a categorical disproof of a statistical

hypothesis. Consider the hypothesis ‘Most crows are black’. We observe 50
crows 49 are grey and one is black. Have we disproved the hypothesis ? The
;

answer is no, for it is possible that this was just bad luck, and that if we could
observe all the crows in the world it would indeed prove to be the case that

most are black. Although we have not disproved the hypothesis, we have
learned something from our study of 50 crows and, if we have to make a
decision about the blackness of all crows, we will be very much better ofT for
having this information. In particular, we are likely to suspect that our
hypothesis is The question of decision-taking is considered
not correct.
below but in the meantime we must ask if it is ever possible categorically to
refute an hypothesis. To do this we need two conditions. First, the hypo-
thesis must admit of no exceptions ; it must say, for example, all crows are
black, or in the language of Chapter 1, it must be a deterministic hypo-
thesis. Second, we must be certain that any apparently-refuting observations
are not due to errors of observation. In this case we need only be certain
that we have observed one grey crow to refute the hypothesis. The observa-
tion of 49 black crows and one grey refutes the hypothesis all crows are
black providing we are sure that we genuinely saw a grey crow. But are we
sure that the odd bird really was a crow ? Are we sure that what looked like
a grey crow was not a black crow that had dust on him?^ Errors in
observation may always be present. For this reason, an hypothesis cannot
1 Even if we satisfy ourselves fully that we saw a grey crow, future generations may not
accept our evidence unless they go on observing the occasional grey crow. After all we no
 52 SCOPE AND METHOD
be refuted on the basis of a single conflicting observation
and indeed it can
never be categorically refuted no matter how many
conflicting observations
we make If we
observe 49 grey crows and only one black
one, our faith
in the hypothesis all crows arc black may well
be shaken and as a prachca)
measure we may choose to abandon the hypothesis (see
below) but we can
never be certain that all 49 cases were not due to erron
of observation
and had we persisted we might have ended up with 999,950 black
crows
and 50 grey ones which would make the hypothesis look pretty good,
since
a measurement error on 005 per cent of our cases might
not seem at all
^
improbable

The Decision to Reject or Accept is Subject to Error

VVe have seen that m general we can neither prove nor refute an hypothesis
conclusively, no matter how many observations we make Nonetheless, wc
have to make
decisions and act as if some hypotheses were refuted (i e we ,

have them) and we have to act as if some hypotheses were proved

to reject
(i e wc have to accept them) Just as a jury can make two kinds of errors,
,

/tndtiig an mnocenf man gut/cy, or fereing a guilty man go free, so can the
statistical decision*taker make two kinds of errors he can reject hypotheses

that are true, and he can accept hypotheses that arc false Like a jury, he
can also make correct decisions

Decision rules U'c noted above that the dectsion to accept or reject an
hypothesis is By using statistical analysis we can control the
subject to error
possibility of making errors even if we cannot eliminate it This is an ex-
tremely valuable thing to be able to do The method of control is to choose
the risk we are willing to take of rejecting an hypothesis if it is in fact correct
Conventionally, >ve use cut-off points of 5 per cent or 1 per cent If we use

longer accept the mass of well documented evidence accumulated several centuries ago on the
existence and power of witches even though it fully satisfied most contemporary observers
Clearly the existence of observational errors even on a vast scale has been shown to be
possible

even though it may not be frequent

1
• Advanced students may notice that the above appears to differ from the view expressed
difference anses from the fact that
by Professor Popper The Logic of Scientific Dtsem/erj Hie
m
1 take all empmcal hypotheses to be statistical
ones because of the universtil existence oferror
reject staustical hypotheses but so
of observation ^Ve do of course make arbitrary decisions to
rules of thumb for practical decision
also do we make arbitrary decisions to accept them These
methodolc^ical questions of whether any hypothesis can be
taking have nothing tc, do with the
conclusively refuted hypothesis can be conclusively proved My answer to
a„d whether any
both questions is no Those who arc not convmced by my arguments may proceed with the
text as long as they are prepared to accept
that most hypotheses in economics are suustical
hypotheses
 THE TOOLS OF STATISTICAL ANALYSIS 53

the 5 per cent cut-off point, we say that we will regard an hypothesis as
rejected if there exists less than one chance in twenty that we could have
made tlie same set of observations if the hypothesis were correct. Using the
1 per cent decision rule wegive the hypothesis a greater measure of reason-
able doubt; we reject hypotheses only if the results we observe could have
happened by chance one time in 100.
Consider an example. When studying expenditure our hypothesis might
have been that the consumption expenditure of US households falls as their
income rises. We tvould then ask what are the chances of making the
observations shown in Table 3.1 even though the hypotheses were correct.
There is always some chance that our sample was untypical of all US
households or that the relationship appears as it is because of measurement
errors. If we calculate that there is less than one chance in 100 of making
the observations in Figure 3.1 although the hypothesized relation actually
held for all US households, then we would abandon the hypothesis and for
practical purposes regard it as refuted.
When action must be taken, some such rule of thumb is necessary. But it
is important to understand, first, that we can never be certain that we are
right in rejecting a statistical hypothesis and, second, that there is nothing
magical about our arbitrary cut-off points. The cut-off point is a device
used because some decision has to be made. Notice also that decisions can
always be reversed should new evidence come to light.

QUANTITATIVE MEASUREMENT OF ECONOMIC

RELATIONS
So far considered whether certain observations support certain
we have
general hypotheses. The actual data does, for example, support the hypo-
thesis that households’ expenditure on food increases as households’ income
increases. This, however, is not enough. It is important to quantify such
qualitative statements. In this case, we should like to be able to say that
American household expenditure on food increases by some definite
amount, say lOff, for ever}' SI. 00 that household income increases.'
Economic theories are seldom of much use until we are able to give
quantitative magnitudes to our relations. For estimating such magnitudes,
our common sense and intuitions do not get us very far. Common sense
might well have suggested that expenditure on food would rise rather than
fall as income rose, but only careful observation is going to help us to decide

whether it typically rises by 10^ or by 20/ for every increase of SI. 00 in

income.
somewhat more complex and this simple one is used
I The actual quantitative relation is

solely for purposes of illustration.

54 SCOPE AND METHOD

One
of the major uses of statistical analysu is to help us to quantify
our
relations In practice, we can use actual observations both
to test the hypo
thesis that two things are related and to estimate the numerical values of the
function describing such a relation, if it exists
Very often the result of a of a theory is to suggest a new
statistical test
hypothesis that ‘fits the facts better than the previous one Indeed, in some
cases just looking at scatter diagrams (or
making a regression analysis) un
covers apparent relations that no one anuapatcd, and leads the economist
to formulate a new hypothesis You should look back to Figure I 1 on
page 1 8 once again, this time to see where such hypotheses enter the picture ‘

A WORD OF WARNING
In this chapter I have been concerned to dispel the common view that cco
nomists cannot be scientific m their use of data because they cannot make
controlled experiments The reader should not be left with the view that the
statistical tasks described in this chapter arc easily accomplished In fact

they are often very difticult and the pitfalls ready to trap the unwary user of
inappropriate statistical procedures are too numerous to mention Indeed a
whole new subject, econometrics, has grown up to amend existing tech-
niques and to develop new ones able to handle the special data problems
which occur in economics and other social sciences To launch into a career
in economic or social research without a full knowledge of the field of
statistical analysis is to take a severe nsk that one’s work will be useless
or

even downngh t misleading, because of ignorance of the possible pitfalls and

because of lack of knowledge of appropnate techniques for avoiding them

that onginatc from dau are $omeUm« called initulivt hypothesei m

conuast
1 Hypothwes
continuous The
to deduclivt one* But in any science the sequence of theory and icsung is
chicken
question of which came fint theory or observation » analogous to the debate over the
and the egg
 appendix to chapter 3

GRAPHING ECONOMIC
OBSERVATIONS

The popular saying ‘the facts speak for Table 3.3

themselves is almost always wrong if
Weight
there are many facts. We need theories Individual Height inLbs
to explain how facts are linked together
and we need summary measures to as-
A 5' 3" 130
B 5' 8" 160
sistus in sorting out what it is that the
C 5' 9" 155
facts do show in relation
to theories.
The simplest means of providing com-
D 5' 11" 180
E 6' 2" 185
pact summaries of a large number of
observations is through the use of tables
and graphs.
In this appendix we are concerned
with the two kinds of graphs that are
most often used to represent observa-
tions in economics, the scatter diagram
and the time series. Scatter diagrams and
time-series graphs are limited to the di-
mensions of a piece of paper; therefore,
they can be used only to depict the
variation of two variables. Height in feet and inches
Fig 3.2

the scatter diagram measures height. Each observation in

Suppose we have taken a sample of 100
men for the purpose of studying the re-
lationship between height and weight.
Table 3.3 shows the information about
five of them, but we have
data about
the other 95 as well.
Figure 3.2 plots the five points on a
graph in which the vertical axis meas-
ures weight and the horizontal axis
Table 3.3 represented by a small
is
circle in thegraph whose coordinates
show the height and weight of the sub-
ject concerned. Thus point B in the
figure refers to the individual called B
in the table, whose height is 5' 8" and
whose weight is 160 pounds. The stu-
dent should label the other four points
in the graph to be sure he understands
exactly what each one stands for. Since
 SCOPE AND METHOD
wc are interested m
the relation of points come to lying on one particular
height to weight in general (rather than line or curve , if the points he exactly
on
B’s height and weight in particular), we a line or curve, the relationship is de-
usually lea\e off the labels identifying terministic, if they are scattered very
the points as refernng to particular widely around the line, the relationship
individuals IS rather weak The line in Figure 3 3
Figure 3 3 shows the scatter diagram shovv? the average relation between
for our full set of 100 observations The weight and height, and the scatter of
five observations of the previous chart the actual observations around the line
appear as dots, the rest as crosses gives an idea of the strength of the re
The usefulness of the scatter diagram lationship * The fact that the points do

Height in feet and inches

Fig 3 3 Relation between heights and weights of 100 individuals

selected at random

IS that, once it is drawn, the relation- not all he on the line indicates that the
ship (if there is one) often leaps to the relationship is not deterministic

eye In Figure 3 3 wc have drawn a hnc
to suggest this relationship
Not only can we see immediately
from the diagram that there is a rela-
tion, but wc can gam some idea of its
Data of the kind used in the above
example are often called cross sectional
dataWc arc (m our example) com
panng the heights and weights of differ-
than
ent people at the same time rather

nature whether it is best described by

inconvenient just to trust our eyes to
a straight line or by a curve, and if by
a I It u
Elementary statistical theory
line, then of what slope We
can also draw luch a line

how a hni to the dau so that the

gam some idea of the strength of this re- shows us to Jit
average relation
line best describes the
lationship by seeing how close the
 4

GRAPHING ECONOMIC OBSERVATIONS 57

Studying the height and weight of a
single individual as he grows from in-
fancy adulthood. Cross-sectional
to
studies are very important in econo-
mics. We may look across households to
see the relationship of consumption
patterns to incomes ; w’e may
countries to see the effect of the
tities of natoral resources on economic
well-being, and so on.
for example, at the years 1908-9, we
see that unemployment was almost con-
stant at 8-5 per cent, whereas in 1891-
92, unemployment rose from 3 per cent
to 6-3 per cent- a large rate of change
look across
quan-

THE TIME SERIES GRAPH

Data are available showing the per-
centage of the unionised labour force
unemployed in the UK
for each year
since 1860. We
can plot this data on a
time series graph rvhere each point re-
fers to a year and to the corresponding
level of unemployment. This series is
shovvTi in Figure 3.4. The graph shows
that there is a very marked and regular
pattern of fluctuations in unemploy- Fig 3. Percentage of the unionised labour
ment over this period. This pheno- force unemployed in Britain, 1861-1913.

menon, the trade cycle as it is often

called, leaps to the eye. A
closer look at over a short period of time. We may
the time series suggests further ques- notice that when unemployment is

tions. How fast is unemployment changing a great deal over time the
changing over time? Is it changing a graph is steep, and when it is changing
great deal or only just a bit? If we look. only a little over time the graph is flat.
 CHAPTER 4

THE PROBLEMS OF
ECONOMIC THEORY

We are now ready ro begin ovr stvdy of economics So far in this 7ncro*
auction we have considered the general nature of social science and the two
major sets of tools of theoretical and statistical analysis This concluding
chapter of the Introduction provides a general vjew of the subject matter of
economics It is intended to give the student an idea of the relation between
the mam divisions of economics which we shall study in subsequent chapters
Theory is meant to relate to problems If the student cannot think of a set
of problems to which the theory he is studying might help to provide

answers, then either he or the theory has failed The student is advised after
readingthis chapter to refer back to it during the course of his study of the
remainder of the book, such references being particularly advisable when he
he has lost sight of the problems to which a particular part of
feels that
economic theory js directed

THE NATURE OF ECONOMIC PROBLEMS

Most of the problems of economics arise out of the use of resources to satisfy

human wants The resources of a soaety consist not only of the free gifts
of nature, such as land, forests and minerals, but also of human resources,
both mental and physical, and of all sorts of man made aids to further pro-
duction, such as tools, machinery and buildings It is sometimes useful to
divide the resources of any country into three mam groups ( 1 ) all those free
gifts of nature such as land, forests, minerals, etc ,
commonly called
NATURAL RESOURCES and known to the economist as land, (2) all human
resources, mental and physical, of both an inhented and
acquired sort,

called by the economist labour, and (3) all those

man-made aids to

further production, such as tools, machinery, plant and

equipment, in-
not consumed its own sake but
cluding everything man-made which is for
 the problems of economic
theory 59
wjiich used up in the process of making
IS
other goods, called by
economist CAPITAL. Economists call the
such resources factors of
because the resources of a society production
are used to produce those
things hai
people desire in order to satisfy their
wants. The things produced
are Llled
commodities. Commodities may be
divided into goods and services:
goods

educatiom This distincHon, however,

should not be exaggerated: any
s valued because of the
good
sennces it yields to its owner. In the
case of an auto-
consist of such things as
transportation,
mnh- h and, possibly, status.^
mobility
In most societies goods and seiraces
are not regarded as desirable in
them-
selves; no great virtue is
attached to piling them up endlessly in
warehouses
never to be consumed.^ Usually
the end or goal that is desired is that
the
individual should have at least
some
of his wants satisfied. Goods and ser-
vices are thus regarded as means by which the r/ii/ of consumer
satisfaction
may be^ reached. The act of making goods
and ser\'ices is called by the
economist production, and the act of using
these goods and services to
satisfy wants is called
consumption.

Scarcity

The human wants that can be satisfied by consuming

goods and services
may be regarded, for all practical purposes in today’s
world, as insatiable.^
The division of resources into land,
labour and capital, and the division of consumption
ommo itics into goods and sendees is a matter of definition. Definitions are to be judged not
grounds of usefulness and convenience. The question : ‘Is this
^ee o d division of resources the right one?’ is one
which has no meaning for the scientist,
e question. Is this
division likely to be a useful one?’ is a question that can fruitfully
be
iscussed. Arguments about definitions are one of the most common sources of futile debates
a fields. Such arguments are so common that they have been given a name,
essenlialisl
argumenls. An essentialist argument takes place whenever we have no disagreement about the
acts of the case but we
argue as to what name to use to indicate the agreed faets. We might,
or example, be in
complete agreement on what goes on in Soviet Russia and Communist
ina but we might get into
an argument as to which shouid be referred to as true Socialism,
uch an essentialist argument is a waste of our time and we would be better to call what goes
on in Russia X and China Y and see if we can get on svith defining some arguments of
su stance (e.g., does X provide more freedom to the e.xpression of dissent than T?).
2 This is intended as a
statement of fact, a statement about what is; it is not intended to
imply any value judgment
about what ought to be.
3 Whether or not it would
ever be possible to produce enough goods and services to satisfy
a human wants is
a question which we need not consider here. It would take a vast increase
m production (a percentage increase in the thousands) to raise all members of society
to the
standard at present enjoyed
by the richer members. It is doubtful that, even if this could be
one, all members of
society would find their wants fully satisfied so that there would be no
one who would desire
more goods.
60 SCOPE AND METHOD
In relation to the known desires of individuals, for better
food, clothing
housing, schooling, vacations, entertainments,
etc ,
the existing supply of
resources is woefully inadequate, it is sufficient
to produce only a small
fracuon of the goods and services that people desire
This fact gives nse to
one of the basic problems encountered m
economics, the problem o^scaraty
since there are notenough resources to produce everything we would like
to consume, there must exist some mechanism by which it is decided
what
will be done and what left undone, what goods will be produced
and what
left unproduced what quantity of each good will be produced, and whose
,

wants will be satisfied and whose left unsatisfied

Choice Choices are necessary because resources arc scarce The decision
to have more of one thing necessanly implies the decision to ha\e less of
something else All societies face this problem and somehow a decision must
be taken on such questions as, ‘How much of each commodity will be pro-
duced and how will it be divided up among the individual households in
the society’’ At this stage the student may ask ‘Who makes such decisions ’
Who actually chooses’’ In most societies many different people and
organisations either make or influence these choices Individual consumen,
business organisations, labour unions, farmen and government officials all

exert some influence One of the differences among various economies such
as jhose of the United States, the United Kingdom, India and the Soviet
Union is in the amount of influence that different groups have upon these
choices

Opporti/mtv cost Because resources are scarce, we are forced to choose

A choice means that y ou have one thing or the other If you choose to have

more of one thing, then, where there is an effective choice it will be neces-
less of the other Think of the members of an
individual
sary for you to have
household with a certain amount of resources answering the question How
shall we use our resources’ If we have more of this then
we must have less
to obtain some-
of that If by cost we mean what must be given up in order
of something else^
thing then the cost of having more bread is having less
some aCttirokncesra'i
Say that in this case the household decides to give up
third of the pnee ofa cinema
the cinema If the price ofa loaf of bread is one
seat then the cost of three loaves ofbread is
one cinema attendance foregone
is three
or, put the other way around, the
cost of one cinema attendance

loaves of bread foregone

whole society If the
Now consider the same problem at the level of the
thereby finds that it
government elects to bmid more roads, and
then
dorvn on its school construction programme,
mile of road If the government
can be expressed as so many schools per
 THE PROBLEMS OF ECONOMIC
THEORY 61
decides that more
resources must be devoted to
defence production then
wi be available to produce civilian
1
less
goods and a choice will have
to be made
between ‘guns and butter’ with the
cost of one expressible in
amount of the other foregone. The terms o"S
economist’s term for expressing
costs in
terms of foregone alternatives is
opportunity cost.

SUMMARY
Our discussion may now be summarised
briefly. Most of the problems
of
economics anse out of the use of
resources, land, labour and capital,
to
satisfy human wants.
Resources are used to produce goods and
services
which are then consumed by households
to satisfy their wants. The prob-
lem of choice anses because resources
are scarce in relation to the virtually
unlimited wants which they could be
used to satisfy. The concept of oppor-
tunity cost emphasises the problem of choice by measuring the cost of ob-
taining a quantity of one
commodity in terms of the quantity of some other
commodity which could have been obtained instead.

BASIC ECONOMIC PROBLEMS

Most of the specific questions pdsed at the beginning of Chapter 1 (and
many other questions as well) may be regarded as aspects of six more general
questions that must be faced in all
economies, whether they be capitalist,
socialist or communist. Economists
have been interested to find out how
decisions on these questions are arrived
at in various sociedes, and how
governments and other organisations can intervene to change
the answers
currently being given.

I Are the country’s resources being fully utilised, or are some

OF them lying idle ? It may seem strange that we should ask this question
at all. Surely, you will say, if resources are so scarce that there
is not enough
of them to produce of these goods that are urgently required, there can
all
be no question of the resources that are available being left
idle. Certainly,
no individual or government would plan to waste resources that are so scarce.
Yet, it is one of the most
disturbing characteristics of free-market economies
that such waste sometimes occurs. When
this happens the resources are said
to be involuntarily unemployed (or, more simply, unemployed). Unemployed
workers would like to have jobs, the factories in which they could work are
available, the managers and owners would like to be able to operate their
factories, raw materials are available in abundance, and the goods
that
could be produced by these resources are urgently required by individuals
m the community. Yet, for some reason, nothing happens: the workers stay
 }

62 SCOPE AND METHOD

unemployed, the factories he idle and the raw matenals
remain unused The
cost of such penods of unemployment is felt
both m terms of the goods and
services that could have been produced by the idle resources, and m
terms
of the terrible effects on human beings who arc unable to find work for
prolonged penods of time ‘
^
It is one of the most important problems of economics to discover
g
why it
ISthat free-market societies produce such penods of involuntary unemploy-
ment which are unwanted by virtually everyone tn the society Having discovered
why this is so, the next problem is to investigate how such unemployment
can be prevented from occurnng in the future
^ §
These problems have long been the concern of economists, and have been
studied under the heading of trade cvcee theory JhesT study wzsgnen
renewed importance by the Great Depression of the 1930’s For more than
ten years almost all Western countries experienced heavy unemployment
In the USA and the United Kingdom, for example, this unemployment
was never less than one worker in ten, and it rose to a maximum of approxi-
mately one worker in four This meant that, during the worst part of the
depression, one-quarter of the country’s resources were lying involuntarily
idle, while many millions of people remained without employment for a
penod of upwards of ten years A great advance was made in the study of
these phenomena with the publication in 1936 of the General Theory of
Employment, Interest, and Money,\>y M
Keynes This book, and the whole
branch of economic theory which grew out of it, has greatly widened the
scope of economic theory and greatly added to our knowledge of the prob-
lems of unemployed resources This branch of economics is called macro
ECONOMICS
More recently, quite high levels of unemployment have been suffered m

the United States Between 1958 and 1964 the unemployment rate flue
tuated between 5 and 7 per cent of the labour force After a long political
debate the remedies which Keynesian economics predicted would cure the
unemployment were adopted and within a very short time unemployment
fell as predicted Here is a sinking example of the
power of economic theory
in allowing us to control our emironment instead of remaining passive

victims of It

produced
2 In what quantities are goods and services being
arises directly out of the scarcity
of resources which creates i

This question

pmoa.l eapencim of un.mptoyniem o, drpr.oioo ihoold .ttooip

1 The itudent with „o
jhe Gnat
to (caio .omc idea of ihi. eapeneocc by leading ooo
or Iwo of ihe niaoy bool.“
Dep,e.i.ooof.hel930.T«..uehbook..„Goo,geO,we.lir».t!W..^
./ (Vikmg 1958) Bod.
Brace World, 1958) aod John Steinbeck. Ji,
S
also available in paperback
 ^

THE PROBLEMS OF ECONOMIC THEORY 63

situation in which the decision to use resources to produce more of one thing
necessitates using fewer resources to produce something else and thus pro-
ducing less of it. The question concerns the allocation of scarce resources

among alternative uses (a shorter phrase, resource allocation, vdll be used

hereafter). The questions ‘What determines the allocation of resources in
various societies?’and ‘What are the consequences of conscious attempts to
change resource allocation?’ have occupied economists since the earliest
days of the subject. Any economy in which resources are scarce in relatioji
to human wants whether it be capitalist, socialist or communist, must hav'e
some mechanism to produce decisions on the problem of resource allocation.
In free-market economies, the majority of decisions about the allocation of
resources is made through the price system and hence the study of the work-
ings of the price system is an extremely important branch of our subject.
This study is dealt with in the theory of price.

3 By \vhat methods are these goods produced? This question arises

w'henever there is more than one way
in %vhich goods
technically possible
can be made. Generally there are many such ways. Agricultural com-
modities, for example, can be produced by farming a small quantity of land
ver>’ intensively, using large quantities of fertiliser, labour and machinery,
or by farming a large quantity of land extensively, using only small quanti-
tiesof fertiliser, labour and machinery. Both methods can be used to produce
the same quantity of some good one method is frugal with land but uses
;

large quantities of other resources, w'hereas the other method uses large
quantities of land but is frugal in its use of other resources. Similar pos-
are available with manufactured goods; It is usually possible to
sibilities

produce the same output by several different techniques, ranging from ones
using a large quantity of labour and only a few simple machines to those
using a large quantity of highly automated machines and only a vety small
number of workers. Questions about why one method of production is used
rather than another, and the consequences of these choices about produc-
tion methods, are dealt with in the theory of production.

4 of goods allocated among the members who

How IS the supply
MAKE UP THE SOCIETY? The distribution of the national product among
various individuals and groups in the society is clearly of great general
interest. Economists have long asked w'hat governs the division of the national
product between various groups, such as labourers, capitalists and land-
owners.^ We also wish to know to what extent active government interven-
1 The important concept of a free-market economy is discussed at greater length in
Chapter 6, page 79.
2 In the eighteenth century when the theory’ of distribution was first developed the three
great social classes were workers, capitalists and land-owners, and the problem of distribution
 64 SCOPE AND METHOD
tion can, within the frame\vork of a free-market
society, succeed in altering
the distribution of income, and what are the consequences
of such inter
ventions ^
Such questions have been of great concern to economists since
the be-
ginning of the subject, and interest m
them ts as active today as it was almost
two centuries ago when Adam Smith and David Ricardo made their path
breaking attempts to solve them These questions come into the subject
of
tjie THEORY OF DISTRIBUTION

5 How EFFICIENTLY ARE THE RESOURCES BEING USED’ This quCStlOn

subdivides into two questions Is the production efficient’ and, Is it allo-
cated efficiently’ These questions quite naturally arise out of questions
2, 3
and 4 Having asked what quantities of goods are produced, how they are
produced and to whom they are allocated, it is natural to go on to ask
whether the production and allocation decisions are efficient ones Produc-
tion IS said to be inefficient if n would be possible merely to re allocate
resources and to produce more of at least one good without simultaneously
producing less of an> other good Existing production ts said to be in
efficiently allocated if it uould be possible to redistribute the existing
production amongst the individuals forming the society, and make at least
one person better off without simultaneously making anyone worse off
There is reason to believe that such inefficiencies exist in all economies
If they could be removed it would be possible to increase the producuon of
everything simultaneously and to make everyone m the society belter off
The importance of such inefficiencies, however, depends on their ^uantitaime
significance It would not be worth while spending time and effort to remove
them unless the cost of so doing was more than made up by the gains result
mg from their removal In fact, not enough is known about the quantitative
significance of such mefficienacs
Questions about the efficiency of production and its allocation belong to

the branch of economic theory called welfare economics A detailed

of
study of very difficult branch of economics is beyond the scope
this

this book Problems of efficiency will, however, be touched

upon at many
points
Questions 2 to 5 are related to the allocation of resources and
the distn

bution of goods and are intimately connected, in a market

economy, to the
system works They are sometimes grouped under
way in which the price
the general heading of micro-econokics

times
to explain how the national product was
»pbt up between these classes In modern
was
we are concerned to explain distnbuDon betswen aU
the vanous groups mwhich we may be

interested
 THE PROBLEMS OF ECONOMIC THEORY 65

6 IsTHE economy’s CAPACITY TO PRODUCE GOODS AND SERVICES

GROWING FROM YEAR TO YEAR OR IS IT REMAINING STATIC? This question
is one of great concern. If the capacity to produce does grow steadily, as
it has in most Western countries over the last few centuries, then a steady
is made possible. That the horror and poverty
increase in living standards
described in the England of 100 years ago by Charles Dickens is no longer
with us as a mass phenomenon is largely due to the fact that capacity to
produce goods and services has grown at about 2 per cent per year since
Dickens’ time while population has grown at a much slower rate. Why the
capacity to produce grows rapidly in some economies, slowly in others, and
not at all in yet other economies is a critical problem which has exercised
the minds of the best economists since the time of Adam Smith. Although a
certain amount is now known in this field, a great deal remains to be dis-
covered. Problems of this type are dealt with in the theory of economic
GROWTH.
There are, of course, other questions that arise, but these six are the major
ones common to all types of economies. Most of the rest of this book is de-
voted to a detailed study of these questions. First, we shall study how
decisions on these questions are made in free-market societies, and, second,
we shall enquire into the (often unexpected) consequences of settling these
questions through the price system.

The Questions Distinguished Diagrammatically

It has been a common source of error in the past to confuse these questions.
An answer appropriate to one has often been accepted uncritically as an
answer appropriate to another question, merely because two or more
questions could be stated in words which sounded similar.
The distinction between the four questions that are most commonly con-
fused call be emphasised by introducing a simple diagram.
Consider the choice that faces all economies today between producing
armaments and producing goods for civilian use. This is a problem in the
allocation of resources: how many resources to devote to producing ‘guns
for defence’ and how many to devote to producing goods for all other pur-
poses. We illustrate this choice in Figure 4.1. On one axis we measure the
quantity of defence goods produced and on the other axis the quantity of
all other goods. Next we plot all those combinations of defence and civilian

goods that can be produced if all resources are fully employed. We join up
these points and call the resulting line a production possibility boun-
dary. Points inside the boundary show the combination of defence and
civilian goods which can be obtained given the society’s present supplies of
resources. Points outside of the boundary show combinations which cannot
66 SCOPE AND METHOD
be obtained because there is not enoi^h productive capacity
to produce
them Points on the boundary are just obtainable, they are
the combination
that can just be produced using all the available supplies
of resources

— Production possibility boundary

Quantity of defence goods

Fig 4 I

The downward slope of the boundary indicates that there is an oppor-

tunity cost of producing more of one type of commodity, cost being measured
in terms of the quantity foregone of the other type of commodity Thus if
wc move from pome a to point b we arc reallocating resources out of civihan
production and into defence production and the quantity of defence pro
Full employment production

Fig 4 2
 THE PROBLEMS OF ECONOMIC THEORY 67

duction rises from Oq to Os while the quantity of civilian production falls

from Op to Or. Thus the opportunity cost of getting qs more defence goods
produced is pr civilian goods sacrificed. When we talk about moving be-
tween points a and b we are talking of the allocation of resources discussed
in question 2.
It is, economy to be at some point inside
of course, always possible for the
the production boundary. If the economy could be at point b
possibility
then it could also be at point c producing less of both defence and civilian
goods than at b, or indeed at any point inside the boundary. The reader
can easily check that, when the economy is located at a point inside the
boundary, production of both types of commodity is less than it would be if
some point on the boundary were attained.
An economy can be producing at some point inside its production possi-
bility boundary either because some of its resources are lying idle (question

1), or because its resources are being used inefficiently in production

(question 5). Figure 4.2 shows the production possibility boundary on the
assumption that one-third of the economy’s resources are, and continue to
The higher the proportion of resources unemployed the closer
be, lying idle.
will the broken line be to the origin.
Let usnow ask; ‘How can this economy produce more defence goods?’
Clearly we must know whether the present position is on the boundary or
inside it. If the economy is on the boundary, then, assuming for the moment
that the boundary cannot be more defence goods can
shifted, the answer is :

only be obtained at the cost of producing goods (e.g., by moving

less civilian

from point a to point b ) If, however, the economy is at some point, such as
.

c, inside the boundary, then more of both goods can be produced simul-

taneously. If the economy is inside the boundary because there is heavy un-
employment, then the measures which succeed in reducing unemployment
will allow the economy to have more of both goods. If, on the other hand,
the economy is inside the boundary because, although existing resources are
fully employed, they are being used inefficiently, then measures which in-
crease the efficiency of resource utilisation will allow the economy to
produce more of both goods.
Finally we come to the question of economic growth. If the economy’s
capacity to produce goods is increasing through time, then the production
possibility boundary is being pushed outwards over time as illustrated in
Figure 4.3.^ In this case, if the economy remains on the possibility boundary,
it be possible to increase the production of all goods over time, moving
will
for example from point a to point d.

1 Anything that increases labour productivity will push the production possibility
boundary
outwards from the origin. This could be done, for example, by the invention of new machines
which increased the hourly output of each worker.
68 SCOPE AND METHOD
Thus we see that if it desired to increase the production of all goods
is
m
an economy, it is necessary to do one of two things If production is
at a
point inside the production possibility boundary, then it may be moved
to
a point closer to, or actually on, the boundary, from c to i in Figure 4
2, for
example If the economy is already on the boundary, then it is necessary to
take steps which will move the boundary outwards so that production can
expand, for example from A to </ in Figure 4 3 It is very important to dis-
tinguish between two sorts of movements (i) a movement from a point
within, to a point on, the boundary, and (ii) a movement of the actual
boundary The conditions for doing the former, dealt with by the theory of
income and employment, are very different from the conditions for doing
the latter, dealt with by the theory of growth The fact that m both cases we
speak of an increase in national income (total production in the economy)

Fig 4 3

can cause confusion But a policy which would succeed in increasing

national income if the object were to move from a point within the bound-
ary to a point on the boundary might be a failure if it were necessary
to

increase national income by moving the actual boundary The student

will

avoid confusion if he uses the phrase economic grow th to refer to an

out-

ward shift of the production possibibty boundary, and the phrase a change
actual production
IN THE LEVEL OF ACTIVITY to refer to a movement of
boundary
away from (activity dcchn/ng) or towards (activity rising) the
due to a change m the proportion of resources employed
 PART 2

the elementary
theory of price
CHAPTER 5

A GENERAL VIEW OF
THE PRICE SYSTEM

All economies are faced with the problem of scarcity. Since there are not
enough resources to produce all the goods and services that could be con-
sumed it is necessary to choose, to decide what will be produced and what

leftunproduced. In short, resources must be allocated between the various

uses to which they could be put. In a market society, the allocation of re-
sources is the outcome of millions of independent decisions made by con-
sumers and producers acting through the mechanism of the market. In the
all

present chapter we shall give a short intuitive view of the working of the
market, and in subsequent chapters we shall formalise this view into a
definite theory of the behaviour of the market and at the same time con-
sider the empirical evidence relating to this theory.

A CHANGE IN DEMAND
By a change in demand we mean a change in the willingness of consumers
to purchase some product. As an example of such a change we shall con-
sider how the market reacts to a change in the tastes of individual con-
sumers. Let us say, for example, that consumers develop a greatly increased
desire for Brussels sprouts and a diminished desire for carrots; it might be a
matter of fashion, sparked off by some quite minor cause, or it might be the
result of a successful advertising campaign on the part of an association of
the Brussels sprout producers: ‘Eat Brussels sprouts, grown above ground.’
Whatever the reason, we can take it that there has been a major shift in
tastes in favour of sprouts and away from carrots.
What will be the effects of this change? Consumers will buy more Brussels
sprouts and fewer carrots. With production unchanged, a shortage of
Brussels sprouts and a glut of carrots will develop. In order to unload their
72 THE ELEMENTARY THEORY OF PRICE
surplus stocb of carrots, merchants will reduce carrot
pnces, on the grounds
that It IS better to sell them at a reduced price than not
to sell them at all
On the other hand, merchants will find that the) are unable
to satisfy all
their customers’ demands for Brussels sprouts, sprouts will become a scarce
commodity and the merchants will raise the price As the pnce nses, fewer
people will be able and willing to purchase sprouts Thus the demand
will
be limited to the available supply by the means of making the commodity
very expensive
Farmers will observe a rise in the pnce of Brussels sprouts and a fall m
the pnce of carrots Brussels-sprout production will be more profitable than
in the past, for the cost of producing sprouts remains unchanged while their
market pnce rises Similarly, carrot production will be less profitable than
in the past because the cost of producuon will remain unchanged while the
pnce fallsAttracted by high profits in Brussels sprouts and deterred by low
profits or possibly losses in carrots, producers wifi expand the production of
sprouts and curtail the production of carrots Thus the change m consumers’
tastes, working through the price system, causes a re allocation of factors of
production, out of carrot production and into Brus$eIs*sprout production
As the production of carrots declines, the glut on the market diminishes,
and carrot pnces will begin to nsc On the other hand, the expansion in
Brussels'sprout production reduces the shortage and the pnce will begin to
fall These pnce movements will continue until it no longer pays farmers

to contract carrot production and to expand the production of Brussels

sprouts When the dust has settled we shall end up in a position in which the
pnce of sprouts is higher than it was onginally, but lower than it was when
the shortage sent pnces soanng without any time for output to be adjusted
We shall also end up in a position in which carrot prices are lower than they
were onginally, but not as low as they were when the initial glut sent prices
tumbling without there being time enough for output to be adjusted
We can now see how the transfer of resources takes place Carrot pro
ducers will be reducing their production, and they will therefore be laying
off workers and generally demanding fewer factors of production
On the
other hand, Brussels-sprout producers will be expanding producuon
and

they will therefore be hinng workers and generally increasing their

demand
Labour can probably switch from carrot to sprout
for factors of production
producuon without much difficulty If, however, there are certain re

sources, in this case a certain type of land, say, that is

much better suited
for sprout growing than for carrot growing,
the demand for, and hence the
increasing sprout pro-
price of, this land will be affected Since farmers are
for those factors that are
duction, they will be increasing their demand
create a shortage and cause the
especially suited for this activity This wiU
price of these factors to nsc On the other hand, carrot producuon will be
 A GENERAL VIEW OF THE PRICE SYSTEM 73

falling, and hence the demand for resources especially suited for carrot
growing will be reduced. There will thus be a surplus of these resources and
their price will be forced down.
Thus, factors particularly suited to sprout production will be earning
more than previously, and they will obtain a higher share of total national
income than before. Factors particularly suited for carrot production, on
the other hand, will be earning than before and so
less will obtain a smaller
share of the total national income than before.

SUMMARY
1 A change in consumers’ tastes causes a change in demand, which in turn
causes a shortage or a surplus to appear. This in its turn causes market prices
to rise, in the case of a shortage, and to fall, in the case of a surplus.
2 The variations in market price affect the profitability of producing
goods — profitability varying directly with price. In search of profits, pro-
ducers will shift their production out of less profitable lines and into more
profitable ones,
3 The attempt to change the pattern of production will cause variations
in the demand for factors of production. Factors especially suited for the
production of commodities the demand for which is increasing will them-
selves be heavily demanded so that their own prices will rise.
4 Thus the change in consumers’ tastes sets off a series of market changes
that causes a re-allocation of resources in the required direction.
The theory that deals with point I is the theory of determination of market
price through demand and supply. Point 2, which concerns the reaction of
producers to market changes, is part of the theory of production. Point 3 is

dealt with in the theory of distribution.

A CHANGE IN SUPPLY
Now consider another change - this time on the side of producers. We
assume that, at existing prices, farmers become more willing to produce
sprouts than in the past less willing to produce carrots. There are many
and
things that could cause such a change. It might be brought about by a
change in the costs of producing the two goods - a rise in carrot costs and a
fall in sprout costs - or it might be brought about by a change in the tastes

of farmers everything else being equal, including the profits to be earned

:

from carrots and sprouts, farmers might just prefer to grow sprouts rather
than carrots.
Now what happen ? For a short time, nothing at all, for the existing
will
supply of sprouts and carrots on the market is the result of decisions made

3*
74 THE ELEMENTARY THEORY OF PRICE
by farmers some time m
the past But farmers now begin to plant fewer
carrots and more sprouts, and soon the quanutics on the market
begin to
change The quantity of sprouts available for sale nses and the quanuty of
carrotsfalls A shortage of carrots and a glut of sprouts results The pnee
of
carrots consequently nses and the pnee of sprouts falls As carrots become
more expensive, fewer people buy them, and as sprouts become cheaper,
more of them will be purchased The nse in carrot pnees and the fall in
sprout pnees now acts as an incentive for farmers to move back into carrot
production out of sprout production
m
We started from a position which there was a shortage of carrots, which
caused carrot pnees to nse The nse in carrot pnees removed the shortage
m two ways, first, by reducing the demand for carrots, which became pro
gressivcly more expensive to purchase, and, second, by increasing the supply
of carrots, which became progressively more profitable to produce We also
started from a position m which there was a surplus of Brussels sprouts,
which caused their pnee to fall The fall in pnee removed the surplus m two
ways, first, by encouraging the consumers to buy more of this commodity,
which became progressively cheaper, and, second, by discouraging the pro
duction of this commodity, which became progressively less profitable to
produce

THE MARKET MECHANISM

These examples illustrate man) of the important features of the pnee sys-
tem The allocation of resources is a result of decisions on the part of both
producers and consumers A change in the conditions either of consumers’
demand or of producers’ suppl> affects the allocation of resources and thus
also the final pattern of production and consumpUon in the economy
The
meckamsm by which these changes occur is through changes in pnees and
profits

It IS oftenremarked that m a free-market society the

is king. Such a maxim reveals only half
the
consumer
truth. Prices are determined by both demand and
supply. A free-market society gives sovereignty to two
decisions
groups, producers and consumers, and the
of both groups affect the allocation of resources

economist as conditions
Under certain very special conditions, known to the
sosereignty and becomes
ofPERFECT COMPETITION, the producer loses his
of the consumer These very
a mere automaton responding to the w.11 case.
special conditions are described m
Chapter 22 Aside from this special
 ;

A GENERAL VIEW OF THE PRICE SYSTEM 75

however, the producer has at his command, and actually does exercise, con-
siderable power economy.
in the allocation of resources in the
This general picture of the working of the price system has left untouched
many problems. Before ive can handle these problems, we must formulate
the ideas given in this chapter into a more precise theory of price. This will
be done in the folloiving chapters.

The Empirical Evidence

There is a great deal of empirical ewdence showing that, for many agri-
cultural commodities and industrial raw materials, the price system works
ver}' much as described in this chapter. In any retail or wholesale produce
market, prices can be obseiv'ed to react to the state of demand and supply,
prices rising when there is a shortage and falling w'hen there is a surplus.
Even the most casual obseia^ation of agriculture >vill enable one to obsen^e
farmers vaiyung their production of different crops as market prices vary.
A much more difficult question is whether or not it is valid to generalise
thisview of the price system into a theory of the prices of all commodities
agricultural goods, manufactured goods, and services. This question must
be postponed until after the theory' of price has been developed more fully.
 CHAPTER 6

THE THEORY OF
MARKET BEHAVIOUR:
SOME PRELIMINARY
CONSIDERATIONS

In this chapter we introduce some concepts and assumptions which form

the basis of a formal theory of market behaviour This is the first step re-
quired to develop the formal theory which occupies us for the next five
chapters

THE DECISION TAKERS

Economics is about behaviour, the behaviour both of individuals and of
groups of individuals We normally assume that anything mc observe in the
world and any relationship we assume in our theories can be traced back
to decisions taken by certain individuals or groups of individuals Who takes
the relevant decisions In elementary theory we have three groups of major
importance, households, firms and central authorities These are
the damatis ptTSonae oT economic theory and the stage on which much of their
play IS acted is called the market

The household By a household we mean all the people who live under
one roof and who make, or are subject to others making for them, joint

financial decisions In the theory of demand we take the household as our

basic atom We assume that the household usually behaves consistently
when faced with choices Like all assumptions of this type the test of its value
IS if predictions which follow from it fit the facts This we
shall consider in

a later chapter In the meantime we should notice that many interesting

problems concerning conflict wilhm the fomily and parental control over

the fate of minors are neglected because we take the household as a basic
 THE THEORY OF MARKET BEHAVIOUR 77

decision-taking unit. These intra-family problems are discussed by other

disciplines such as sociology, anthropology and psychology and we should
not expect to find problems handled within the field of economics. How-
all

ever, it is very important to remember that when economists speak of the

consumer or the individual they are in fact referring to the group of indi-
viduals composing the household. Thus, for example, the commonly-heard
phrase consumer sovereignty really means household sovereignty. These two
things are in fact quite different. It is one thing to say that individuals
should be free to decide their own fate and quite another thing to say that
the head of the household should be completely free to decide the fate of all
the members of the household some of whom will usually be minors.
The above is just a warning not to confuse households with individuals
in certain general discussions. In proceeding with the theory of market be-
haviour, we need only note that our theory assumes that individual house-
holds are consistent decision-taking units so that in their market behaviour
they behave as if they contained only one individual. The conflict problem
of how such consistent decisions are taken by the household (by paternal
dictatorship or by complete democracy) is not considered in our theory.

The firm; The firm may be defined as the unit which hires the services of
factors of production and usesthem to produce goods and services which it
sells either to other firms, or to households, or to the central authorities ;
the
firm is employment of factors of pro^
thus the decision-taker regarding the
duction and the production of goods and services. It cannot decide what its
sales shall be, but, through advertising and other media, it can try to influ-
ence the purchasing decisions of households which themselves determine the
firm’s sales. Since firms make decisions about production they are often re-
ferred to as producers. The firm is considered in much more detail in
Chapter 18. In the meantime we note the assumption analogous to the one
made for households ; that the firm makes consistent decisions in relation to
the choices open to it, and that the internal problems of who reaches parti-

cular decisions and of how they are reached can be ignored. In short the
firm is taken as our atom of behaviour on the supply side just as the house-
hold is taken as our atom of behaviour on the demand side.

Central authorities: This comprehensive term includes all public

agencies, government bodies and other organisations belonging to or under
the direct control of the government. It includes such bodies as the central
bank, the civil service, commissions and regulatory bodies, the cabinet, the
police force other bodies by which control can be exercised over the
and all

behaviour of firms and households. It is not important for the student to

draw up a comprehensive list of all central authorities but only to have in
 78 THE ELEMENTAR\ THEORY OF PRICE
his mmd the general idea of a group of organizations that exist at the centre
of legal and political power and exert some control over individual
decision
taken and over markets It is not a basic assumption of economics that
the
central authontics alwa>-s act in a consistent fashion as if they
were a single
individual Indeed conflict bclN^ccn difTcrent central bodies is the subject
of
much intermediate analysis of the theory of the control of the economy

THE INDIVIDUAL MARKET

Onginally the word market designated a geographically precise spot where
certain things were bought and sold Petticoat Lane (for many consumen’
goods) and Covent Garden (for wholesale fruit and produce) arc two world
famous examples of markets in the everyday sense Indeed much of eco-
nomic theory was onginally based on an attempt to explain pnee behaviour
in such markets Why, for example, can you sometimes obtain at the end
of (he day tremendous bar^gains, and at other times only get wliat you want
at pnccs which appear exorbitant in relation to pnees ruling a few hours
before’ The student who takes the trouble to visa any nearby street market
will secmany other interesting forms of behaviour most of which can be
explained by the theory of price that he is in the process of studying
As theories of market behaviour were developed they were quickly ex-
tended to cover such commodities as wheat Supplies of wheat produced
anywhere in the world can be purchased almost anywhere else in the world
«nd Its pnee tends toward unilbrmity the world over Clearly when we talk
about ‘the wheat market’ we have extended our concept of a market well
beyond the idea of a single place to which the housewife goes to buy some
thing The theory was also extended to cover future markets where things
which have not yet even been produced are bought and sold
To get a satisfactory definition' ofa market is not an easy task and some
of the complications arc discussed m Chapter 22 For present purposes we
define a market as an area over xthth buyers and stUers negotiate the exchange ofa
uell-defned commodity We
have already noted that the actual geographical
area covered by a single market will vary greatly with the commodity
In
case of
the case of wheat, the market is the whole Western world, in the
one
strawbemes, it may only be a small area including and surrounding
city, while in the case of hancuts, it may be a much smaller area than one

city
which the
Throughout Part II we shall confine ourselves to markets in
them has
number of buyers and sellers is large enough so that no one of
true and
1 By taihng of a MUifactory
dcfiniuon wc do not irwan to imply ihat there are
but only that the problem u to include what we wiot to
false definitions (see note 1. page 39)
the world of expenenw
include and to get a theoretical construct which can be related to
to hold
we will be able to identily the places where we expect the predictions of our theory
 THE THEORY OF MARKET BEHAVIOUR 79

any appreciable influence on price. This is a very rough definition of what

economists call competitive markets. Thus, the theory' developed in
Part II is the theory of competitive markets. In Part V we shall consider
the behaviour of markets that do not meet this competitive requirement.

A FREE-MARKET ECONOMY
A free-market economy is a collection of individual free markets. Such an
economy is one in which the allocation of resources is determined by produc-

tion, sales and purchase decisions taken by firms and households. The way
in which these decisions influence the allocation of resources by causing
market changes has been discussed in Chapter 5.
At the opposite extreme from a completely free-market economy is a
centrally-controlled economy in which all the decisions about the
allocation of resources are taken by the centnil authorities, and in which
firms and households produce and consume only as they are ordered.
Neither the completely free-market economy nor the completely con-
trolled economy has ever existed, at least in recent histoiy'. In practice all
economies are mixed economies in the sense that some decisions are taken
by firms and households, and some by central authorides. What does vary
between economies is the degree of the mix. In some economies the influence
of the central authorities is substantially lower than it is in others. Not only
may the average amount of central control vary between economies, it may
vary between markets within one economy. Thus, in Britain the day-to-day
behaviour of the stock market is left free from central control, while the
market for housing is subject to quite substantial amounts of regulation and
control by the central authorities.
The economic theory that we are developing is about the behaviour of
free markets, but can also deal with many of the types of central control
it

over markets that are commonly found in Western economies. We shall use
the phrase free-market economy to indicate economies for which the de-
cisions of individual households and firms exert a substandal amount of
influence over the allocarion of resources. The dmding line is an arbitrary
one and we must always remember that every shade of mix of central and
decentralised control exists, and that the economies of Poland and Russia
differ from those of France and the U.K. only in the degree to which the
central authorides exert an influence.^
I Free-market economies are sometimes called capitalist economies and we shall occasion-
ally use capitalist as a synonym for free market. The latter term is, however, the more
descriptive, since in this dimension economies are not differentiated by the extent of their use
of capital (indeed there is probably more capital per head in Soviet Russia
than in some
Western countries), but by the extent to which individual rtiarkets are controlled or not by the

central authorities.
 CHAPTER 7

THE ELEMENTARY THEORY

OF DEMAND

We seen that {fie marlct price of a commodity is influenced by the

demand ofhouseholds to purchase the commodity, and by the supply of the
commodity which firms ofTer for sale Tlie first steps, therefore, in de\e]op*
ing a formal theory of market prices are to consider the determinants of
households’ demand and firms’ supply Demand is considered this chap- m
ter and supply in the next
Both demand and supply are flows * We arc not concerned with a simple
isolated purchase but with a continuous flow of purchases, and we must
therefore express demand as so much per penod of tunc, say, ont orange per
day or, what is the same thing, sneii oranges per week or 5d5 per year

THE DETERMINANTS OF AN INDIVIDUAL HOUSEHOLD’S

DEMAND FOR A COMMODITY
By the households’ demand we mean the amount of some commodity which
an individual household is prepared to purchase This is a working defini-
tion to allow us to begin our study More precise definitions arc introduced
on pages 92 and 93 In the meantime let us consider the most important of
the factors influencing demand

1 The household’s demand for a commodity will be influenced by the

market price of that commodity In most cases we would expect that the
higher the price of the commodity, the lower would be the quantity
demanded
1Set Chapter 2, page 36 Such a view of demand raises difficulties with purchases of
durable consumer s goods when we deal with a single household Such difficulties disappear

when we consider the total market demand, for iC makes pertcetJ)' good sense to think of the
total demand for, say motor cars as being a flow of so many per month or per year
 THE ELEMENTARY THEORY OF DEMAND 81

2 The household’s demand for a commodity will be influenced by the size

of its income. In most cases we would expect that the larger the house-
hold’s income, the greater would be the quantity demanded.
3 The household’s demand commodity will be influenced by the
for a
prices of other commodities. Insome cases we would expect the demand
for one commodity to incretise as the price of a second commodity in-
creased (e.g., butter and margarine), while in other cases we would
expect the demand for one commodity to decrease as the price of a
second commodity increased (e.g., petrol and motor-cars).
4 A household’s demand for a commodity will depend upon the tastes or
preferences of its members. If, for e.xample, an individual member of
the household is influenced by some change of fashion, the household
may increase its demand for some goods and decrease its demand for
other goods, even though income and all market prices have not changed.

This list of factors influencing the household’s demand may conveniently be

summarised, using the notation developed in Chapter 2. What %ve have said
isthat demand (i.e., the amount of a commodity the household is prepared
to purchase) is a function of (i.e., depends upon) the price of the good in
question, the prices of all other goods, the household’s income and its tastes.

This statement may be expressed in symbols as follows:

~ ^{Pn>P\-> • •
'ipn-it ^

where D„ is the household’s demand for some commodity labelled ‘com-

modity n', where p„ is the price of this commodity, where />i, .,p„-i is a . .

shorthand notation for the prices of all other commodities, where Y is the
household’s income and T the tastes of the members of the household.^
This is quite a complicated functional relationship, and we shall not
succeed in developing a simple theory of demand or price if we consider
what happens to demand when these things — prices, income and tastes — all
change at once. To get around this problem we use a device that is very
frequently employed in economic theory. We assume that all except one
of the terms in the right-hand side of the above expression are held constant;
w'e then allow this one factor, say to vary, and consider how demand
(Z)„) varies with it, on the assumption that all other things remain un-

changed, or, as the economist is fond of putting it, ceteris paribus. We then
allow some other term, say income (T), to vary, and consider how, ceteris
paribus, demand varies as income varies. We can now consider the relation

1 This functional notation is merely a shorthand notation; it is pot of itself mathematics.

If you still find this troublesome you should read pages 21-4 and 33-4 of Chapter 2 and its
Appendix now.
 82 THE ELEMENTARY THEORY OF PRICE
between demand and each of the vanables on the nght
hand side of the
expression, taking them one at a time ^

1 The Relation between the Demand for a Commodity and the Price
of that Commodity,^

In the case of almost all commodities, the household’s demand increases as

the pnee of the commodity falls, income, tastes and all other prices remain-
ing constant As its price falls, a commodity becomes cheaper reJalive to its
substitutes, and it is therefore easier for the
commodity to compete against
these substitutes for the purchaser’s attention Generally, therefore, a fall in
the price of a commodity causes the household to buy more of that com-
modity and less of the otner commodities which compete with it, while a
rise in price causes the household to buy less of that commodity and more

of competing commodities Thus, the household does not always buy the
same bundle of goods it substitutes one commodity for another its budget
, m
as prices ^change.If, for example, Brussels sprouts become very cheap, the

members of the household will be induced, up to a point, to buy more

sprouts and less of other vegetables whose prices arc now high relative to
the price of sprouts
The relation that we have been considering between the demand for a
commodity and the price of that commodity may be illustrated on a graph

1 The Kudenc is often disturbed by his frequent encounien with ctkrvpenbus argumenu m
elementary economic theory Itcertainly imporunt to know how any two things are
u
related to each other (e g ,
how demand for a commodity is related to its own pnee) and
the
in order to deal with these quesuons ceUm panhas argumenb are necessary When employing
such -elationships, however, one must never forget the assumption that the relation holds only
if other things remain unchanged Many senous entns have resulted from the application of a

ctUris panbus argument to real world situations in which the other things did not and indeed

could not remain unchanged

Partial derivatives which are the mathematicians equivalent of ctttni parthus arguments
the
have a long and useful career to their credit The student should now reread section 1 1 of
Appendix to Chapter 2 (pages 41 2) in which the problem of trims panius reasoning is
discussed
4^' wr can ^oeciCv what
i* let

we are holding constant by writing,

T= r
! ,P. i=P\. ’Pl-i

vanables e T* might
where the superscript o indicates a particubr level for each of
the (i

at say, £l,000 per

mean for a particular household that lU incomes arc being held constant
notation looks rather forbidding at fint but it is just a way of specifying the other
annum) This
what is allowed to vary u wntten to the left of the vertical
things that are being held constant
Zinc and what is held constant is noted to the right of it
 THE ELEMENTARY THEORY OF DEMAND 83

by plotting the price of the commodity on the vertical axis, and the quantity
demanded on the horizontal axis. A cur\'e can then be drawm showing the
quantity the household r\'ill \vish to purchase at various alternative market
prices.

This curve, which shows the relation between the

price of a commodity and the amount of that com-
modity the household wishes to purchase, is called a
DEMAND CURVE, and it is drawn on the assumption that
income, tastes and all other prices remain constant.

An example of a demand curv'e is shoum in Figure 7.1. In this example,*

the household ^s^ll purchase 30 lb of commodity X
per month if the price is

0 20 40 60 80 100 120

Quantity of X in lbs per month

Fig 7.1 A demand curv'e.

6r a pound; it muU purchase 120 lb per month if the price is 2s, but only
10 lb per month if the price is lOr. If the price rises to 15r, the household
ceases to purchase thecommodity, while if the price falls to zero, that is, if
the good becomes free, it consumes 140 lb per month. The student can
check other points on the cur\'e to find out what quantity the household
would wish to purchase at various market prices.
A single point on the demand curve indicates a single price-quantity combination. For
example, the point a indicates that if the price were 5r the household
would be prepared to buy 45 lb per month, while the point ^ indicates that if
the price were 2s 6d the household would be prepared to buy 1 15 lb per month.

1 A particular numerical example is used to provide practice in using graphs. One must be
prepared, howev'er, to drop the numbers and to reason about demand curves in more genera!
terms. The reason for this is that in real-world applications we usually have knowledge about
the approximate shape of the demand curve, but we seldom have sufficiently precise knowledge
to plot a curve exactly on a numbered graph.
84 THE ELEMENTARY THEORY OF PRICE
The whole demand curve is a puture of the complete functional relation between
quantity demanded and price An economist often speaks of the conditions of
demand m a particular market being given or being known When he does
this he not referring just to the parUcular quantity that is being
IS
demanded
at the moment (i e not just to a parUcular point on the demand curve), he
,

IS refernng rather to the whole demand curve, to the complete

functional
relation whereby desired purchases are related to all possible alternative
prices of the commodity

2 The Relation between the Demand for a Commodity and the Prices
of Other Goods, ‘ On=f(p, Aj-i)
There are three possible relaUons between the demand for one good and the
prices of other goods a fall in the pnee of one good may lower the house
hold’s demand for another good, it may raise n or it may leave it unchanged
If a fall in the price of one good, Y, causes a fail m the demand for another
good, vY, the two goods, X and 1', are said to be substitutes When the
price of one good falls, the household buys more of it and less of goods which
are substitutes for H, thus the demand for a good vanes directly with the
price of Its substitutes This relation is illustrated in Figure 7 2(i) The
curve slopes upwards, indicating that as the pnee of its substitute nses, the
household’s demand for the commodity rises, while when the pnee of its
substitutefalls, the demand for the commodity falls Examples of goods
which are probably substitutes arc butter and margarine, Brussels sprouts
and cabbage, cinema seats and theatre seats, public transport and pmate
cars These goods are often said to be competitive with each other, the
terms competitive goods and substitutes being used synonymously
If a fall in the price of one good, Z, raises the demand for another good,
X, the two goods, Z and ^Y, are said to be complements In this case, when
the price of one good falls, more of that good is consumed and also more
obtain
of those goods which are complementary to it This relation will
between goods which tend to be consumed together, goods such as motor
and cups and saucers, bread and butter, rail tnps to
petrol,
Austna
cars
and skis This
illustrated is Figure 1 2(n) the curve slopes downwards
m ,

indicating that a fall the price of one good leads to a nse m the demand
m
^
for a complementary good
1 Or to spenfy ihe other things which are held constant

D. = {(pi i)|#. =
r= r
I r= v
that the definition given in the
2* Readers fam.l.ar with more advanced works wiH realne
where
used .n intermediate theory books,
present text does not conform exactly to one.
 THE ELEMENTARY THEORY OF DEMAND 85

i ii

Fig 7.2
(i) The relation between the demand for a commodity
and the price of a substitute.
(ii) The relation between the demand for a commodity
and the price of a complement.

3 The Relation between the Demand for a Commodity and the

Household's Income/ D„=i{Y)
Normally we would expect a rise in income to be associated with a rise in
the demand for most goods. There are two possible exceptions. In some
cases a change in income might leave demand completely unaffected. This
would be the case with goods the desire for which is completely satisfied
after a certain level of income is obtained. Beyond this level, variations in
income would have no effect on demand. This is probably the case with
many of the more inexpensive foodstuffs. It is unlikely, for example, that
thedemand for salt would be affected by either an increase in a household’s
income from ^2,000 to ^^2,100 per annum, or by a decrease in its income
from ;^2,000 to j()l,900 (although purchases might well be influenced by

definitions often involve sliding budget lines around a single indifference surface. Such defini-

tions may
be of use in developing theory but they are completely nonoperational. Also, since
in practice most income effects are very small, the more usual definition will be indistinguish-
able from the one we have adopted, in any practical problem of measurement (i.e., if it is a
substitute on the theoretical definition it will produce the empirical result on which we have
relied to define a substitute).
1 Or to specify more fully;

jD„ = f(r)i/.i, =
1 T= T
 THE ELEMENTARY THEORY OE PRICE
income changes if income were as low as, say,
j{;200 per annum) In other
cases It IS possible for a nse m
income beyond a certain level to lead to
a
fall in a household’s demand for a commodity Such a relation
is likely to
occur when one commodity is a cheap but inferior subsutute for some other
commodity An example might be an mfenor sort of black bread
When
incomes arc low, money might be saved by using black bread
exclusively
But at higher levels of income the consumer might feel that
he was able
to change to white bread Thus, as income rose
beyond a certain amount,
the demand for black bread would fall and would possibly
reach zero
as income reached a le\el at which the consumer
ceased to worry about

Household income per period of time

Fig 7 3 The relation between the demand for a commodity and the

level of a household s income

spending a few pence more or less per day on bread Commodities the

demands for which fall as income nses arc called by the economist inferior
GOODS
These relations are illustrated in Figure 7 3 The three curves indicate

different functional relations between income and the demand for a com-
modity, and all prices being held constant Curve (I) illustrates what
tastes

IS possibly the most common case the one m which a

nse in income bnngs
a case
about a nse m purchases at all levels ofincome Curve (2) illustrates
(income Oe) and
in which purchases nse ^vith income up to a certain point
Curve
then remain unchanged as income vanes above that amount (3)
income
illustrates the case of an mfenor good purchases of which rise with
up to a certain level (at income Oi where purchases are Oi/), but then fall
as the household s income nses beyond that level
 THE ELEMENTARY THEORY OF DEMAND 87

4 The Demand for a Commodity Depends upon the Tastes or

Preferences of the Members of Society
If itbecomes fashionable amongst middle class households to have a second
car, the flow of expenditures on cars will increase. This does not mean that
everybody rvill buy a second car, but some people will, and demand will
increase. Some changes in tastes are passing fads, hke rock and roll and
black leather boots; other changes in taste are permanent, or at least long
lasting, like the rising popularity of car owmership in Britain, or the switch
to filter cigarettes and ball-point pens.
The economist often regards tastes as given, and changes in these tastes
as exogenous and therefore as outside his province. Tastes are sometimes
thought of as arising out of the basic wants and needs of human beings and
as such as being more nearly in the realm of the biologist and the psycholo-
gist than of the economist. Changes in taste may, however, be the result of
economic activities. The whole purpose of advertising, which is a major
industry in most Western countries, is to change people’s tastes, and there
is little doubt that it succeeds. Why else w'ould Americans buy electric can

openers? Other people influence taste also. Keeping up with the Joneses
creates a bandw'AGOn or demonstration effect; the quantity and
quality of a good one person buys depends upon the quantities and qualities
other people buy. A man may develop a taste for fine wine after tasting
some, but he may also develop it after discovering that serving it enhances
his prestige. In any case, people have tastes and and these pre-
preferences,
ferences change. When they do, the demand for some commodities increases
and for others declines.

MARKET DEMAND CURVES

Up to now we have discussed how' the of an individual household
demand
depends on such things as prices and income. In the theory of price we are
concerned with the total demand for some commodity on the part of all
households. For each individual household there a relation betw'een the
is

price of a commodity and the quantity of that commodity which it will de-
mand. The MARKET demand is merely the sum of the demands of all the
individual households. The relation between the demand curves of house-
holds and the market demand curves is illustrated in Figure 7.4, where, for
simplicity, w'e deal with a market containing only two households. We
assume that we know the complete demand curve for each household and
we show these in Figures 7.4(i) and (ii). From these individual demand
curves we have derived the market demand curve which merely shows how
much will be demanded at each price by both households. Geometrically,
the market demand curve in (iii) is derived by a horizontal summation of
 ’"E elempntarv theory or
price
die two .nd.vtdual curvetm {.) and (ti) At a pnee ofSr, for example
house-
“ household (ti) demands Oi, the
total Lmand ,s
ol - Or, which quantity ts plotted
Oa+ 06
(7»+
tn Figure 7 4(ui) against
the once
household () demands OJ, while
household (u)
demand
demands Or, and total demand is 0d+0,=0/ Thus the market demand

‘ " ill

Ftg 7 4 The derivation of a market demand eune by the aggregation

of the demand curves lor individual households

curve IS to be thought of as the (horizontal) sum of all the demand curves

of the households in the market *
In practice we seldom have information about individual demand curves
although vve often do have evidence about the general shape of market dc-

I When su nming curve* students someumrs become coafused bcf»een \ertical and hon
zontat summation Such a confusion can only resuli from the application of memory rather
chan commoci sense to one * economics CmtJrr utat maid be meant by vetuca! nmmdtio’i Lay
off the equal distances Oa and Od id curves (} and (u) representing the same quant ty
demanded by each consumer These quantiues correspond to prices 9s in (i) and 1 Ii in (ii)

.\<?K inFigure (ill) layoff OY=OaseOfiandp}<rtKagiiastOiepncr9f-b-ih=20/ Tibj*poujt

now relates a certain quantity to the sum of the prices vvhKh households (i) and (ii) are
separately prepared to pay for this comnxidity Clearly this information is of no interest to us
m the present context Faery graphical apeialan cart be Iraaslaled w/« uords The advantage of
graphs is that they make proofs easier, ihe disadvantage is that they make possible to make
it

silly enon To avoid these the student ibotild always translate snXo words any graphical opera
lion he has performed and ask himself Does this make sense and is this what I meant to do’
For example a market demand curve is meant to tell us total purchases at each price and
hence it is obtained from individual curves by adding up the fusnhl es demanded by each con
sumer at given pnees not by adding the which each ccBisumer would pay for some given
quantity But see Chapter 38 where vertical summation of demand curves is required
 THE ELEMENTARY THEORY OF DEMAND 89

mand cur\'es. This derivation of market demand cur\-es by summing indi-

vidual curves is a theoretical operation. W’e do this because we wish to
understand the relation between curves for individual households and
market curves.
^Vhen we go from the individual household’s demand curv’e to tlie market
demand curve we must add two new determinants of demand to the list
given on pages 80-1.

5 The Demand for a Commodity Depends Upon the Size of the

Population

As population grows, more people need to be fed, clothed, housed and en-
tertained, and thus demand increases as population does.* Economists treat
population change as another exogenous determinant of demand.

6 The Demand for a Commodity Depends Upon the Distribution of

Income Among Households

Imagine a society, an oil-rich sheikdom, say, in which the average income,

though seemingly high, has been obtained by averaging the very' low in-
comes of a large number of poor households and the veiy' high incomes of
the few enormously wealthy ones. IVe would expect this society to have a
very different pattern of demands from a society in which the same average
income is distributed more equitably. Closer to home we would expect that
any measure which redistributed income from single persons to married
couples would change demand in favour of furniture, baby goods and other
things bought by couples with children, and away from the sort of thing
consumed by bachelors.

The Importance of the Market Demand Curve

The market demand curve relates the demand for a product to the price of
that productand it is the most emphasised of all the relations in the demand
function shown on page 81. tVhy? Much confusion exists on this subject.
fVe do not imply {or believe) that price is the most important determinant of demand.
In the earlier sections of this chapter, we discussed other important factors,
and if economists had to choose a single most important v'ariable, it would

1 The relation is not quite as simple as this since we not only need mere people. the\ must
be able to buy things before demand increases. But a change in total population without an)
change in the percentage of total resources unemployed tvill shift most market demand curces
to the right.
 y

TnE ELEMENTARY THEORY OF PRICE

probably be incorac ‘ (Wc do not, howcYer, have to choose,
as ivill become
'
ImlZ r "T'”'“ T“” L
^P'-rsentmg the

wh« det"'”"
of what
or determine marlet “T"'
i"' price, thtn tt h convenient to have pnee
ft'O'y
as one
01 our two \anablc5

Shifts in the Market Demand Curve

We must now consider the elTect on the demand curve of a
change in each
of the other factors which were hetd constant
when we drew the curve
Ihesc effects arc, of course, implicit in what
has been said about the rela
tion between demand and each of these other
factors

fig 7 5 An increase in demand

1 The effect on the demand curve of a change in income It has

already been argued tint, m the case of most commodities, a nse in income,
cetfru paribus, will cause an increase in demand Therefore, if household
income nses, we shall find (hat, v^hateser the pnee wc consider, there will
be an increase in the amount that is demanded at that price Graphically,
the whole demand curve shifts to the nght The curve shifts for example,
from DiDi to ZJjDj m Figure 7 5 The curve Z),Z)| shows the relation be-
tween the pnee of the commodity and demand on the assumption that in-
come 13 held constant at some level, Yi , the curve £> 2^2 shows the relation
between demand and pnee, on the assumption that income is also held con-
stant, but at some other level, 1 2 > which is greater than Y The shift of the
curve from Dy to Di indicates an incrcaso m
desired purchases at every

possible pnee For example, at the pnee Oa desired purchases increase

by
from
the amount be from Ob to Oe Another way of showing that the shift
Dy to £>2 does represent an increase m
demand is to show that, for any

1 Se« Chapter* 1 1 and 56 (pages M8-51 and 81 1-15) Qn the relauon between growth and
the allocation of resource* for a partial justification €>f thu statement
 THE ELEMENTARY THEORY OF DEMAND 91

given quantity, a higher price \vill be paid than previously. For example,
the quantity Ob can be sold at the price Oa when the demand curve is
DijDj, but the same quantity can be sold for the much higher price Oe when
the demand cur\'e is D2 D 2 .'

In the case of an inferior good, a income will cause a reduction in

rise in

the quantity demanded at each market price, and the whole demand curve
moves to the left.

2 The effect on the demand curve of a change in the prices of

OTHER goods: Here the effect depends on whether the good, whose price
changes, a complement or a substitute. Consider, for example, the effect
is

on the demand cur\'e for petrol of a rise in the price of cars. When the price
of cars rises, fewer cars will be bought and hence less petrol will be bought
at each possible price of petrol. Thus the rise in the price of a complementary'
good, cars, shifts the whole demand cur\'e for petrol to the left, indicating
that at each price less will be bought than previously. Now consider the
effect on the demand of a rise in the price of public trans-
cuiA'e for petrol
port. In this case, households will be more inclined to use their own cars
and more petrol will be purchased at each price than previously. Thus a
rise in the price of a substitute good shifts the demand curve to the right,
indicating that, at each price, a larger amount will be demanded than
previously.

3 The effect on the demand curve of a change in tastes: This

problem, of course, is quite simple. A change in tastes in favour of a com-
modity will mean that, at each price, more will be demanded than pre-
viously, so that the whole demand curve will shift to the right. On the other
hand, a change in tastes away from a commodity will mean that, at each
price, less will be demanded than previously, so that the whole demand
curv'e wll shift to the left.

SUMMARY
Figure 7.6 summarises the preceding discussion in which we have con-
sidered the effects on the demand curve of changes in the other things which
are assumed constant when one curve is drawn. It is, of course, possible to
do the same things for the curves illustrated in Figures 7.2 and 7.3, and the
Thus a rightward shift in the demand curv’e indicates an increase in demand in the sense
1

both that more is demanded at the same price and that a higher price would be paid for the

same quantity. It is, of course, true that the amount demanded at the point y on is less than

the amount demanded at the point xon D,. Thb comparison merely shows that, in spite of the
increased desire to purchase the good, a sufficiently large rise in price can reduce the quantitv
actually demanded to an amount lower than it was originally.
 the elementary theory of price
reader should check that he understands
the analysis by showing the shifts
m these cura-cs caused by variations m
the other factors that were
assumed
to be constant when the particular
curve was constructed (For example
what will happen to the curves in Figure 7 3 if there
is a fall the price of m
the product question m

« H

Ftg 7 6
(i; A rise in demand more u demanded at each pnee This can be caused by
(1) a nse m income
(2) a rise m the price of a substitute
(3) a fall in the price of a complement
(4} a change in tastes m faxour of this commodity
(ii) A fall in demand - less is demanded at each price This can be caused b)
(1) a fall m income
(2) a fall in the price of a subsutute

(3) a rise in the pnee of a complement

(4; a change in tastes against this commodity

Movements along Curves versos Shifts of Curves

most important in economics to distinguish between a movement along a

It ts

demand curve and a ffii/t of the whole curve A movement along a demand
curve indicates that a different quantity is being demanded because the
price
quantity
has changed A shift of a demand curve indicates that a different
willbe demanded at each possible pnee because something else
either m
There is no
comes, tastes or the price of some other good has changed
quite different
generally agreed terminology to distinguish between these two
occurrences a moxement along one curve and a shift of the
whole curve
This absence of agreement on the use of words can be confusing
When the
 THE ELEMENTARY THEORY OF DEMAND 93

economist speaks of an increase or a decrease of demand, he is usually re-

ferring to a shift of the whole curve, because he is more concerned with the
whole functional relation between demand and price than with the particu-
lar quantity that happens to be demanded at any one moment. We shall
follow this usage, and when we speak of an increase or a decrease in demand
we shall be referring to a shift in the whole curx'e - to a change in the quantity
that will be demanded at each possible price. When we refer to a movement
along a curve, to a change in the quantity demanded because price has
changed, we shall refer to a change in the quantity demanded, specifically, to
an increase in the quantity demanded, indicating a movement down the
curve because of a fall in price, or a decrease in the quandty demanded,
indicating a movement up the curve because of a rise in price.

THE DEMAND FOR PETROL: AN EXAMPLE

The preceding discussion of the factors influencing demand may be reviewed
by considering, as an example, the demand for petrol. The demand for this
commodity may be expected to varj’ inversely Math its price (Figure 7,1) as ;

the price falls, more petrol will be consumed. This will occur because exist-
ing car owners will use more petrol, because new purchasers of cars will
worr)' less about obtaining cars with low petrol consumption, and because
some non-car owners will now feel they are able to afford to run a car. For
all these reasons the demand may be expected to rise as its price
for petrol
falls. The demand for petrol may also be expected to vary inversely with the
price of cars (Figure 7.2(ii)). As the price of cars falls, more households will
purchase them and thus there will be increased purchases of petrol (the
price of petrol remaining unchanged) Petrol and
. cars are thus complemen-
tary goods. On the other hand, the demand for petrol can be expected to
vary directly with the price of public transport - a fall in the price of public
transport leading to a fall in the demand for petrol, and a rise in the price
of public transport leading to a rise in the demand for petrol. If the price
of public transport rises, car owners can be expected to use their own
vehiclesmore frequently and public transport less frequently and it is pos-
sible thatsome non-car owners will be induced to buy cars because public
transport is now more expensive. Public transport and petrol are thus sub-
stitutes for one another. Finally, the demand for petrol -will vary directly
with household incomes (Figure 7.3, curve I). A rise in consumers’ incomes
will lead to a rise in petrol consumption. This will occur because car owners
will use their existing cars more frequently, because some households \vill
switch to more expensive cars which generally use more petrol per rmle than
do the less expensive ones, and because some non-car owners will now pur-
chase cars as their incomes rise.
 CHAPTER 8

THE ELEMENTARY THEORY

OF SUPPLY

By the supply of a commodity wc mean the amount of that commodity

that producers are able and willing to offer for sale Like demand, supply is
a flow, It IS so much per day, week, month or year

VVe shall make a very superficial study of supply in this chapter, estab
Ushing only what is necessary for a simple theory of price In Part JV we
shall devote considerable attention to the theory of production, which is
the branch of economics concerned with the determination of supply

FACTORS INFLUENCING SUPPLY

We now introduce five hypotheses about the most important factors
influencing supply

1 The SUPPLY OF A COMMODITY DEPENDS UPON THE GOALS OF FIRMS If

drug companies prefer to engage in the production of medicines rather than
rat poison because it makes them feel more important m
society, we expect
more medicines and less rat poison to be produced than if producers held
all commodities in equal regard If producers of some commodity want to

sell as much as possible, even if it costs them some profits to do so, more ^Mll

be sold of that commodity than jf they wanted to make maximum profits

If producers are reluctant to take risks, v>« would expect smaller production
of goods whose production is nsky
In standard economic theory we assume the goal of the firm is to make as
much profit as is possible The implications of this hypothesis, the implica
tions of alternative hypotheses and the consequences of the rejection
of the

‘profit-maximismg hypothesis’ are considered in great detail m

Part IV

2 The supply of a commodity depends upon the price of that

COMMODITY CeUrts paribus, the higher the price of the commodity, the more
 THE ELEMENTARY THEORY OF SUPPLY 95

profitable will it be to make that commodity. We expect, therefore, that the

higher the price, the greater udll be the supply.

3 The supply of a commodity depends upon the prices of all

OTHER commodities: Generally, an increase in the price of other com-
modities will make production of the commodity whose price does not rise
relatively less attractive was previously.
than it We thus expect that, ceteris

paribus, the supply of one commodity would fall as the prices of other
commodities rise.

5
4 The supply of a co.MMODiTy depends upon the prices of factors
OF production; A rise in the price of one factor of production will cause a
large increase in the costs ofmaking those goods which use a great deal of
that factor, and only a small increase in the costs of producing those com-
modities which use a small amount of the factor. For example, a rise in the
price of land will have a large effect on the costs of producing wheat and
only a ver^' small effect on the costs of producing motor vehicles. Thus a
change in the price of one factor of production will cause changes in the
relative profitability of different of production and this will cause pro-
lines
ducers to shift from one line to another, and so cause changes in the supplies
of different commodities.

The supply of a commodity depends upon the state of tech-

nology: The enormous increase in production per worker that has
occurred in industrial societies over the last 200 years is very largely due to
improved methods of production. These in turn have been heavily influ-
enced by the advances of science. But the Industrial Revolution is more than
an historical event; it is a present reality. Discoveries in chemistr)' have led
to lower costs of production of well-established products, like paints, and to
a large variety of new products made of plastics and synthetic fibres. The
new electronics industry rests upon transistors and other tiny devices that
are revolutionising production in television, high-fidelity equipment, com-
puters and guidance-control systems. Atomic energy will one day be used
to build canals and to extract fresh At any time what is
water from the sea.
produced and how it is produced depend upon what is known. Over time,
knowledge changes and so do the supplies of individual commodities.

THE SUPPLY FUNCTION

We may summarise the preceding discussion as follows: the supply of a
commodity is a function of the price of that commodity, the prices of all
 96 THE ELEMENTARY THEORY OF PRICE
Other commodities, the pnccs of the factors of
production, technology ard
the goals of producers Or, m
the notation of Chapter 2,

= iiPnspu >Pn-i,Fu ,F„ G, T)

where is the supply of the commodity /?, js the price of that commodit)
Pu ypn-i IS shorthand for the prices of all Other commodities, F], ,f,
IS shorthand for the prices of all factors of production G the tastes of pro-
ducers and T is the state of technology

Supply and Price

For purposes of a simple theory of pace we msh only to hnow how th(
supply of a commodity varies with us own price, all other things being helc
constant ‘
We are only concerned, therefore, with the ceferts paribus rcla

— ^p„) There is much that needs to be said on the relation be

tion,^ S„
tween supply and price. For the moment we shall content ourselves witl
the intuitively plausible hypothesis that, cetms panbus, the quantity ofi
commodity produced and offered for sale will increase as the price of th
commodity rises and decrease as the pnee falls (i e ,
quantity supplied ani

price will vary directly with each other) This hypothesis has a stroni

common sense appeal, since the higher is the pnee of the commodity, th
greater are the profits that can be earned, and thus the greater is the in
centive to produce the commodity and offer it for sale The hypothesis i

known tobe correct in a lar;ge number of cases

We shall proceed for the next few chapters assuming it to be generall;
correct The exceptions to the hypothesis and their implications will bi

studied when we come to the theory of production

The supply curve This hypothesis may be illustrated on a graph tbai

measures price on the vertical axis and quantity on the horizontal one Thf
curve, which is drawn in Figure 8 I, is called a supply curve, and it shows
the quantity producers will wish to make and to offer for sale at various

1 We do ihis because we arc

concerned at this staijc to develop a theory of price \V e do no
mean lo imply that price is the quaniiiauvety most important detemiaaat of supply Over an)
roost important determinant although as we shall set
long period technology w probabJy the
in Chapter 21 the question of the
exieni to which technological changes themselves occur in

response to price changes is a difficult one to answer

clause explicitly we should write
2 To state the esltrii par, bus
 THE ELEMENTARY THEORY OF SUPPLY 97

.ternative prices of the product. When the economist speaks of supply he is

iferring to the whole supply cur\'e, which shows what would be supplied at
ach price.

Quantity per period of time

Fig 8.1 A supply curve.

A movement from one point on the curve to another point on the same
:urve will be referred to as a change in the quANTiTY supplied and it

ndicates the response of firms to a change in the market price of the com-
modity. Thus, in Figure 8.1, Oc is supplied when the price is Oa while a
ise in the price to Ob causes a movement along the supply curt'e with the
quantity supplied increasing to Od.

SHIFTS IN THE SUPPLY CURVE: A shift in the whole supply curve must be
iue to a change in some factor other than the price of the commodity and
t will be referred to as a change in supply. Just as with demand, it is

:xtremely important to distinguish such shifts in the whole curve from a

movement along a curve due to a change in the price of the commodity.
An increase in supply is illustrated in Figure 8.2(i). A movement from
?!to $2 represents an increase in supply in that at any price more is supplied
:han previously. At the price Oa, for example, the amount Od wall be sup-
pliedwhereas previously only the amount Oc would have been supplied.
Fhe rightward shift of the curve also means that the same quantity wll be
3ffered at a lower price than previously. The price of Oa, for example,
.vould call forth a supply of Oc when the supply curve was S^, whereas a
3rice of only Ob is all that is now' necessary to call forth the quantity of Oc.
Both of these comparisons indicate an increased willingness on the part of
producers to make the commodity and offer it for sale. A
shift of the supply
:urve in the opposite direction, as shown in Figure 8.2(ii), indicates a
reduction in supply.
What can cause a supply curve to shift? Clearly, the curve is shifted by
anything that changes the desire of firms to supply the commodity at a
pven market price. Such a change can result from a change in any of the

4
 98 THE ELEMENTARY THEORY OF PRICE

factors that affect supply other than market pncc The major possibihut
aie ^urtimansed rn Figure 8 2

erf
Quantity per period of time
e f

Quantity per period oftir

I 11

Fig 8 2
(i) An increase m supply - producen wish to make and sell more at each pr
This can be caused by
(1) Improvements in technology
(2) Decreases in the pnees of other commodities

(3) Decreases the pnees of factors of producuon used m making

m
commodity
(4) Some kinds of changes in the goals of producers
each pr
(ii) A decrease in the supply producers wish to make and sell less at

This can be caused by

(1) Loss m technical knowledge (unlikely)
(2) Increases in the prices of other commodities
(3) Increases in the pnees of factors of production used m making '

commodity
(4) Some kinds of changes m the goals of producers

I
 HAPTER 9

"HE ELEMENTARY THEORY

3F MARKET PRICE

-4\ this wc shall combine our theories of demand and supply into a
chapter
dieoiy of the determination of market prices. \Vc shall confine our attention
to competitive markets which arc markets in which there are so many
buyers and sellers that no one of them can by himself exert a significant
influence on what happens.

the DETERMINATION OF EQUILIBRIUM PRICE IN A

COMPETITIVE MARKET
The relation between q^uantitv demanded and supplied at
VARIOUS prices: What happens in the market for a particular commodity,
say, asparagus?
Goods are sold by firms and bought by households and mc
have already developed in the last two chapters an elementary^ theory of
supply and demand. now bring these two theories together and use
We
them to
develop our theory of market price.
In Figure 9.1, we plot the price of asparagus, on the verdcal axis, and we
draw both a demand and a supply cur\'e on the same diagram. The market
e
demand curve shows the amount of asparagus that households woul
to buy at each market price, assuming that populadon, incomes, tastes a
that qua
all other prices
remain constant; and its downward slope shows ^
demanded and price are assumed to vary inversely with each o er
wi ^
supply curve shows the amount of asparagus that producer
grow and sell at each market price, assuming that their goa ,

other
technology, the prices of factors of produedon and the pnees °
e ^ ourv'e
commodities remain unchanged. The upward slope o
ass
shows that the quantity supplied and the market price are
directly with each other.^ do«-n»-ard
1 tVe shall assume throughoutchapter that market-dema
this
^ ^ vertical supph’
^nd that market-supply curves slope upward. Limiting cases.
 THE ELEMENTARY THEORY OF PRICE
Look at the point at which the two curves intersect This point corre-
sponds to a market price of Is per pound Tlie amount
demanded is 1 000
units and tlic amount supplied is also 1,000 units Thus, at
the pnee oHi
the amount consumers wish to buy u exactly the
same as the amount pro’
ducers wish to sell Provided that the demand curve
slopes downward and
the supply curve slopes upward throughout their entire
ranges, there will be
only one price, Is in this case, at which the quantity
demanded is equal to
the quantity supplied
Now consider any price higher than Is, say At this price, consumers
ivish to buy 400 units, whereas producers wish to sell 1,450 units The

Quantity per period of time

Ftg 9 1 Determination of equilibnum pnee and quantity m a compeltuve market

quantity supplied exceeds the quantity demanded by 1 ,050 units It is easy

to sec, and you should check a few examples to make sure that you do, that
for any price above Is quantity supplied exceeds quantity demanded Such
situations are described as having excess supply The higher the price,
the greater the excess supply
Noiv consider prices below 7s, say 2i At this pnee consumers desire to
purchase 1,850 units and producers desire to sell 200 units There is an
excess demand causing a shortage of 1,650 units of asparagus Again it is clear,

and again you should check one or two examples, that at all pnees below
7 j the quantity demanded exceeds the quantity supplied Such situations
arc desenbed as having excess deuano The lower the pnee, the larger
the amount of excess demand
and a honzontal demand curve are sometimes of intemt They require some modification of
the statementsmade hereafter but the reader should have no difficulty m making the appro
pnaie modifications when they are required
 THE ELEMENTARY THEORY OF MARKET PRICE 101

Price changes when demand does not eq,ual supply: VVc now intro-
duce the hypothesis that, when there is excess supply, the market price will
fall. Producers, unable to sell some of tlicir goods, may begin to ask lower
prices for them; purchasers, obsen'ing the glut of unsold eommodities, may
begin to ofTer lower prices. For cither or both of these reasons prices will fall.
This hypothesis is illustrated in Figure 9.1 by the arrow indicating a down-
ward pressure on price at all prices above Is.
We also introduce the hypothesis that, when the quantity demanded ex-
ceeds the quantity supplied, market price will rise. Individual purchasers,
unable to fulfil all their requirements, may begin to offer higher prices in
an effort to get more of the available goods, and suppliers, able to dispose of
more than their total production, may begin to ask higher prices for the
quantities that they have produced. For cither or both of these reasons prices
nill rise tvhen demand exceeds supply. This hypothesis is illustrated in
Figure 9. 1 by the arrow indicating an upward pressure on price for all prices
below Is.

The equilibrium price: For any price above Is, according to our theory,
the price tends to fall, and any price below Is,
for rise. At
the price tends to
a price of Is, there is neither an excess of quantity demanded creating a
shortage nor an excess of quantity supplied creating a glut quantity supplied ;

is equal to quantity demanded and there is no tendency for the price to

change. The where the two curves intersect, is the price that
price of Is,

equates quantity demanded and the quantity supplied; it is the price to-
ward which the market price gravitates; and it is the only price at which
there is neither a shortage nor a surplus. This price is called the equi-
librium price. The term equilibrium means a state ofbalance; according
to our theory, such an equilibrium occurs when demanders desire to buy
the same amount as suppliers desire to sell. Since there is neither excess
supply nor excess demand there is no cause for price to change.
When demand equals supply we say that the market is in a state of
equilibrium. When demand does not equal supply we say that the market
is in a state of disequilibrium.

Summary. We may now summarise our simple theory of price as follows:

Hypotheses
1 Demand curves slope downward continuously.
2 Supply curves slope upward continuously.
3 An excess of the quantity demanded over the
quantity supplied causes price to rise an excess o ;

quantity supplied over quantity demanded c-

price to fall.
 ,

102 THE ELEMENTARY THEORY OF PRICE

Implications
1 There is no more than one price at which quantity
demanded equals quantity supplied. In the language
of economic theory, equiltbrium is unique.
2 If either the demand or supply curve shifts, the
equilibrium price and quantity will change. (The
actual changes are considered below.)‘

SHIFTS IN DEMAND AND SUPPLY

We may now ask what is the effect on the equilibrium price and quantity
bought and sold of shifts in the demand and supply curves ’ In Figure 9 2
the originaldemand curve is D, and the supply curve is Si Now assume
that thedemand curve shifts to Dj This increase in demand might, for
example, be the result of a nse in incomes The original equilibnum price
IS Oe and the quantity Oa When the demand curve shifts, excess demand
develops because, at pnee Oe, the quantity demanded is now Oc, whereas
the quantity supplied remains at Oa As a result of the excess demand, oc,

pnee will nse toward the new equilibnum price of 0/ At this price the
quantity demanded equals the quantity supplied The new equilibnum
quantity bought and sold is Ob, the nse m
pnee from Oe to 0/ reduces the
quantity demanded from Oc to Ob, whereas it increases the quantity sup*
plied from Oa to Ob

1 We conclude that a nse in the demand for a com*

modity (i.e., a rightward shift of the demand curve)

causes an increase m
both the equilibrium price
and the equilibrium quantity bought and sold.

When the demand decreases (iwhen the demand curve shifts to the left)
c ,

there will be a decrease both in the equibbnum pnee and m the equilibrium
quantity bought and sold This can ako be seen in Figure 9 2 if u c visualise
a shift m
the demand curve from to Di Equilibnum pnee decreases
from 0/ to Oe and equibbnum quantity from Ob to Oa

2 We conclude that a fall m

the demand in a com-
modity (i.e., a leftward shift of the demand curve)

causes a decrease in both the equilibnum price and

the equilibrium quantity bought and sold.
1 ume u was thought that the following inference could be drawn from these
For a long
hypotheses the market will be tlaiU in the seme that, if the pnee moves
away from lU equi
libnum it will move back toward and will
level, eventually return to the equilibnum level
of lubility is dealt with in
This inference cannot be drawn from this theory The problem
Chapter 12
 A FORMAL ANALYSIS OF ELASTICITY 123

point on the curve at which we made compare the elasticities of the two
our measurement. curves at that price. Since the curves
Figure 10.7 shows a straight-line are parallel, the ratio AqjAp is the same
demand curve by way of illustration. If on both curves. Since we are comparing
we wish to measure the elasticity at elasticities at the same price on both
point 1, we take our p and q at that curt'cs p is the same and the only factor
3
point and then consider a price change, left to vary is q. On the curve farther
taking us, say, to point 2, and we from the origin quantity is larger (i.e.,
measure our Ap and Aq as indicated. Oqz >Oqi) and hence pjq is smaller and
Now, the slope of the straight line join- thus e is smaller.
ing points 1 ApjAq (if you have
and 2 is It follows from Theorem 2 that paral-
forgotten this, refer to the Appendix to lel shifts of a straight-line demand curve
Chapter 2, pages 38-40), and the term lower elasticity (at each price) when
in equation (1) is AqjAp, which is the the line shiftsoutward, and raise
reciprocal of ApjAq. We conclude, elasticity when the line shifts inward.
therefore, that the first term our
in The elasticities of two intersecting
elasticityformula is the reciprocal of
the slope of the straight line joining the
two price-quantity positions under
consideration.
We now develop a number of theo-
rems relating to straight-line demand
curves.
1 The elasticity of a downward-sloping
straight-line demand curve varies from in-
finity (oo) at the price axis to zero at the
quantity axis. We first notice that a
straight line has a constant slope so that
the ratio ApjAq is the same anywhere
on the line. Therefore, its reciprocal, straight-line demand curves can be compared
AqjAp, must also be constant. We can at the point of intersection merely by com-
now infer the changes in e by inspecting paring slopes, the steeper curve being the less
the ratio pjq. Where the line cuts the elastic. In Figure 10.9 we have two
price axis, quantity is zero so the ratio intersecting curves. At the point of
pjq is infinity, thus e=cx).^ As we move intersectionp and q are common to
down the line,
p falls and q rises both curves and hence the ratio pjq is
steadily, thus
pjq is falling steadily so the same. Therefore e varies only with
that e is also falling. At the
q axis the AqjAp-, on the steeper curve ApjAq is
price is zero so the ratio pjq is zero. larger than on the flatter curve, thus
Thus €=0. the ratio AqjAp is smaller on the steeper
2 Comparing two straight-line demand
curvesof the same slope, the one farther from 1 This requires the dubious operation of divi-
the origin is less elastic at each price than we
ding by zero. If we wished to be more formal
could define elasticity at the intercept on
the one closer to the origin. Figure 10.8 the

shows two parallel straight-line demand price axis as the limit approached by
elasticity

functions. Pick any price, say Op, and as q-*0.

124 THE ELEMENTARY THEORY OF PRICE
4
curve than on the flatter curve, so that 5
which, by substitution from (2), gives
elasticity is loiver
Any straight-hne supply curve through
the origtn has an elasticity of unity Such a
supply curve is shown in Figure 10 10
IVttk a straight line demand curve, the
Consider the two triangles with the elasticitymeasured from any point
and p, q, ac-
sides p, q the S curve, and Ap, Aq cording to equation
(1) above, is independent
of the direction and magnitude of the change
in prue and quantity This follows im
mediately from the fact that the slope
of a straight line is a constant If we
start from some point p, q, and then
change pnce, the ratio AqjAp will be
the same whatever the direction or the
6
size of the change in
p

Quantity

Fti 10 9

t Quantity

Fig 10 11

The Theorem 5 does not hold

result in

for any demand function other than a

Fig 10 10
straight line Figure 10 II shows a de
mand curve that is not a straight line
and the 5 curve Clearly these are simi- We desire to measure the elasticity
lar mangles Therefore the ratios of from point 1 Whatever changes we
their sides are equal, i e ">ake. the ratio pjq is given but the
,
ratio AqjAp will vary according to the
/oi sac and the direction of the price
q Aq change We have alread) seen that
AqjAp is the reaprocal of the slope of
Elasticity of supply is defined as
the hne joining the two points con-
sidered If we lower pnce by a large
*
?!
amount so that we move from 1 to 3,
 A FORMAL ANALYSIS OF ELASTICITY 125

the ratio ^qj^p is the reciprocal of in fact the differential-calculus concept

slope of the line joining 1 and 3. If we of the derivative of quantity with re-
make the price cut smaller, so that we spect to price.
go from 1 to 2, the slope of the line is This elasticity is the one normally
larger than that joining 1 and 3, so that used in economic theory. Equation (1)
its reciprocal will be smaller. Thus our may be regarded as an approximation
measured c from the point p, q will be to this expression. It is obvious by in-
smaller if we make a small price cut specting Figure 10.12 that the elasticity
than if we make a large one. By a repe- measured from ( 1 ) will come closer and
dtion of the same argument we can closer to that measured from (5) the
show that, if we we
raise price, so that smaller the price change used to calcu-
move from 1 we will get a smaller
to 4, late the value of (1). In (1), change the
elasticity than when we lowered price, price so that we move from a to some
whereas, if we move from to 5, the1 point b; the ratio AqjAp is the recipro-
measured be even smaller
elasticity will
that it was when we moved from 1 to 4.
Thus our elasticity measured from
point 1 toU vary depending on the
direction and the magnitude of the
price changes which we happen to
make. This result is very inconvenient.
The reason for it is that, when we take
a big Ap, we are averaging the reaction
of Aq to Ap over a whole range of the
demand curve, and, depending on the
range that we take, the average reaction
will be different.
If we wish to measure the elasticity
at a point we need to know the reaction
of quantity to a change in price at that cal of the slope of the straight line
point, not over a whole range. call We joining a and b. The smaller the price
the reaction of quantity to price change change that we make, the closer point
at a point dqjdp and we define this to b comes to point a. The closer b comes
be the reciprocal of the slope of the to a, the closer the slope of the line
straight line (i.e., AqjAp) tangent to the joining a and b comes to the slope of the
demand curve at the point in question. line tangential to the curve at a; to see
In Figure 10.12 the elasticity ofdemand this, compare the slopes of the lines
at a is the ratio
pfq (as it has been in joining a and b' and a and b" with the
allprevious measures) now multiplied slope of the tangent T
in Figure 10.12.
by the ratio AqjAp measured along the If the slopes of these two lines get closer
straight-line tangent to the curve at a. together so also do the reciprocals of
This definition may now be written the slopes and, thus, so do the elastici-
ties measured by and (5). Thus, if
(1)
we consider (1) ss an approximation to
diminish as the size
(5), the error will
The ratio dqjdp as we have defined it is of Ap diminishes.
 CHAPTER 11

SOME PREDICTIONS OF THE

THEORY OF PRICE
In this chapter wc shall study the theory of pncc that we have developed
over the last five chapters to see if it can be made to yield useful predictions
about actual behaviour m the world When developing our formal theory
we concentrated first on the determination of equilibrium prices, and,
second, on the effect on equilibrium prices of vanous shifts in demand and
supply curves The propositions that we have denved Chapters 9 and 10m
and others that we shall derive in this chapter can be viewed m two different
ways First, they may be considered merely as logical implications of our
theory of price For example, the proposition that an increase in demand
raises the equihbnum pnee and quantity traded is a logical deduction from
our theory and, unless generations of economists have all made the same
we must accept it as being incontrovertible From the
gross error of logic,
second point of view, however, these propositions may be regarded as pre
dictions about what will happen m
the world under certain stated circum-
stances From this second point of view the correctness of our theory is an
empmcal matter
In order to make contact between our theory and the real world, wc
advance three hypotheses (1) that the assumptions of our theory (eg,
about the shapes of demand and supply curves) adequately describe rela-
some curve shifts, the actual
tions that exist in the real world, (2) that unless
price m
the real world stays at or near the equilibnum pnee, and (3) that,
if there is a change in the equihbnum pncc, the actual pnee wi/1 move
fairly quickly towards the new equihbnum If these hypotheses are correct
then the propositions of our theory will provide useful predictions about how
real pricesand quantities will actually behave under certain conditions, if
contradicted by the
not, the predictions of our theory will frequently be
*
evidence
I The $tudeni should now re read two earlier passages
on page ^
14 It IS of course quite
a) The section on ‘True in Thewy and not in Practice*
but that the
possible that a proposition foHows logically from the postulates of a theory
 SOME PREDICTIONS OF THE THEORY OF PRICES 127

The simple theorv’ of the

determination of market prices through demand
and supply is extremely powerful and can be applied to a large number of
real-world situations. The cases studied in this chapter are examples in-
tended both to illustrate how demand andsupply theory' can be used, and
to give the reader practice in using it. It is a mistake to try' to commit these

particular cases to memory'; the reader should seek to understand them, so

that when either slightly' different or totally new situations turn up (as they
always do) he will be able to apply' the theory to them and discover what is
predicted about each situation that arises.

PRICE AND WAGE CONTROLS

1 Maximum Price Legislation

It is very' common in wartime, and not unknown in peacetime, for govern-

ments to pass laws fixing the maximum prices at which commodities may be
sold. In this section we shall confine ourselves to the setting of maximum
prices (sometimes called ceiling prices) while in the next section we shall go
on to consider the effects of setting minimum prices. What would you pre-
dict to be the consequences of a system in which some prices are not left
completely free to be determined by demand and supply ?
Figure 11.1 shows the demand and supply curves for some commodity.
The equilibrium price is Oa and the equilibrium quantity is Ob. We now
wish to predict what would happen if the central authorites were able to
enforce a maximum price at which this commodity was to be sold. If the
maximum price were set above the equilibrium price, the intervention would
have no effect. The equilibrium price would still be attainable and the
market equilibrium would be in no rvay inconsistent yvith the maximum
price law. On the other hand, suppose the maximum price is set at a level
below the equilibrium one, say at Oc. The equilibrium price would no
longer be legally obtainable. Prices must be reduced from Oa to Oc, and as
a result the quantity demanded will expand by bd, from Ob to Od', the
quantity' supplied, on the other hand, will fall by be, from Ob to Oe. Thus a

proposition is refuted by experience. We then conclude that the tvhole thcor)' is a bad one and
that at least one of its assumptions about behatiour must incorrectly describe
what actually
happens.
b) The section of supply and demand on pages 104 and 105.
on the ‘laws’
Of course, given
the theory, certain conclusions about equilibrium prices necessarily follow. But this does nothing
to establish whether or not these conclusions allow us to predict correcdy what will happen in

the tvorld. The ability of a theoiy to predict is a question that can only be answered on
empirical grounds - i.e., by actually looking to see if what does happen is what the theorj’
predicts will happen.
 )

128 THE ELEMENTARY THEORY OF PRICE

shortage of the commodity wtU develop,
the quantity demanded exceeding
the quantity supplied (In Figure 11
I the excess demi ind IS
equal
We now have our fint predicuons about the effect of price to ed
control m a
competitive market

The setting of maximum prices will either

have no
(maximuin price setatorabove the equilibrium)
effect
or It will cause a shortage of the commodity
and
reduce both the price and the quantity actually
bought
and sold below tbeir equilibrium values

In the case of effective price ceilings production

is not sufficient to satisfy
everyone who wishes to buy the commodily Price is not allowed to change

SO as to allocate the available supply amongst the would be purchasers (sec

Chapter 5 on how the free market docs this) from which it follows that some
other method of allocation will have to be adopted Our theory does not
predict what this other method will be, but it is not difficult to enumerate
those alternatives which experience has shown to be likely to arise If shops

sell their available supplies to the first customers that arrive, then people
are likely to rush to those stores which are rumoured to have supplies of any
commodity for which there is a severe shortage, long queues will develop
and on the basis of being lucky or knowledgable enough
allocation will be
to gain from the principle of first come fint served * Another system may
develop if shopkeepers themselves decide who will get the scarce com
modities and who will not Goods may be kept under the counter and sold
only to certain customers The shopkeeper might sell only to regular

1 In wartime Europw the rumour that tome shop was selling supplies of some very scarce
commodity was sufficient to cause a local stampede Housewives often spent days tracking

down such rumours and then hours standu^ m line before being able to gam entrance to a

shop Usually the supplies would be exhatistcd wl^ many housewives remained unserved
 SOME PREDICTIONS OF THE THEORY
OF PRICES 129
customers who bought a wide range of goods.' The
storekeeper might sell
only to people of a co our or religion
or with other affiliations of
which h
approved. All kinds of rules could be
adopted by the storekeeper and
may be given the general name ‘allocation the^
by sellers’ preferences’.
If the central authonties dislike
the somewhat arbitrary system of alloca-
Uon that grows up, they can ration the
goods, giving out ration coupons
sufficient to purchase the
quantity Oe in Figure 11.1. The authorities can
en determine, as a conscious act
of policy, how the available supply
is to
be allocated; the coupons might be
distributed equally amongst the popula-
tion, or they might be
distributed on the basis of age, se.x,
marital status,
number of dependants, or any other criterion
that the authorities wish to
adopt, thus we are led to
predict the following;

Where there is a feeling against allocation on the

basis
of firstcome first served and of sellers’ preferences,
effective price ceilings will give rise
to strong pressure
for a centrally administered system
of rationing.

Next we observe that, under certain circumstances,

price control with or
without rationing is likely to
give rise to a black market.^ For many
pro ucte there are only
a few large producers but veiy' many retailers and,
although it is easy to police
the producers, it is difficult even to locate all
t ose who are,
or could be, retailing the product, much less to police them.
t ough the Central Authorities may be able
to control effectively the
pnce that producers get for their
product, they may not be able to control
e ectively the
price at which retailers sell to the public, and they certainly
cannot control the sale of
ration coupons by those who prefer money to the
rationed good. What would
you predict would happen in this case.^ First,
e amount produced
would remain unchanged at Oe in Figure 1 1 because 1 .

t e producer
would continue to receive the controlled price for his product.
t the retail
level, however, a black market would arise, because purchasers
would be willing to pay very
much more than the controlled price for the
imited amounts of the
commodity that were available. If the whole quantity
were sold on the black
market, it would fetch a price of Of per unit. The
total amount paid by consumers would be Oegf', of this the total amount of
the illegal receipts
of black marketeers would be chgf. The theory predicts

I In wartime Britain,
^
to move from one town to another meant losing one’s status as a
regular in many shops. Unless one was a long-term regular at some shop it was very difficult
indeed to obtain
cigarettes or beer, both of which were subject to price control while being
unrationed.
2 A black market’is one in which goods are sold illegally at prices which violate the legal
restnctions (either
above a legal maximum price or below a legal minimum).

5
130 THE ELEMENTARY THEORY OF PRICE
lhat the potential for a profitable black market will always exist whenever
efiective price ceilings are imposed The actual growth of such a market
depends on there being a few people willing to nsk heavy penalties by
running a black*market supply orgamsalion and a reasonably large number
of persons prepared to purchase goods illegally on such a market It is an
interesting comment on the strengths of various human motives that there
has never been a case documented in which efleclivc price ceilings were not
accompanied by the growth of a black market
It IS unlikely that all goods will be sold on the black market both because

there are some honest people in every society and because the central
authorities always have some power to enforce their pnee laws Thus we
would normally expect not the extreme result given above but rather that
some of the limited supplies would be sold at the controlled price and onI>
some would be sold at the black market price ‘
An economist’s evaluation ofa black-market situation can only be made
when u is known what objectives the central autlionties were hoping to
achieve with their pnce-control policy If they are mainly concerned mth
an equitable division of a scarce product, it is vcr> likely that cfTectivc pnee
control on manufacturers plus a Iargel> uncontrolled black market at the
retail level produces the wont possible results If, however, they are mainly
interested m restricting the total $uppl> available for consumption in order
to release resources for other more urgent needs, such as war production,
the policy works effectively if somewhat unfairly Where the purpose is to
keep pnees down, the policy is a failure to the degree that the black market
succeeds in raising prices and a success to the extent to which transactions
do take place at controlled pnees
There is much evidence confirming these predictions which we have
shown to follow from our simple theory of price Practically all belligerent
countnes m both the First and the Second World Wan introduced schemes
setting ceilings on pnecs well below frec-market, cquilibnum levels These
schemes were always followed by shortages, then by either the introduction
of rationing or the grow th of some method such as allocation by sellcn’
preferences, and finally by the n<e of some sort of a black market These
stheiwts weTe nrote tffeettve hv Iwuvtiwg cowvvimptioa tlian in controlling
pnees, although they did restrain pnee increases to some considerable
extent Peacetime price controls have been less frequent and have probably
been less successful in obtaining their major objectives Many countnes have
tried to control rentals of houses and apartments for pnvate use Such rent

the fraction of the

1 The itudent might try to demoiistialc for hnmelT that the higher i>
will be the black market pnee
available supply that is sold at the coniroOed pnee ihe higher

remainder This is not however, an easy proposiuon to

demonstrate formally
for the
 SOME PREDICTIONS OF THE THEORY OF PRICES 131

control schemes have usually produced the same results : a shortage, private
allocation systems and a black market.

2 Minimum Price Legislation

Governments sometimes pass laws stating that certain goods and services
cannot be sold below some stated minimum price. In many Western
countries today there are minimum wage laws which specify ‘flows’ for the
wages to be paid to different kinds of labour. Resale price maintenance,
which exists in many countries, gives the manufacturer power to prevent
the retailer from selling below prices set by the manufacturer. Before read-
ing on, the student should ask himself what our theory predicts about the
effects of minimum price laws.

Fig 1 1.2 Minimum price controls.

The case of a commodity subject to minimum price legislation is illus-

trated in Figure 11.2. The free-market equilibrium price is Oa, and the
equilibrium quantity traded is Ob. If the minimum price is set below the
equilibrium price then it has no eflfect on the market. The attainment of the
free-market equilibrium and the fulfilment of the minimum price law are
perfectly compatible. On the other hand, if the minimum price is set above
the equilibrium, say at Oe, the free-market equilibrium will be legally un-
obtainable. The and at that price there will be an
actual price will be Oe,
excess of supply over demand. Suppliers would like to sell Om, but pur-
chasers are only willing to buy Oj at the price Oe. The actual amount
bought and sold will thus be Oj and there will be excess supply ofjm. This
leads us to our first prediction about minimum prices.

1 For example, a grossly inflated sum may be paid for a few shoddy bits of furniture. In this

case, the landlord is receiving the difference between the controlled rent and the free market
one as a lump sum payment at the beginning of the tenancy.
 132 THE ELEMENTARY THEORY OF PRICE
The setting of ntintmum
prices will either have no
effect (minimum price set below the equilibrium) or
it of the commodity to develop
will cause a surplus
with the actual price being above its equilibrium level
but the actual quantity bought and sold being below
its equilibrium level.

In is, at the legally enforced price, no scarcit) of the con-

this case there
trolled commodity Therefore we do not predict that alternative allocative
systems will grow up There will, however, be a shortage of purchasen and
potential suppliers may compete m various ways for the available customcn
Methods of pnee cutting will be searched for, some of which find loop-holes
in the law and some of which merely flout it For example, clubs and other
organisations have grown rapidly in order to take advantage of cheap trade
rates which the airlines arc not allowed to offer the single passenger There
will be no opportunity for a set of black-market operators to take over the
distnbution of the product since there is nothing to be gained by buying at

the controlled price and selling at the frec-market pnee There will of
course be an incentive for an individual producer to sell his product at less
than the controlled price as long as hi$ alternativeis not to sell it at all Thus

in this case we predict the absence of an organised black market but the
existence of some clandestine selling by individual producers at pnees below
the legal minimum
As an example of minimum pnee pohaes consider the case of minimum
wage laws which are found m
most Western countries Using our theory to
apply to the labour market » a bit of a jump m
the dark but we can note in
passing that the theory developed in Part V does allow us to use a down-
ward-sloping demand curve for each type of labour If wc hypothesize that
our theory of competitive markets will apply to a labour market, we have
the following predictions about minimum wage laws when applied m only
a few markets of the economy

1 Where the law is effective it will raise the wages of some of those who
remain in employment
2 It will lower the actual amount of employment (by bj in Figure 11 2)
3 It will create a surplus of labour which would like to but cannot obtain
jobs in the occupation affected {jm in Figure 1 1 2)
4 It will create an incentive for some workers to try to
evade the law by
offering to work at wages below the legal one
5 It will not lead to the rise of an independent group of
black marketeers

who buy at the controlled pnee and sell at a black market one
There is ample evidence confirming most of these predictions The illegal
 SOME PREDICTIONS OF THE THEORY OF PRICES 133

offering of their services for part-time and evening work is a well established
reaction of many workers to union regulations of agreed wages. The em-
pirical validity of prediction (2) is a matter of dispute when minimum wage
laws are applied across the whole economy. We cannot go into this con-
troversy here but we shall raise it again in Chapter 34.
It is remarkable how many predictions our simple theory yields about the
effects of price control. It is also interesting to note that these predictions
have been shown to be accurate in such a large number of cases. It is also
interesting, and not a little depressing, to see how often governments are
prepared to pass price-control laws without showing any apparent apprecia-
tion of the likely effects of such measures.

A Digression on Methods of Allocating Scarce Commodities

In the previous sections we methods of

raised the question of alternative
allocating scarce commodities amongst potential consumers. Since it is
almost always true that people would like to have more of a commodity
than is in fact available, it is necessary to have some way of rationing the
available supply. In a dictatorship this may be done by the central authori-
ties. In a free-market society this is done by the price mechanism. When

there is excess demand, price rises and this encourages production and dis-
courages consumption. Price continues to rise until, at its equilibrium level,

the rate of consumption is equal to the rate of production. Thus market

price does the rationing. If price is held constant at a level below the
equilibrium one, the available quantity must be allocated in some other
way by storekeepers on the basis of sellers’ preferences, by queues on the
:

basis of the earliest arrivals, or by government rationing, with coupons

distributed on any one of a number of principles.
Because goods are scarce, there must be some system of allocating them
amongst potential consumers. It is sometimes argued that the price system
provides the best (in an ethical sense) way of doing this. Certainly there is
nothing in positive economics that can prove that the price system as it
functions in the real world is ethically the best method of allocating
scarce goods. Positive economics attempts to show the consequences of
allocating such goods by various methods. Any decision about what method
ought to be adopted will be a better-informed one if taken in the light of
knowledge of the actual effects of different methods.
It is sometimes said that the price that equates demand and supply, Oa
in Figures 11.1 and 11.2, is the natural price, while other prices are artificial
ones. This is very emotive language which is likely to give the impression
that the natural price is in some sense the best price. All that can be said
on the basis of positive economics at this stage is that the price Oa is the
 *34 the elementary theory op price
one that equates tfie quantity demanded to the quantity suppficd through
the mtehanvTn oj pnet If other pnets are enforced, alternative
methods of
equaling demand to supply will have to be employed

TAX incidence
What IS the effect of lavea placed on the sale and purchase of commodities,
such as the excise tax on jewellery, or petrol or whisky?
Do such faxes leave
prices unchanged or do they cause prices to nsc? Docs the producer
pay the
tax or IS he able to pass it on to the comutner through
higher pnees ? Many
such age-old controversies are to be found the field of tax theory m
As a fint step in discos enng whit our theory predicu about these
issues,
wc must consider the effect ofa tax on the jupplj ofa commodity Gsnsidcr
the example tn Table 1 1 1 which shows the supply schedule for an imaginary
commodity

Table 1 1 I

C<ilumK Co/omn 2 Cofumn 3 Column 4

1
j

If fhesuppljff receive the pnee if a tax of Is t$ If a tax equal to

listed below he will offer for sale placed on the sale 25 per cent of the
the quantity listed below of this commodity, price obtained by
« must be sold for seller is levied on
the price listed in this commodity it

cofumn if the must be loW for the

producer is to tt- pnee listed m col-
tape the amount umn (4) if the pro-
listed in column (1) 1
ducer is to rtctive the
pnee listed m col-

1
umn (1)

tons JuppUed
pnee per toa per tnonlk per ton
h
\

1
0 —
500 3s 2s 6d
3j 1,000 4s 3s 9d
4s 1,500 '

5t 5s Od
1

6j 2,500 7s 7s 6d
j

lOr 1
4,500 llj 123 6(1
1

The supply schedule indicated by columns I and 2 is graphed m Figure 1 1 3,

and is labelled S The schedule shows the relation between the pnee that the
 :

SOME PREDICTIONS OF THE THEORY OF PRICES 135

producer obtains for his commodity, and the amount that he is willing to
If no tax is levied, then the seller receives the whole market price for
sell.

which the commodity is sold. If, however, a tax is levied on the sale of a
commodity, then the seller will receive on each unit that he sells the market
price of the commodity minus the amount of the tax. In order that he should
receive the same amount per unit as he was receiving prior to the tax, the market price
must be raised by the full amount of the tax. This is illustrated in column 3 of
Table 11.1 with the example of a specific tax of U per unit. If the producer
is prepared to sell 500 tons when he himself obtains 25 per unit, then

500 tons will be supplied at a market price of 25 when there is no tax, but
at a market price of 35 when the tax is levied. If the producer is prepared

Fig 11.3 The effect of a tax on the

supply curve of a competitive
industry'.

when he obtains 65 per unit, then 2,500 tons will be supplied

to sell 2,500 tons
at a market price of 65 when there is no tax, and at a market price of Is
when the tax is levied. Thus, assuming that the seller’s tvillingness to supply
the commodity in response to his own personal remuneration is unchanged,
then, after a tax has been levied, every quantity offered will be associated
with a market price higher by the full amount of the tax than the one
previously required. This shift in the relation between market price and
quantity supplied is illustrated in column 3 of Table 11.1 and by the supply
curve Si in Figure 11.3.

The on a commodity, therefore, is to

effect of a tax
shift every point on the supply curve vertically up-
wards by the amount of the tax.

In this example we have used a fixed tax of I5 per unit. Precisely the same
conclusion follows if we examine a tax that is a fixed percentage of the value
136 THE ELEMENTARY THEORY OF PRICE
the supply cur\e shifts vertically upwards by the full amount of the tax
Column 4 of Table 1 1 1 provides the data and the student should plot for
hvmsclC the supply cvitve after ttxes (For the remainder of this section we
shall forbrcvit) deal only with the specific tax of a certain sum on each
unit sold but the student should repeal the argument and draw the graphs
for the other case as well
\Vc saw m Chapter ^ that pnee depends on demand as well as on supply
Before we can say anything about the price we must consider the demand
curve This is done in Figure 11 4 The original equilibnum price is 5s
wliile the quantity traded is 2 000 units If,following the imposition of the
tax the price rises h\ the full amount of the tax from 5s to 6s then the

Fig U4 The effect of a lax on

the pnee of a commodity

Quantity

quantity demanded would fall and there would be an excess of supply over

demand This would cause the price to fall until it reaches the equilibrium
point at which the new supply curve cuts the original demand curve In the
example of Figure 4 the new equilibrium price is 5s Bd This is the price
1 1

that will be paid by consumers while when the tax of Is per umt is deducted
the producers will receive a unit pnee of 4s Bd Thus in this example the
tax has the effect of raising thepnee paid by consumers by Bd and lowering
the price received by producers by 4</ The incidence of the tax in this
case IS said to fall two thirds on the consumer and one third on the pro
ducer The term incidence is used to desenbe who pays the taxes levied
the
The following general prediction can easily be shown to follow from
theory by replacing the numbers of the previous specif c example with
the

letters used m general reasoning

As long as the demand curve slopes downwards and
the supply curve upwards, the imposition of a tax will
 :

SOME PREDICTIONS OF THE THEORY OF PRICES 137

raise the price paid by consumers and lower the price

received by producers in both cases by an amount
less than the amount of the tax.

The influence of elasticity of demand and supply. Figure 11.5

repeats the supply curves of Figure 11.3, but combines them with two
different demand curves representing two extreme cases: a perfectly in-
elastic demand curve in Figure 11.5(i) and a perfectly elastic curve in
Figure can be seen that, in the case of the perfectly inelastic
11.5(ii). It

curve, the equilibrium price increases by the full amount of the tax, while,

0 2.000 4,000 0 2,000 4,000 0 2000 4,000

Quantity Quantity Quantity

ii iii

Fig 11.5 The effect of a tax on price and quantity given demand curves
of various elasticities.

in the case of the perfectly elastic curve, the equilibrium price is unchanged
in spite of the shift in supply. This suggests the following general predictions

1 The more inelastic is the demand for a commodity,

the greater will be the rise in the price paid by the
consumer and the less the fall in the price received
by the producer as a result of the imposition of any
given tax.

2 The more inelastic is the demand for a commodity,

the less will be the fall in quantity bought and sold
resulting from the imposition of the tax.

The derivation of this prediction is given in Figure 1 1.5 (iii). Consider a de-
mand curve, D, intersecting the original supply curve, S, at an equilibrium
price of 5s. Note the post-tax equilibrium price given by the intersection of

5*
138 THE ELEMENTARY THEORY OF PRICE

the original demand curve with the new supply curve S, Now consider
piloting the demand curve through the original equilibrium point as shown
in Figure 1 1 5(iii) Clearly, the steeper, and thus the more inelastic, is the
demand curve, the greater is the nse in price paid by consumers and the
smaller is the fall in the price received by producers
We have relied on graphical dialysis to derive the above predictions, but
the argument can be stated verbally without too much difficulty The case
of the completely inelastic demand curve means that consumers insist on
buying the same quantity of the commodity whatever the market price If
the pnee rises by the full amount of the tax, there will still be no change in
the quantity demanded, and, since there will also be no change in the
quantity supplied (because the pnee received by producers will be exactly
the same as it was before the tax was levied), then the price that equates
demand and supply will be higher than the original one by the full amount
of the tax In this extreme case the incidence of the tax falls entirely on
consumers
The case of the perfectly elastic demand curve means that consumers will
purchase as much of the commodity as they can obtain at the going price
but will purchase nothing at all at any higher price Thus, any increase in
price will reduce demand to zero The only possible market price u the
original one, and, providing some producers are willing to sell goods at that
price, then the original price must also be the new equilibrium one In this
case the incidence of the tax falls entirely on producers
Empirical evidence relating to these predictions is not easy to obtain If
prices usually stayed constant then all we would have to do would be to

observe price changes m competitive markets after the raising or lowering

of tax ratesUnfortunately (for the purpose of testing these predictions),
however, prices are constantly changing because of continual shifts in
demand and supply due to the growth of real incomes, changes in tech-
nology, changes m tastes, inflations and deflationsand a host of other
factors To discover, in such a changing situation, what is the influence of
taxes or prices a very difficult job of economic measurement and not one
is

which can be accomplished by casual observation

It IS impossible here to summarise such evidence as does exist and we
can
only assert the conclusion which would, we believe, command general even
ifnot universal, assent In the case of those raw materials and agricultural
products which are sold on a competitive market and m
the case of property

for rental, it is probably the case that tlic burden is shared

between the
consumers and producers and that the division of the burden is roughly
what one would expect on the basis of die elasticities of demand and supply
competitive
In the case of most manufactured goods which are not sold on
markets (see Chapter 26) it is less certam how the burden is shared and
it is
 SOME PREDICTIONS OF THE THEORY OF PRICES 139

at least possible that taxes are passed on completely to consumers through

higher prices.^

THE PROBLEMS OF AGRICULTURE

The ‘agricultural problem’ is one of the most perplexing problems facing
policy makers in Western countries. In many countries the average level of
income in agriculture is substantially lower than what it is in other sections
of the economy. In Britain, for example, farmers provide about 1 -5 per cent
of the population and 3-5 per cent of the labour force, but they earn only
2-5 per cent of the total income of the country. In many countries problems
of overproduction loom large, agricultural supply shows a more or less con-
tinual tendency to outrun demand and thus there is a tendency for free-
market prices to be depressed to the disadvantage of farmers. Whenever the
central authorities intervene to help farmers by supporting prices, output
soars and the granaries and store-houses of the country begin to bulge with
unsold agricultural surpluses. Such is the complexity of interventions in
agriculture, and the belief that a free-market would produce very different
results that would be disastrous to at least some farmers, that the six
countries of the European Economic Community (France, Germany, Italy,
Holland, Belgium and Luxembourg) took many years of hard negotiations
to obtain a measure of agreement on harmonising agricultural policies,
which they had relatively little difficulty in doing so with industrial policies
and in agreeing to complete free trade in manufactured goods by 1970.
Indeed at the time of writing (February 1966) the agricultural problem
threatens to damage seriously, possibly even to destroy, the Community.^

1 Agricultural Price Fluctuations

The production of many agricultural goods is subject to quite large varia-

tions due to factors completely beyond human control lack of rainfall,
: in-

vasion of pests, floods and other natural causes are capable of reducing out-
put to a level well below that planned by farmers, while exceptionally
favourable conditions can cause production to be well above the planned
level. We may now ask what oiir theory predicts about the effect of these
unplanned fluctuations on the price of agricultural commodities and on the
revenues earned by farmers for the sale of their crops.
1 Either because producers are not trying to maximise profits or because long-run supply
curves are perfectly elastic.

2 The whole problem is given added importance by the fact that in most countries farmers

have a degree of political influence out of proportion to their actual numbers in the population.
Most rural constituencies have a much smaller population than most urban ones. Therefore it
takes fewer farm votes to eleet a member of parliament than it does urban votes.
140 THE ELEMENTARY THEORY OF PRICE

A supply curve is meant to show desired output and sales at each market
price unplanned vanations in output then actual production
If there are
and diverge from their planned level The supply curve drawn m
sales will
Figure 1 1shows the total quanuty farmers desire to produce and offer for
6,
sale at vanous pnces If the pnce were Oa, then planned production would
be Ob, but actual production would vary around this planned amount,
owing to causes bcjond the farmers’ control Two demand curves are drawn
in Figure II 6 one is relatively elastic and the other is relativel> inelastic
over the price range from Of to 0/* In a world m
which plans were alwa)s
fulfilled,price would settle at the equilibrium level of Oa with output Ob
But unplanned fluctuations m
output will cause the actual price to fluctuate
If, for example, the crop is poor so that the actual production is Oc, then a

Fig H
6 Variations in pnee caused
by unplanned vanations m
supply
operating on elastic and unelastic
demand curves

shortage will develop, pnces wiHnse to Ot »n the case of demand curve D„

and Og in the case of curve In each case the quantity demanded will be
reduced to a point at which it is equal to the available supply If, on the
other hand, conditions are particularly favourable, actual production will
exceed planned production, a surplus will occur and pnce will fall For
example, when production is Od, price falls to Of in the case oF curve D,
and to Oh m
the case of curve Z), In each case the fall in pnce is sufficient
to increase the quantity demanded sulflciently to absorb the extra un-

planned supply but the fall in pnce is large in the case of curve D, and small
in the case of curve Z?*
We have now derived the followmg prediction about the effect of un

planned fluctuations supplym

Unplanned fluctuations m
supply will cause price
variations in the opposite direction to the supply
changes (the bigger the supply the lower the price)
and, for given supply fluctuations, the price fluctua-
tions will be larger the lower is the elasticity of
demand for the product-
 SOME PREDICTIONS OF THE THEORY OF PRICES 141

We have relied on a geometrical analysis in deriving the above prediction,

but does have a strong commonsense appeal. High elasticity means that
it

the quantity demanded is very sensitive to price changes. When there is a

crop failure, a small increase in price will be sufficient to choke off the
quantity demanded until it is equal to the greatly reduced quantity
supplied when there is a bumper crop, a small fall in price will be sufficient
;

to call forth the additional demand necessary to buy up the larger supply.
Low on the other hand, means that the quantity demanded is not
elasticity,

ver)' responsive to price changes. People insist on trying to purchase roughly

the same amount of the commodity and, when there is a crop failure, a very
large increase in price is necessary to reduce the quantity demanded to the
level of the available supply. On the other hand, when there is a bumper

crop, a very large price reduction is necessary to persuade consumers to

purchase the whole available supply.
Now consider the effects on the revenues received by farmers from the
sale of their crops.* Here the relations sound a bit more complex, but they
all follow immediately from the results established on page 140. If the good
in^question has an elasticity of demand greater than unity, then unplanned
increases in supply raise farmers’ revenues while unplanned decreases lower
them. If the demand for the product is inelastic, consumers’ total expendi-
ture on the product, and thus farmers’ revenues will rise when price rises
and fall when price falls. Thus, good harvests will bring reductions in total
farm revenues while bad harvests will bring increases in farm revenues!^
The reason for this is that when demand is inelastic, a given percentage
change in the harvest will cause a much greater percentage change in price.
To illustrate, a production might cause a 50 per cent rise
20 per cent fall in

in price, in farm revenue would rise by 20 per cent, while

which case total
a 20 per cent rise in production might cause a 40 per cent fall in price, in
which case total farm receipts would fall by 28 per cent.
If the elasticity of demand happened to be unity then farmers’ revenues
would not vary as output and prices varied because every change in output
would be met by an exactly compensating change in price so that total
expenditure would remain constant.
We now have the following predictions:
1 Unplanned fluctuations in output can cause every
conceivable type of fluctuation in farmers’ revenue.
1 While we can make predictions in this section about the revenues, such receipts are
closely related to the incomes of farmers. We can without risk of serious error extend these

predictions to incomes.
2 This is only true if demand is inelastic over the relevant range. It does not follow that every
individual farmer’s income must rise (after all, some farmers may have nothing to harvest), it

only follows that the total revenues earned by all farmers must rise.
 142 THE ELEMENTARY THEORY OF PRICE
2 Farm revenue and farm output will vary in the
same direction whenever demand for the product
IS elastic
3 Farm revenue and output will vary in the opposite
directions whenever demand for the product is
inelastic
4 The fluctuations in revenue will be larger the
further does the elasticity of demand for the pro-
duct diverge from unity in either direction

Evidence on these predictions is fairly abundant Unplanned fluctuations

in supply do occur frequently in agriculture Where the prices of such goods
are left to be determined by the free market, large price fluctuations do

occur In the case of many agricultural goods the demand is quite inelastic
In these cases we find very large price fluctuaUons together with the peculiar
situation that when nature is unexpectedly kind and produces a bumper
crop farmers see their incomes dwindling while when nature is moderately
unkind so that supplies fall unexpectedly larmers incomes nse The self
interests of the farmer and of the consumer appear to be exactly opposed
in such cases

2 Agricultural Stabilisation Programmes

In free market economies, agricultural incomes often tend to fluctuate
around a low averagt level Agricultural stabilisation programmes have two
goals to reduce the fluctuations and to raise the level of farm incomes Most
countries now operate some sort of scheme to reduce agricultural fluctua
tions The two policies, producing stable incomes and producing reasonably
high incomes, can, as wc shall see, often conflict
Figure 11 7 shows the demand and supply curves for some agncuhural
product The supply curve shows planned production at each pnee and if
production could be planned with certainty, pnee would settle down at the
equilibrium level of On Although planned production is at Oi actual pro
duction fluctuates around that level, say between the quantities Oc and Od
In a free market, prices will thus fluctuate between Oe and 0/
m
One method of preventing these fluctuations prices and incomes is for
the individual farmers to form a producers association which tries to even
out the supply actually coming on to the market m
spite of vanations in

production ^ There is no point in an individual farmer holding some of his

are concerned here only with action to stabilite pnees After
we have studied the
1 We
producers organisa
theory of perfect competition in Chapter 22 we shall return to the topic of
and study their attempts to raise the average level of their members incomes At
this
tions
time we shall develop a theory of the inherent insUbihty of producers cooperatives
 SOME PREDICTIONS OF THE THEORY OF PRICES 143

production off the market in an effort to force up the price. Since one
farmer’s production is a completely insignificant part of total production,
the farmer who sold less would only reduce his income without having any
noticeable effect on price. But farmers get together and agree to vary
if all
the supply coming on to the market, then, collectively, they can have a
major effect on price.
Under the conditions illustrated by Figure 11.7, a producers’ association'^
might be quite successful in keeping the price at Oa and incomes at the
level indicated by the area of the rectangle Oaxb. What would the associa-
tion’s policy have to be ? Any excess of production over Ob would have to
be stored away unsold. If, for example, production for one year were Od,
then bd would have to be added to the association’s stocks while Ob was sold
at the price Oa. Any deficiency of production below Ob, on the other hand,
would have to be made good by sales out of the association’s stocks. If pro-
duction were Oc, for example, then cb would be sold out of stocks, making

Fig 11.7 Government policies

designed to stabilise price in
the face of unplanned
fluctuations in supply.

total sales again equal to Ob at a price of Oa. In this way the producers’
association could keep sales, priceand incomes stabilised in spite of fluctua-
tions in production. Provided that the level of sales to be maintained (Ob in
Figure 11.6) is equal to average production, then the policy could be carried
on indefinitely. If, on the other hand, an attempt were made to keep the
price too high, so that sales were less than the average amount produced,
then, over a number of years, additions to stocks held would tend to in-
crease. The successful policy is the one that keeps sales constant at Ob (by
adding to, or subtracting from stocks) and, since income accrues to the pro-
ducers when the goods are actually sold on the market, revenues ^vill be
stabilised at Oaxb.
In fact both the supply and the demand curves will be shifting to the
right over time, the supply curve because of improvements in technology
which lowers farmers’ costs of production, and the demand curve because
 144 THE ELEMENTARY THEORY OF PRICE
of the general rise m household income due
to economic growth The *

agricultural sectors of most Western economies tend toshow rapid techno-

logical change but
from low income elasticities of demand, so that
to suffer
generally the supply curve will be shifting to the right faster than
the
demand curve In these circumstances the equilibrium price will be falling^
and It takes great foresight, self-control and discipline to stabihse prices
‘about a falling average level rather than a constant one
What will happen if a producers' association is not formed or is not suc-
cessful but the central authorities attempt to stabilise the incomes of farmers
by entering the market themselves, buying m the open market and adding
to their own stocks when there is a surplus, and selling in the open market,
thus reducing their stocks, when there is a shortage? If the central authori
ties wish to stabilise farmers’ incomes, what policy should they adopt ? Should

they aim, like the producers' association, at keeping prices constant at all
times? Before reading further, the student should attempt to work out for
himself the consequences of a government policy designed to keep price
fixed at the level Oa in Figure 1 1 7 by buying goods when production is in
excess of Ob and by selling goods when production falls short of Ob The
central authorities are assumed not to consume any of the commodity but
only to hold stocks, thus all their purchases are added to their own stocks
and all of their sales are made out of these stocks
If the average level of production around which the year-to year figure
fluctuates is is no reason why the authorities should not
Ob, then there
successfully stabilise the price at Oa indefinitely This policy would not,
however, have the of stabilising farmers incomes Farmers will now
result
be faced with an infinitely clastic demand at thepnee Oa whatever the total
quantity produced, they will be able to sell it at the price Oa if the public
will not buy all the production, then the authorities will purchase what is
leftover If total production is Od, then Ob will be bought by the public and
bd by the authorities to add to their own stocks Total farm income in this
case will be the amount indicated by the area of the rectangle Oayd (the
quantity Od multiplied by the pnee Oa) If total production another year m
ISonly Oc, then this quantity will be sold by farmers and the central authori-
will remain at Oa Total
ties also will sell cb out of their stocks so ihat price

farm income will then be the amount indicated by the rectangle Oaze
(quantity Oc multiplied by the pnee Oa) U is obvious that if prices are held

page 90 for » duciuHon of the effecu of changes in income on the demand

1 See Chapter 7,

2 In making this prediction we are assuming all other prices to be constant Ifwe place it

into the more usual context of a general inflationary nsc in all

puces then the prediction is that
nsmg more slowly than the average level of all prices Sec Chapter
14
agricultural prices will be
on the significance of relauve putts
 SOME PREDICTIONS OF THE THEORY OF PRICES 145

constant and farmers sell their whole production each year, then farmers’
incomes will fluctuate in proportion to fluctuations in production. This
government policy therefore will not eliminate income fluctuations but it
will rev'erse their direction. Now bumper crops will be associated with high
incomes while small crops will be associated with low incomes.
What, then, must the policy of the central authorities be if they wish to
stabilise farmers’ incomes through their own purchases and sales in the open
market? Too much price stability causes incomes to vary directly with pro-
duction, as in the case just considered, while too little price stability causes
incomes vary inversely with production as in the free market case origin-
to
ally considered. It appears that the authorities should aim at some inter-
mediate degree of price stability. If they allow prices to vary exactly in
proportion to variations in production, then incomes will be stabilised. A
10 per cent rise in production should be met by a 10 per cent fall in price,
and a 10 per cent fall in production by a 10 per cent rise in price.
The policy necessary to achieve the requisite price fluctuations is de-
scribed in Figure 1 The analysis becomes a bit
1.8. difficult at this stage and
the student should draw his own graph, building it up
by step as the
step
argument proceeds, using Figure 11.8 merely as a guide. D
and S should
first be copied from Figure 11.8. As before, planned production is Ob, and
the price which equates demand with planned production is Oa. Actual
production, however, fluctuates between Oc and Od, and these fluctuations,
given the very inelastic demand curve D, could cause price to fluctuate
between Or and Oi.
Now construct through the point x (the equilibrium point when actual
production is equal to planned production) a curve of unit elasticity through-

out whole range. This constructed curve is the dotted rectangular

its

hyperbola in Figure 1 1.8, labelled 17= 1. If Oi is produced and sold at price

Oa, total income is that indicated by the rectangle Oaxb. The dotted curve
now tells what must happen to the market price if production and sales are
allowed to vary but income is to be held constant at Oaxb.
Consider first what happens if production is Od. Market price must be
held at Oj { = dn) if income is to be unchanged. But, at market price Oj, the
public only wishes to purchase Oq ( =jp) and it is therefore necessary for the
authorities to buy up the remaining production, qd {=pn) and add it to their
stocks. Farmers’ total sales are Od at price Oj, and, since the dotted curve is

a rectangular hyperbola, it follows that income Ojnd is equal income Oaxb.

to

Now consider what must happen if production is equal to Oc. If farm

income is to be unchanged, then the price must be allowed to rise to Ot (by
construction the area of rectangle Otke is equal to the area of rectangle
Oaxb). But at the price Oi, the public will wish to buy Om {
= ie) so that the

central authorities must sell cm (=ke) out of their stocks.

 THE ELEMENTARY THEORY OF PRICE

Fig II 8 Government policies designed to stabilise income

in the face of unplanned fluctuations in supply
 SOME PREDICTIONS OF THE THEORY OF PRICES 147

If this policy is it will have the following results. First, there

successful,
will be smaller fluctuations in the price of this product than there would be
if price w'ere determined on a completely free market. Secondly, total in-

come of the producers will be stabilised in the face of fluctuations in pro-

duction. Finally, the stabilisation scheme should be self-financing. In fact,
if we ignore costs of storage, the scheme will show a profit, for the authorities

willbe bu)Tng at low prices (below Oa) - the lower the price the more they
buy - and they will be selling at high prices (above Oa) ~ the higher the
price the more they sell. Whether or not the scheme actually shows a profit
wall depend on the costs of storing the crops from the periods of glut when
they are purchased until the periods of shortage when they are sold. In any
case, this scheme has the financial advantage over the previous one in which
the authorities completely stabilised prices, because in that case there
would necessarily be a loss since all purchases and sales would be made at
;

the same price, Oa, there would be no trading profit to set against the costs
of storage.

Problems with stabilisation policies. The above analysis is meant

merely to illustrate the many types of schemes that could be operated and
to show how, once the details of the schemes are specified, our theory of
price can be used to predict the consequences of each scheme. If such
schemes have all of the advantages outlined above, why is it that there is so
much trouble with most government stabilisation schemes, which are actu-
ally operated ? One of the major problems with these schemes arises from
the absence of perfect knowledge, combined with political pressure applied
by farmers. Demand and supply curves are never known exactly, so the
central authorities do not know what will be the average production forth-
coming over a number of years at various prices. The central authorities do
not, therefore, know exactly what level of income they can try to achieve
while also keeping sales from stocks approximately equal to purchases for
stocks, over a large number of years. Since farmers have votes there is strong
pressure on any government to err in the direction of fixing the income to be
stabilised at too high a level. If the level of income, and hence price, is fixed
too high, then the central authorities will find it necessary to buy unsold
crops most of the years and will find only a few years when production is
low enough for sales to be made out of stocks. In this case, stocks will build
up more or less continuously, and the time will come when no more can be
stored. When this happens, the stored crops will either have to be destroyed,
given away or dumped on the market for what they will bring, thus forcing
the market price down to a very low level. If the crops are thrown on the
market and allowed to depress the price, then the original purpose for which
the crops were purchased, price stabilisation, is defeated. If the crops are
148 THE ELEMENTARY THEORY OF PRICE
destroyed or allowed to decay, then this means that the efforts of a large
quantity of the country’s scarce fectors of production (the land, labour and
capital that went into producing the stored goods) have been completely
wasted ‘ Furthermore, the authonties’ plan will now show a deficit, for
goods will have been purchased which cannot be sold at all This deficit will
have to be made up by taxation which means that people in cities will be
paying farmers for producing goods which arc never consumed
When schemes get into this sort of difficulty, the next step is often to tiy
to limit the production of each farmer Quotas may be assigned to individual
farmers and penalties imposed for exceeding the quotas Or else, as has been
done many times in the past, bonuses may be paid for leaving land idle, for
ploughing crops under without harvesting them, or for other means of
cutting back on production Such measures attempt to get around the
problem that too many resources are allocated to the agricultural sector by
the device of preventing these resources from producing all that they arc
capable of producing The interesting question of the morality of such
measures takes us outside the scope of positive economics which docs not, of
course, mean that the student should not pursue the question

Resources in Agriculture The Long Period Problem

Even if the temptation to set too high a price is avoided there is still a
formidable problem waiting to wreck many agricultural stabilisation pro-
grammes This problem results from the fact that the productive capacity
of almost all economics is growing over time In the United Kingdom, the
actual increase in production has averaged almost 2 per cent per year over
the last 100 years The rise m
production has been the result of the increased
productivity of the working force, owing to better health, better working
conditions, and more and better capital equipment Workers in most sectors

of the economy can produce more per head than they previously did If the
allocation of resources were to remain unchanged, there would be an in-
crease the production and hence in the supply of each commodity
m in

proportion to the increase in productivity in that industry If the pro-

the
ductivity increases are spread generally throughout the economy, then
increases in actual production will also be generally spread
throughout the

economy
The real incomes of the population will also increase, on the average at a

rate equal to the production increase How will the people wish to consume
their extra income^ The relevant measure m this case is the income

much
1 To say nothing of the moral problems involved in so much waste in the face of so
with hungry mouths
poverty throughout the world If they arc given away or sold to countnes
they often cause a political storm
 SOME PREDICTIONS OF THE THEORY OF PRICES 149

of demand, which measures the effect of increases in income on

elasticity
the demands for various goods. If, to take the simplest case, all goods have
unit-income elasticities of demand, then the proportion in which the various
goods are demanded will not change with income, and an x per cent rise in
income will lead to an x per cent rise in the demand for ever>' good. It is
known, however, that income elasticities vary considerably among goods,
and that most goods tend to have different income elasticities at various
levels of income. For example, at the level of income achieved in advanced
industrialised countries, many
have very low income
foodstuffs elasticities,
and many manufactured goods have high income elasticities.
Assume that productivity e.xpands more or less uniformly in all indus-
triesthe demands for goods with low income elasticities will be expanding
:

slower than the supplies excess supplies will develop, prices and profits will
;

be depressed, and it will be necessary for resources to move out of these in-
dustries. Exactly the reverse will happen for goods with high income
elasticities: demand will expand faster than supply, prices and profits will
tend to and resources will move into the industries producing these
rise,

goods. Table .2 illustrates the point just made. It gives a simple numerical
1 1

example of an economy divided into an agricultural and a manufacturing

sector. Originally, resources are divided equally between the two sectors.
Productivity then doubles in both sectors. The incomes of all consumers
double and the income demand for manufactured goods is
elasticity of
higher than the income elasticity of demand for agricultural goods. The rise

Table 11.2

Agrimliure Manufacturing

Production originally was 50 50

Production after productivity
change, if there were no re-
allocation of resources, would
be 100 100
Income demand is
elasticity of •50 1-5

Therefore quantity demanded

after rise in income is 75 125
Therefore surplus or shortage is 25 (surplus) 25 (shortage)

in productivity causes a surplus equal to one-quarter of the agricultural

production, and a shortage equal to one-quarter of the manufactured-goods
production. Thus it will be necessary for resources to move out of the

agricultural and into the manufacturing industries. Furthermore, if the

productivity increases are going on continuously, there will be a continual

 150 THE ELEMENTARY THEORY OF PRICE
tendency to\Nard excess supply ofagncultuntl goods and excess demand for
manufactured goods
In a free market economy this re allocation will take
place under the
incentives of low prices wages and incomes m
the declining sector and
high prices wages and incomes in the expanding sector Look
at Table 1 1 2
again Because supply exceeds demand m
the agricultural sector prices will
fall and incomes of producen vsill fall Because too much
is being produced
there will be a decline in the demand for farm labour and the other facton
of production used in agriculture and the earnings of these factors will
decline At thesame time exactly the opposite tendencies will be observed
in manufacturing Here demand is expanding faster than supply prices will
rise incomes and profits of producen will be rising there will be a large
demand for the factors of production used in manufacturing industries so
that the price of these factors and consequently the incomes that they cam
will be bid upward In short manufacturing will be a buoyant expanding
industry and agriculture will be a depressed and contracting industry

In a free market society, the mechanism for a con

tmued re allocation of resources out of low elasticity
industries into high*elasticity ones is a continued
depressing tendency on prices and incomes in con
tracting industries and a continued buoyant tendency
on prices and incomes in expanding industries

Now what is the effect that all this will have on the kind of government
stabilisation policy considered earlier^ Generali) there are two motives

behind agricultural stabilisation policies first to secure a stable leiel of

income and second to provide a htgh level of income Frequently in a
wealthy community where real incomes are expanding year by year many
feel that the agricultural iccxor aught to share in this prosperity Stabilisation
programmes often aim at providing the farmer with an income on a panty
with incomes earned m the urban sector of the economy
Positive economics has nothing to say about the ethics of such a policy

it merely tries to discover its consequences The mam problem is that a

programme that succeeds in giving the rural sector a high level of income
may frustrate this re allocation mechanism and unless some other means of
between
reducing the size of the rural sector is found the discrepancy
supply will conunuc to grow until it reaches
unmanageable
demand and
dimensions If productivity continues to increase m
the rural sector while
the excess of supply
income elasticity of demand for its products is low then
government
over demand will get larger and larger as time passes If the
insists on trying to maintain agncultural prices
and incomes it will find
 SOME PREDICTIONS OF THE THEORY OF PRICES 151

that, as time goes by, it is necessary to purchase ever-larger surpluses. If

incomes are guaranteed, there will be no monetar)' incentives for resources
to transfer out of the agricultural sector. Unless some other meansis found

to persuade resources to transfer, then a larger and larger proportion of the

resources in the industry will become redundant. If, on the other hand, the
government does not intervene at all, leaving the price mechanism to
accomplish the resource re-allocation, be faced with the problem of
it will
a more or less permanently depressed sector of the community. The govern-
ment may not be willing to accept all of the social consequences of leaving
this sector to fend for itself.

The reader should not jump to the conclusion that economics proves that
governments ought not to interfere %vith the price mechanism because the
risks Such a conclusion cannot be proved’, it is 3. judgment, which
are too large.
depends on a valuation of the gains, losses and risks of such intervention.
Positive economics, by providing some insight into the workings of the price
mechanism, can be used to predict some of the consequences of such inter-
vention and thus to point out problems that must be solved in some way or
another if the intervention is to be successful. If the problem of re-allocating
resources out of the rural sector is not solved, then intervention to secure
high and stable levels of farm incomes will be unsuccessful over any long
period of time.
 CHAPTER 12

THE ELEMENTARY DYNAMIC

THEORY OF PRICE

It takes five years for a newly planted rubber tree to reach

maturity and to begin to yield latex It tikes five months for a newly
hatched chicken to reach matuniy and to begin to produce eggs It takes
time before the production of an> commodity can be increased, and the
time required varies greatly from one commodity to another Does it
matter’
The thing to notice is that there is a gap in time between the de-
first

cision to produce more of a commodity and the actual increase in produc-

tion The length of this gap will depend on a number of facton, but the two
most important are the extent to which there is any productive capacity
that IS not being used at the moment, and the time that it takes to expand
productive capacity If the owner of a rubber plantation wishes to produce
more rubber in response, say, to a nsc in the market pnee of rubber, he will
be able to increase actual production fairly rapidly if there exist stocks of
mature trees not now being tapped for raw rubber But once such stocks of
unused capacity are utilized (and ofien they will not exist at all) then it will
take at least five years before the desire to change production is translated
into an actualit) The gap between a change in the desire to produce goods
and a change m
actual production is called a supply lag
Every commodity has its own charactenstic, and often quite complex,
supply lag To take one further example, an increase in the demand for
raw milk can be met to some extent almost immediately by diverting milk
from other uses, to a greater extent within 27 months by not slaughtenng
calves at birth and waiting until they reach matunty to produce milk and
to an cver-increasing extent as the laiger population of adult cows give
birth to a larger number of calves First, therefore, producers divert existing
supplies from other uses, second they lower the slaughter rate among newly
 THE ELEMENTARY DYNAMIC THEORY OF PRICE 153

born calves and finally they are able to increase the total number of births
once there has been an increase in the total number of mature cows.

STATIC AND DYNAMIC ANALYSIS

Before we enquire further into the importance of the supply lag,
we should
note that in everything we have done in previous chapters the lag in supply
has been ignored. We must now look more critically at the method that
has been employed in previous chapters. The theory' of taxes provides an
example of this method and you should now refer back to Figure 1 1 .4. A tax
on a commodity shifts the supply curve vertically upwards by the amount
of the tax. The new equilibrium price is above the old one, but not by the
full amount of the tax. This theor)' thus produces the predictions that the

effect of a tax is to raise the unit price paid by consumers by less than the
full amount of the tax, and to lower the unit price received by producers,
again by less than the full amount of the tax.
These conclusions refer to a comparison of the new equilibrium position,
y, with the original equilibrium position, x. A moment’s reflection will re-
veal that this is the method of analysis used throughout all of the previous
chapters. We wish to form an hypothesis about the effect of some change in
the data, possibly the introduction of a tax or a change in the conditions
of demand. We start from a position of equilibrium and then introduce the
change to be studied. The new equilibrium position is determined and it is
compared with the original one. The differences between the two positions
of equilibrium must be due to the changes in the data that were introduced.
This analysis, which is based on a comparison of two positions of equilibrium,
is called comparative static EquiLiBRiUM analysis. This is a rather

cumbersome expression, and it is usually abbreviated to comparative

STATICS.
Theories based upon comparative statics will be of most interest if our
concern is to predict where a market will settle down after all of the effects
of some change have been worked out. Since, in an ever-changing world,
such a position of complete rest will never be observable it might appear
that comparative statics is useless, but this is not the case. Static theory
makes such predictions as ‘in response to an increase in demand price and
quantity will rise’. These predictions will pass test as long as price and
quantity move in the direction of their equilibrium values, in what has been
called ‘pursuit of equilibrium’, even though the actual equilibrium position
may never be reached. The above example concerns a prediction about be-
haviour over time. We also have predictions about how various parts of the
economy will compare at a moment of time. Comparative statics does thus
154 THE ELEMENTARY THEORY OF PRICE
have a substantial range of usefulness * What it can never do,
however, is
to give us predictions about how the vanables will
behave as they move
from one position of equiUbnum toward another We might ivish to
predict,
for example, how price and output of rubber would change
from ytar to year
in response to a change in demand m
a situation m which the final equi
librium could not be reached until at least 15 years after the original
de-
mand shift occurred Such predictions cannot be derived from static theories
In some cases it is even possible that the market will never settle down m a
position of equilibrium, and, in such extreme cases, the predictions derived
from static theory are very likely to be contradicted by the actual behaviour
of the market
In summary we cansay that although theones based on the technique
of comparative static analysis are adequate for dealing with many problems,
they cannot be used to handle two important classes of problems First, they
cannot be used to predict the path that the market will follow w hen moving
from one equilibrium to another, and second they cannot predict whether
or not a given equilibnum position will ever be attained Indeed when an
equilibrium is not obtained, predictions based on the assumption that it will
be are likely to turn out to be empirically false
For purposes of studying behaviour in situations other than equilibrium,
we use DYNAMIC ANALYSIS which may be defined as the study of the behaviour
of systems, single markets or whole economies in disequilibrium situations

MARKET FLUCTUATIONS
In Chapter 5 we discussed in an intuitive fashion the effect of an increase in

the demand for Brussels sprouts In Chapter 9 wc were able to formalise

some of this discussion in terms of the theory of equilibrium price You
should re-read the discussion on page 105 before reading on Wc were
now
not able at that time to formalise the discussion of the movement between
the initial and the final equilibrium because we did not have a dynamic
theory Let us consider this problem in more detail with reference to
Figure 12 1

The Jise m
the demand for Brussels sprouts raises the equilibnum price
from Oe to Ov and the equilibnum quantity bought and sold from Oa to Of
We can expect supply to react to any pnee change only after a considerable
lapse of Ume When the original increase m demand occurs it will not be
In
possible to produce any more sprouts unul another crop can be raised

argumenU cannot prove that predictions derived from sutic

I Of course such general
is only meant lo dispel the
belief thai il is
theones will pass test The discussion in the text

obvious that they won t pass test and hence are useless TTie actual value of such lUUC theones
depends on their ability to generale predictions which actually do pass lest
 THE ELEMENTARY DYNAMIC THEORY OF PRICE 155

the intervening period, the whole effect of the rise in demand will have
to
be taken out in a price increase. If supply continues to come on the market
at a rate of Oa while the demand rises from to Z)2, then the price mil
rise to Of. This is the price that equates the new demand to the unchanged
supply. If price reacts quicklyit may rise to Of within a week or two after
the rise in demand and stay there until more sprouts can be raised. But at
the price Of farmers would like to produce and sell Ob sprouts. They may
^vell plan to produce at that rate instead of at the equilibrium rate, Oc -

indeed, there is as yet no signal to tell them that Oc is the ‘correct’ rate at

which to produce. If, ivhen the newly planted sprouts reach maturity, the
rate of production suddenly expands to Ob, then price will fall drastically.

Indeed, the price that will clear the market is now Od. But this will lead
after a lag to a contraction in the quantity supplied. At this point we may
well begin to wonder whether the market ivill ever reach equilibrium.

THE COBWEB

We shall now introduce an elementary dynamic theory that accounts for

some of the aspects of behaviour discussed above. In this theory, we assume
that producers’ output plans are fulfilled, but with a time lag, and we show
how planned changes in supply can give rise to oscillations in market
behaviour.^
All supply decisions take time to implement, so that supply coming on to

1 In Chapter 1 1 ,
we applied comparative-static analysis to a very simple case of agricultural-
price fluctuations. In that example we assumed that planned supply did not change and that
the price-fluctuations were caused by unplanned exogenous changes in supply.
156 THE ELEMENTARY THEORY OF PRICE
the market at any one time is always the result of decisions made in the
past,
whereas decisions made about production in the present always have
their
effect on the actual supply coming forward to the market
only at some Ume
m the future In cases in which the time lag is short, it can often
be ignored
successfully, but m
many cases in which the lag is longer it is of critical
importance In order to study the ideas of dynamic theory, we shall intro-
duce only the simplest possible time lag, but even this will be quite sufficient
to upset and sometimes even to destroy the smooth working of
the market
adjustment mechanism This simple time lag is one in which this year’s
price has no effect whatsoever on this year’s supply, and in which the full
adjustment to this year’s price is made all at once next year ‘

Farmers look to the existing market price when deciding iihat crops to plant this year,
and thus next year s supply depends on this year s price and this year's supply depends
on last year's price *

Such lags are typical of agncuUural products, such as wheat, oats and
barley, that give one crop annually
A market subject tosuch a simple one*year time lag is illustrated in
Figure 12 2(i) The demand curve shows the relation between the price
ruling in any year and the quantit> that will be demanded in the same year,
the supply curve shows the relation between the price ruling any one m
year and the supply that will come on to the market in the following year
The pnee that equates demand and supply is Oa At this price, Ob units
will be produced and sold
What will happen if this equilibrium is disturbed ’ If in one year, year /,
the price is Oc, farmers will plan to produce Od in the following year In that
year (year t + Od will come on the market, and, in order that Od may be
1)

sold, the price must fall to Oe The pnee of Oe will induce farmers to produce
the quantity Of When this quantity comes on the market m
the following

year, year t+2, the price will nse to Og This price wll call forth a supply
of OA the next year, year /+3, and this will depress the price below Og It is
clear from this that, in the market desenbed by Figure 12 2(i}, the pnee
and
quantity will oscillate around their equilibnutn values in a senes of dimmish
mg fluctuaUons, so that, if nothing further disturbs the market, pnee and
quantity will eventually approach their equilibrium levels
Now consider the case illustrated in Figure 12 2(ii) Exactly the same

1 In this case we say the Ume lag i» a me period Ug ifu takes several periods (years m this

a single pricechange to occur we say the time lag is a

case) for the full supply reaction to
distributed one (i e , U is spread over lime)
(page 38). =!(/>, i). which is read
2 In the terminology of the Appendix to CHiapter 3
the previous time penod
supply at time period t depends on the pnee of the product ruling m
1—1, where time penods are measured in yean
 THE ELEMENTARY DYNAMIC THEORY OF PRICE 157

argument as the previous paragraph applies here, and the text of that para-
graph should be re-read to describe the process in this market. In this case,
the oscillations get larger and larger so that the equilibrium is never restored.
The market in Figure 12.2(i) has an adjustment mechanism which is
STABLE, while that of Figure 12.2(ii) has one which is unstable. A stable
equilibrium is one which will be restored if it is disturbed, thus the actual
price and quantity will tend towards their equilibrium levels (Figure 12.2(i))

Fig. 12.2 A stable and an unstable cobweb.

 '

TJIE ELEMENTABY THEORY OF PRICE

an umtiblc equilibnum is one which will not be restored if
it is disturbed
thus the actual pnce and quantity will tend away
from their equilibnum
levels (Figure 12 2(ii)} What is the diflcrencc between these two markets
which makes one stable and the other unstable’ The studei t should try to
answer this question for himselfbefore reading further
The difTerencc between the two markets is in the relation between the
slopes of the demand and
the supply curves In Figure 12 2{i) the demand
curve than the supply curve The absolute quantity demanded
is flatter

changes more as price changes than docs the absolute quantity supplied
Any excess demand or supply can be eliminated with only a small pnee
change and the price change in turn causes only a very small change in
supply in the following year and hence the supply change has only a small
effect on next year s price In Figure 12 2(ii) the supply curve is flatter than
the demand curve the quanuty supplied responds more to pnce changes
than does the quantity demanded When there is excess supply a large
pnce fall is necessary to call forth the required demand This pnce fall

causes a large reduction in next years supply (because supply is very re

sponsive to pnce) Next year there is a large shortage and a very big pnce
increase is necessary to reduce demand to the level of the available supply
This pnce nse causes a tery large increase in quantity supplied in the
follotving year and so we go on in a senes of altemaung periods of ever
increasing surplus and shortage
In the case of the unstable equilibnum the osciJlauons get bi^er and
bigger There is nothing iti our theory so far to prevent these osallations
from becoming infinitely large In practice hoxvever we should not expect
this to happen we should expect that the oscillations would tend to reach
limits A full theory of such a market would require an analysis of these
limits

What we have established, however, is that, in the

unstable case, the operation of the competitive pnce
system itself does not tend to remove any disequi
libnum, It tends rather to accentuate it

We shall call a system in which this happens an unstable system

This cobweb model a very simple one in which supply plans are always
is

fulfilled (actual quantity suppbed always equals

planned quantity suppbed)
planned supply one year depends solely on the price nihng in the previous
m
equate the quanuty currently
year and the market pnce » always such as to
It is evident that more
being demanded with that currently being supplied
however easily
1 Thsargumenm of course only an intu uve one MathemaUcal analys s
is steeper than the demand curve
shows that the market is in fact stable if the supply curve

and unstable if the supply curve s flatter than die demand curve
 THE ELEMENTARY DYNAMIC THEORY OF PRICE 159

interesting and complex models arise if: (i) we allow actual supply to
deviate from planned supply due to such uncontrollable factors as weather
conditions ;
(ii) we allow for a time lag in the adjustment of price to the state

of excess demand and (iii) we allow for some form of learning process on
;

the part of farmers so that planned supply does not depend solely on the
price of the previous period. Situations of this sortbecome quite complex
and they cannot be handled \sTthout the help of mathematical analysis.^
The study of the simplest cobweb model does introduce dynamic theoiyq
and illustrates its value by providing a reasonably satisfactory explanation
of an interesting real-world phenomenon. It also shows in a fairly dramatic
way that even very simple competitive markets may show unstable oscilla-
tory behaviour.^ We shall return to the problems of dynamics many times
throughout this book.

1 The interested student should consult Chapter 1 of R. G. D. Allen Mathematical Economics,

Macmillan (1966).
2 It was thought for a ver^' long time that the static theory of Chapter 9 summarised on
page 101 permitted the inference that the market would always be stable. (See note l,page 102.)
The present analysis shows why this inference cannot be drawn. That excess demand should
cause price to rise and excess supply should cause price to fall is not sufficient to ensure

stability. Such models as the one illustrated in Figure 12.2(ii) contain all the assumptions listed
in Chapter 9 yet they are unstable. In these models excess demand pushes the price up towards
equilibrium but it pushes it too far- excess supply pushes price down towards equilibrium but it

also pushes it too far. Thus an endless series of oscillations around the equilibrium is possible.
 CHAPTER 13

A POSTSCRIPT AND A PREVIEW

THE INTERRELATIONSHIPS OF MARKETS A POSTSCRIPT

In Part II we have developed a theory of the behaviour of individual, com*
petmve markets
The economy docs not, however, consist of a senes of self contained ’

markets functioning in isolation It is to be viewed, rather, as an interlocking

system m
which anything happening in one market has profound effects on
many other markets and could, potentially, influence every other market in
the economy Thus our study of the behaviour of one market in isolation is

only a step towards understanding the behaviour of a whole economy

first

We also need a theory of how the individual markets are linked together
and of how they act and react on each other We shall consider this, albeit
very briefly, in Part VI, but in the meantime it is desirable to obtain some
intuitive picture of the linking together of the various parts of the economy
Consider, for example, the effects of a rise in the demand for motor cars
Thi« will be metfairly soon by a nse in car production If the rise m demand

is and judged to be permanent, there will also be a planned

considerable,
increase in capacity in the car industry Employment will rise and an
attempt may be made to attract labour from elsewhere by ofTenng higher
earnings Thus, one of the first impacts on other industnes will be a loss of
labour and possibly a need to raise wages m order to compete with the car
industry for labour This may cause profits to fall m these other industnes
The employment m the car industry may occasion some geo
increased
graphical movement of labour In this case there would be a nse in the
demand for housing m the car centres and a corresponding fall in demand
elsewhere New housing construction in the car-producing areas would lead
to a rise m the demand for construction workers and materials Quarnes and
 A POSTSCRIPT AND A PREVIEW 161

brickworks may have to take on additional labour and expand output.

Further, there will be a rise in the demand for raw materials used in car
construction and the of this may be felt in such diverse spots as the
effects
glass-making areas of the Midlands, the steel-manufacturing sections of
Wales, and the rubber plantations of Malaya. If new investment in plant
and equipment takes place in the car industries, there will be a rise in the
demand for many and bottlenecks may develop,
capital goods; shortages
and other industries which use these goods may experience increases in their
costs and troubles with delivery dates. There will also be a change in con-
sumers’ expenditure because some people’s earnings will be increased and
other people’s reduced. Thus the effects of this one change will spread out
through the economy rather like the rippleswhich spread out over the
smooth surface of a pond after a pebble has been dropped into it.

THE PRICE SYSTEM IS A CONTROL MECHANISM

The example considered above illustrates two very important points. First
the various markets of the economy are interrelated a single initial change
:

in demand has numerous effects throughout the economy. Second, adapta-

tions to the initial shift take place without being consciously coordinated by
some single central authority. When shortages develop, prices rise and
profit-seeking entrepreneurs are led to produce more of the good in short
supply. When surpluses occur prices fall and supply is voluntarily con-
tracted. The price system produces a series of automatic signals so that a
large number of different decision-taking units (firms and households) do,
in fact, produce coordinated reactions to some change.

The price system was not consciously created. With

a price system it is not necessary to foresee and to
coordinate all necessary changes such changes occur
;

automatically as a result of the separate decisions

taken by a large number of individuals, each seeking
his own private profit, but all responding to changes
in demands and prices.

Having grasped this idea, the student must beware ofjumping to the con-
clusion that the price system has been shown to be the best system of regu-
lating the economy; one must beware of equating the word automatic,
which we have used, with the phrase ‘perfectly functioning which we have;

not used. It is easy enough to control the heat in your house by means of an
automatic thermostatic control, but it is equally easy to have such a badly
designed or imperfectly functioning system that the heat control you actually

6
 162 THE ELEMEHTARY THEORY OF PRICE
achieve worse than you would have had by ‘stoking up’ and
is
‘damping
down by hand To observe that the price system functions
automatically,
I e without conscious centrahsed co ordination does not
,
tell us how well
It functions We have seen, for example, in the case of the cobweb
(see
Chapter 12), how the automatic workmg of the price system can produce
violent fluctuations in pnee and output The theory of the price system is
the subject of a large part of this book, the problem of assessing
the efficiency
of this automatic system, particularly in comparison with the efficiency of
other systems, is a most important problem which belongs
to a later stage of
the study of economics
The student who believes that behaviour in a free market economy is un-
planned and uncoordinated must dispel this notion It was the great
discovery of the early eighteenth century economists that a competitive
price system produced a coordination of effort m
which by seeking their
private gams and responding
to such public signals as pnccs, costs and pro-
fit produced coordinated reactions to changes in demand
rales individuals
and supply The existence of a coordinating mechanism is beyond dispute
m
The question of how well it works comparison with alternative coordinat
mg systems has been a matter of dispute (br 200 years and is still a great un
settled social question

DEMAND AND SUPPLY A PREVIEW

In Part II we have developed a theory of the behaviour of competitive
markets, which are markets in which no single buyer or seller is important
enough to exert a significant effect on the behaviour of the market In order
to develop this theory of market behaviour we first introduced a theory of
households’ demands in Chapter 7 and then, m
Chapter 8, a theory of
firms’ supply In order to take our theory further we must now consider m
much more detail the theory of demand and the theory of supply Demand
IS considered in Part III and supply in Part IV
The theory of demand only
occupies us for two chapters because at the level of an introductory work
there is not very much demand theory that is required In particular we
never need to depart Irom our competitive assumption that eac/r Acrasei’nJitf
IS a price-taker, being totally unable to influence
by any action of its own
the market prices of the commodiues it purchases The theory of
supply

occupies us for amuch longer time This is because many interesting issues
of economic policy are encountered in the theory of supply, and
also be-
that a
cause very senous complications are encountered when it is realised
very large proportion of production is m
the hands of firms m
noncompeti-

tive situations situations are noncompetitive

These the m
sense that each

firm able to exert a significant influence on price and total quantity

bought
is
 A POSTSCRIPT AND A PREVIEW 163

and sold by altering its oi\ti price-output decisions. Indeed, although we do

not need to alter any of the hypotheses about demand introduced in Chap-
ter 7, itbecomes necessary, as a result of the existence of these noncompeti-
tive situations, to abandon the basic hypothesis of Chapter 8 that there
always exists a simple relation between market price and firms’ supply. But
first we must turn our attention to the theory of demand.
 PART 3

THE INTERMEDIATE
THEORY OF DEMAND
CHAPTER 14

SOME BASIC THEOREMS

AND PREDICTIONS

Is the effect on a household’s consumption different if its income rises than

if the prices it pays for goods fall ? Does an increase in the general level of

prices hurt everybody ? Does it hurt anybody ? Can we predict the effect of
changes in relative prices on households’ behaviour? To answer these and
other similar questions, we must go behind the demand curve, whose
determinants we studied in Chapter 7, to look in more detail at the be-
haviour of millions of independent decision units whose aggregate behaviour
is summarised in the market demand curve.

In this chapter we consider the different effects on a household’s con-

sumption of changes in relative prices, absolute prices, and incomes. In the
following chapter we show some alternative methods of deriving from
assumptions of household behaviour the basic prediction that a household’s
demand curve for a product slopes downwards. We have already dis-
cussed in Chapter 6 the general assumptions made about the household
which is the basic decision unit on the side of demand. The student should
now review pages 76-7 of that chapter to refresh his memory on these basic
assumptions.
Our main purpose in this chapter is to distinguish between the effects on
a households’ expenditures of changes in income, relative prices and absolute
prices. We shall consider a single household faced with the choice between
only two goods, X and T/ In Figure 14.1 the quantity of Xis measured on

1 As in the last section of Chapter 4, the choice between two goods is sufficient to display

most of the problems in which we are interested in this chapter. The argument can be general-
ised to any number of goods, but when this is done, graphical analysis can no longer be applied,
and some quite difficult mathematics is required. A numerical illustration is used in Figure 14.1
to make the argument easier to follow. The numbers of the specific illustrations are later sup-
pressed so that the argument can proceed in general terms.
 the intermediate tueory of demand
the horizontal axis and the quantity of } is
measured on the vertical axis
Any point on this graph represents a combination of the two goods
Point a
for example, represents 40 uniu of .V and 60 units of K

Ikl

pHmnH
ihki

Lwss:r,
inMHHnni
Tflummir-
Fig H / Shifls in llie budget line caused b> changes m household income

THE budget line

As a first step m
our analysis we construct a budget line, which shows all
those combinations of the goods thit arc just obtainable, given the house-
hold’s income and the prices of the two commodities Avsume, for example,
'

that the household’s income is per month, that the pnee of Y is £2

per unit, and the pnee of A i$ £4 What combinations of \ and T are open
to the household^ First, it could spend all its mc)ne> on F, obtaining each
month 601' and no A (this combination is indicated by point a on Figure
14 1) It could also buy 30A and no I 'point h on the figure) Other
combinations open to it arc 581 and lA, 561 and 2A, 541 and 3A, in fact,

J Thu budget line w analogous to the production possibility boundary shown m Figure 4 1

on page 66 The budget line show* the combinalions of goods available to one household
itl income while ihe carher curve shows the
given prices and combinations of goods avilable
to the whole society given its techniques of prodaciion and ns supplies of resources
 . F

SOME BASIC THEOREMS AND PREDICTIONS 169

since X
costs twice as much per unit as Y, the household must forego the
purchase of 27 in order to be able to purchase one more X. In the language
of Chapter 4, the opporlunity
cost of A' is 2Y. All the possible combinations of

X and Y open household are shown by the straight hne ab in Figure

to the
14.1. In constructing this budget line we ignore the effect of the household’s
stock of wealth on its purchases. Strictly, we are assuming that the house-
hold has only a negligible stock of such ivealth so that its consumption
decisions depend only on its current income and its prices. A correction is
necessary for households with large stocks of wealth but this is not of major

importance most households and need not detain us at this level of

for
analysis. The reader should pick a few points on this budget line and check
that they indicate combinations of X and Y that exactly exhaust the house-
hold’s income. A few points between ah and the origin (e.g., the point 20
and 10 A"’) should then be checked to see that they less than exhaust the
household’s income. Finally, he should check that points above the line ab
(e.g., 50X and 30 F) represent combinations for the two goods that cannot

be bought, given the existing prices and the household’s income.

It is assumed that these goods are the only ones available to the household
and that the household will be on its budget line and never inside it, be-
cause, unless the wants of its members are totally satiated, the members will
always prefer more goods to fewer goods. It is also assumed, for purposes of
*
the present argument, that there is no saving.

Changes in Income

We may now ask what happens budget line when income and prices
to the

change? If, for example, the household’s income is halved from £\2{} to
per month, prices being unchanged, then the amount of goods the
household can buy will also be halved. If it spends all its income on F, it
will now get 30 F and no X (point c ) if it spends all its income on X, it will
;

get 15 A’^ and no F (point d) All possible combinations open to the household
appear on budget line cd. Note that this line is parallel to budget line ab,
but closer to the origin. Since prices are unchanged, the household must still
give up 2F for every additional X it wishes to purchase. The slope of the
line indicates this opportunity cost, for it shows how much F must be given
up another unit of X. The fact that ab and cd both have the same
to get
slope indicates that the opportunity cost of X
in terms of F is the same in

both situations.
If the household’s income rises to ;^180, it will be able to buy more of
both commodities than it could previously. If it buys only F it can now
1 Saving could be made a variable in which case it would appear explicitly on one of the
axes.

6*
170 THE INTERMEDIATE THEORY OF DEMAND
have 90J (poini #} >f it buys only A it can have 45A (point/) ifii divides
Its income equally between the two goods it can have 45} and
22iV
(point All the combinations of X and 1 now available to the household
appeir on budget line ef
m
We conclude that variations the household s income with prices con
stant shift the budget line parallel to itself inward toward the ongin when
income falls and outward away from the origin when income uses The
reader should now draw bis own figure simdar to Figure I and sketch M
m on It budget lines for incomes of £V00 and 3^40 per month

Changes in Prices

\\ e may now consider what happens when paces change Since Figure 14 I

js rather cluiicred we transfer the line f/ (illustrating income ;Ol80 pace

of *=£4 and pace of } to Figure 14 2 Now assume that the pace

w the budget I fle caused by changes in prices

/>/ 2 SI res
 ;

SOME BASIC THEOREMS AND PREDICTIONS 171

of Yremains unchanged while the price of X

changes. Say, for example,
that from
it falls to ^2. The household can now buy either 90X, if it
devotes its whole income to that commodity, or 90 F, if it purchases only Y.
The combinations available to the household are given by the budget line
eh. A few points on this line should be checked to see that the combinations

ofX and F indicated exactly exhaust the household’s income of £180 when
X and F are both priced at £2.
Now assume that the price of X falls to £1, the price of F remaining
constant at £2. The new budget line indicating all the combinations avail-
able to the household is ei. A comparison of the budget lines ef, eh and ei
shows that changing one money price changes the slope of the budget line
the lower is the price of X the flatter is the budget line.
Next let us try changing both prices in the same proportion. If we start
from budget line eh (income £180, X and F priced at £2) and double both
prices, then we will halve the amount of both goods that the consumer can
purchase. The budget line now becomes^ because the income of £180 will
now buy 45 F and no X, 45 A and no F, or any combination of X and F on
the straight line joining these two points.
Now let us go back to line eh (£180 income and a £2 price for X and F)
and halve both prices. The consumer can now have twice as much of both
commodities as previously and the budget moves outward to ki.
line
We notice that changing both prices in the same proportion shifts the
budget line parallel to itself in the same way as a change in income shifted
it. Before reading further the student should go back to Figure 14.1, take

budget line ab (income=£120, F=£2 and A=£4), and determine what

price changes would shift the budget line to cd and to ef.

Relative Prices and Opportunity Costs

The reason why changing both prices in the same proportion shifts the
budget line parallel to itself is that the slope of the budget line indicates the
opportunity cost of one commodity in terms of the other. An equal pro-
portionate change in both prices leaves this opportunity cost unchanged.
X
For example, if F costs £2 and costs £4 then 2F must be foregone in order
to be able to purchase lA; if F costs £4 and X costs £8 it is still necessary
to forego 2 F to be able to purchase one more X. In fact, as long as the price
of Xis twice the price of F, it will be necessary to forego two F in order

to be able to purchase one more X. More generally, the amount of F that

must be given up to obtain another unit of X
depends only on the relation
between the price of Y and X. If
the price
of we take the money price of and X
divide it by the money price of F we have the opportunity cost of in X
172 THE INTERMEDIATE THEORY OF DEMAND
terms of 1 (the quantity of J that must be foregone in order to be able to
purchase one more X) This may be written

A opportunity cost of Xm terms of Y,

Py

where ^and />, are the money prices of A' and Y It is apparent that
changing both prices in the same proportion leaves the ratio
pjpy un-
changed In economics, this ratio is called a relative price, which is the
term used when any price is expressed as a ratio of another price Relame
prices must be distinguished from money prices, which are called absolute
prices For example, if we say that the price of X is £2 we are speaking
of an absolute pnee, if we say that the price of A is twice the price of J', we
are speaking of a relative price This one relative pnee lpx~2p.) is consistent
with an infinite number of absolute prices as long as the absolute pnee of A
is twice that of 1 for instance, P,"£l and Px=£2 will dojust as well as

will p,=£lO and px-£20 or 000.000 and ^,=^2,000,000

Conclusions Irom the Analysis

\Ve have now reached a number of conclusions that are of extreme import-
ance for demand theory The student should re-read the relevant parts of
this chapter if he cannot prove for himself cacJi of these propositions

1 A change in money income with money prices

(and thus necessarily relative prices) constant shifts
the budget line parallel to itself, inward toward the
origin when income falls and outward away from the
origin when income rises.
2 An equal percentage change in all absolute prices
leaves relative prices unchanged, and, if money in-
come remains unchanged, it will shift the budget line
parallel to itself, inward toward the origin when prices
rise and ouWard away frouv the origin when, prices
fall.

3 Multiplying all money prices by the same con-

stant, while holding money income constant, has
exactly the same effect on the budget line as multiply-
ing money income by 1/A, while holding money prices
constant. For example, doubling all money prices has
the same effect on the budget line as halving money
income.
 : 1

SOME BASIC THEOREMS AND PREDICTIONS 173

4 A change in relative prices causes the budget line

to change its slope.
5 An equal percentage change in all absolute prices
combined with a percentage change in income of the
same magnitude, and in the same direction as the
price change, leaves the budget line exactly where it
was before the changes occurred.
All of these conclusions except the last one were illustrated in Figures 14.
and 14.2. As an illustration of conclusion 5 consider thebudget line eh in
Figure 14.2. This budget line was originally obtained from an income of
;^180 and prices of £2 both for X and Y. Now consider doubling the house-
hold’s income to ^(^360. This shifts the budget line to ki, since the household
can buy twice as many units of both commodities as it could previously.
Now double both prices from £2 to £A. This cuts the household’s house-
hold’s consumption possibilities in half and it is back with budget line eh.
(A few numerical calculations will verify the fact the ,^180 with a £2 price
for X and Y can buy exactly the same combinations of goods as can ^{^360
with a £A price for both commodities.)

Predictions from the Analysis

The five conclusions listed above are matters of logic. The effects of various

changes on the budget line are incontrovertible. In order to translate these

conclusions into predictions about household behaviour, we advance the
hypothesis that a household’s market behaviour depends solely on the tastes
of the members of the household and on the location of the household’s
budget Another way of wording this hypothesis is to say that the
line.

household’s behaviour depends on real rather than on money variables. This

behavioural hypothesis along Math the five propositions above will allow us
to make testable predictions about the behaviour of households. Conclusion
3 leads to the following prediction

The change in a household’s market behaviour will be

the same if either its income changes by 7. or if aU
money prices change by \[k.
And conclusion 5 leads us to predict that:

The household’s market behaviour will be unaffected

if its money income and all money prices change
simultaneously by some multiple

Strictly speaking, these predictions apply only to households that do not

have any significant quantities of bonds, cash or other assets whose value is
 174 THE INTERMEDIATE THEORY OF DEMAND
f«cd money uniti and whose real value thus changes when
in
money pnccs
change The extension of demand theory to cases in which
households have
and properly belongs in a more advanced text
assets IS difficult
In the
meantime, we note that the above predictions apply to those
households for
which currently earned income is the mam determinant of
current expendi
ture We should also note that the great majority of all households fall in
this class

REAL AND MONEY INCOME

On the basis of the previous analysis we can now make the important
distinction between real income and money income Money
income
measures a household s income in terms of some monetary unit so many
dollars or so many pounds sterling, real income measures a households
income in terms of the command over commodiues that the money income
confers A nsc in
money income ofx per cent combined with an x per cent
nse m all money pnees leaves a household s ability to buy commodities
and hence its real income, unchanged When we speak of the real value of
a certain amount of money, we are referring to the goods and services that
can be bought with the money we are referring that is to the purchasing
power of the money

Allocation of Resources the Importance of Relative Prices

Our theory of pnee predicts that the allocation of resources depends on the
structure of relative prices If the money value of all pnees incomes debts
and credits were doubled there would according to our theory be little
noticeable effect The economy would function as before The same set of
relative pnees and real incomes would exist and there would be no in
centive for any reallocation of resources the only difference would be that
the money level of all prices and incomes would be doubled
This prediction is an implication of our theories of the behaviour of
households and of firms We have already seen that doubling money pnees
and money incomes leaves the household s budget line unchanged and so

according to the theory of household behaviour the change gives households

no incentive to vary the quantity of each commodity that they purchase
As far as the producer is concerned if all prices both of final goods and of
factors of production double, then the relative profitability of different
lines of production will be unaffected as indeed will the real level of profits
in all lines of production ^ Thus producers will have no incentive to alter
1 What forexample happens to the real valucofacash hoard of £100 if all prices double’
2 Since all pnees and costs will have doubled money profits will have
doubled but the
purchasing power of these profits will be just what it was before the change occurred
 SOME BASIC THEOREMS AND PREDICTIONS 175

production rates so as to produce more of some things and less of others.

If, on the other hand, relative prices change, then our theor)' predicts

that resources will be re-allocated. Households will buy more of the cheaper
goods and less of the expensive ones, and producers will expand production
of those goods whose prices have risen relatively and contract production of
those goods whose prices have fallen relatively (since they would be
relatively less profitable lines of production).
The theory' of prices and of resources allocation that we have developed
so far is thus a theory of relative, not of absolute, prices.

Inflation and Deflation the Importance of Absolute Prices

:

The average level of money prices is called the price level. If all money
prices double, we say that the price level has doubled. An increase in the
price level is an inflation, a decrease is called a deflation. If a
called
rise in all money prices and incomes has little or no effect on the allocation

of resources, it may seem surprising that so much concern is expressed over

inflation. Clearly, a person who spends all of his income, and whose money
income goes up at the same rate as money prices go up, loses nothing from
inflation. His real income is unaffected. There are, however, two problems
that make inflation more serious than it might seem from the above dis-
cussion. First, in our society, many contractual obligations are fixed in
money terms. Second, when money incomes rise during an inflation they
do not same speed there are important leads and lags in the
all rise at the ;

adjustment process and these affect the relative positions of different groups
in the society.
To consider the first point, notice that a vast number of contracts is fixed
inmoney terms. If today I borrow from the bank enough money to buy a
new car, I do not promise to pay back in two years time enough money to
buy a new car; I promise, rather, to pay back the same sum of money as I
borrowed. If prices have doubled in the meantime, I will pay back to the
bank only half as much as is needed to buy a car at that time, while if
prices have halved I will pay back enough money to buy two cars! If I pay
money for an annuity or for life insurance, the company enters into a legal
obligation to pay me or much money when I retire or die. The
my heirs so
purchasing power of this sum of money will depend on what prices are ruling

at the date in the future when the money is paid. Old age pensions, all

borrowing and lending contracts, mortgages and hire-purchase agree-

ments, are but a few examples of contractual obligations fixed in money
terms, the real value of which fluctuates with the price level.
One dramatic example is found in wartime borrowing by governments. A
household which bought a British War Bond for £75 in 1940 may have been
 176 THE INTERMEDIATE THEORY OF DEMAND
pleased lo know that it could cash it in for
^^100 m
1950 But this ^flOO in
1950 bought It onl> ^41 worth of goods at 1910 prices
In other words, by
not buying ^50 worth of goods in 1910 the household
secured the right to
buy £4i worth of goods 10 years later This is a rate of
interest o[ minus 2
per cent per year over the period often yean' To take another example,
a
British Naval Lieutenant Commander who joined the Service at
18, in 1915,
and served his country through two world wars, retiring on a pension of
;{^800 m 1945 would have found that by 1955 the real value of his pension
had shrunk drastically Had the price level remained unchanged since he
entered the Service m
1915, his pension would have bought ;^^1,716 worth
of goods at 1965 prices Since it is very unlikely that the possibility of such a
dramatic change in the pnee level ever occurred to him, it is not stretching
a point to say that in return fora lifetime of service through two wars, he
was defrauded of over 50 per cent of what he had a right to expect on retire
ment, and was given a standard of living once his useful life was over not
much above a bare subsistence level
Thus we see that mnaiion lowers the real value of all debts by reducing
the quantity of goods and services that can be purchased with a given
amount of money The person who has borrowed money gains from in
flation because the sacrifice in terms of goods and services in repaying the
debt IS reduced whereas the person who has loaned money loses because
the real value of themoney relumed is less than the real value of the money
originally given up By the same token anyone who saves money buys
ordinary life insurance or in any other way disposes of his savings so that
he gets a fixed money return loses by inflation An inflation of 3 per cent
per year will reduce the purchasing power ofa sum of money to one half its
original amount in 23 5 years Thus one of the effects of inflation is to
reduce the real value of all debts savings and fixed money incomes
The second consequence of inflation ihal was mentioned above arises
from the fact that all adjustments do not take place instantaneously If the
price level were to double, most money incomes would
probably double

eventually so that the real value of all incomes would be restored to their
originallevel In the process however some groups will manage
to increase

vnetease theirs only slowly

their incomes quickly, vvhereas other groups will
In a state of continuous inflalion those groups whose money incomes nse

quickly gam at the expense of those whose monev

incomes rise only slowly
discussed in
The theory of the detcrnuoation of the absolute price lev el is
Chapter 51 of Part VIII
CHAPTER 15

THEORIES OF
HOUSEHOLD BEHAVIOUR

I N this chapter we shall consider two of the three most important theories
of household behaviour. The prime purpose of the discussion is to derive
the prediction that the household's demand cuiA^e for any product is doivn-
ward sloping from assumptions about household behaviour. We wish to do
this because we wish our relations back to decisions on the
to trace all of
part ol the decision-taking units described in Chapter 6. For many purposes
in positive economics we can take the downward sloping demand cun'e as
a basic assumption and when we do this we can rest secure in the knowledge
that a vast amount of evidence e.xists that conforms with the hypothesis that
demand curves slope downward. For completeness, however, we wish to
trace this relation back to basic household behaviour. In the process of
doing this we hope to develop some very important subsidiarj' insights.
The theoiv' that we describe first is the oldest of the theories of household
beha\ iour, that of marginal utilily. T'he distinction in this theory' between
marginal and total utility seems to me to be one of the most important
distinctions in all of economics. Failure to appreciate its significance has
caused great confusion in public debate and has rendered useless many
I have chosen to discuss
otherwise carefully prepared sociological studies.
this theor>' of household behaviour because of the important by-products
that result from the critical distinction between marginal and total utility.

The second have described at length is the theory' of revealed

theory that I

preference. I have chosen this theoiy^ rather than the more popular
indifference-preference theory because it seems to me that if all we want is
to distinguish in an operational w'ay the income and the substitution effects
and then to derive the prediction that demand curv'es slope downward,
revealed preference theory is the most economical method. At a later stage
in his study the economics specialist will have to master indifference cur\'es
 THEORIES OF HOUSEHOLD BEHAVIOUR 179

whether to have a bit more of one thing or the other. If we find that we
are overspending a bit, or if our income falls a bit, we have to decide what
to cut down on, to have a bit less of this or a bit less of that. Choices in
the world are thus rarely conditioned by total utilities; it is marginal
utilities that are relevant in deciding choices concerning a bit more or a
bit less.

The Hypothesis of Diminishing Marginal Utility

We now introduce the basic hypothesis that for any individual consumer the
value that he attaches to successive units of a particular commodity will
diminish steadily as his total consumption of that commodity increases, the
consumption of all other commodities being held constant. This is called the h\q3 o-
thesis of diminishing marginal utility. .Another way of stating the hypothesis
is to say that the total utility that he derives from consuming the com-
modity increases the more he consumes, but at a diminishing rate so that,
for example, the 100th unit consumed per month will add less to his total
utility than did, say, the 10th unit.

If we assume we have some index with which to measure utility we can

illustrate these hypotheses in the followdng figures. Figure 15.1 (i) shows that
the total utility derived from consuming the commodity rises as more of it is
consumed but Figure 15.1(ii) shows that the amount of utility derived from
consuming each additional unit of the commodity gets less and less the more
of the commodity that is already being consumed. For instance when Oa is
consumed per month total utility is ab but marginal utility is only ac, shoiving
that when ab is already consumed the utility of consuming a bit more is

quite low.^

Water, an Example: As an example of the difference betw'een total and

marginal udlity consider the example of ivater. Some minimum quantity
of water is absolutely necessary to sustain life and a person would give up
all his income and wealth to obtain that quantity: we may say that the

marginal of that quantity is infinite. Much more than this bare

utility

minimum can be drunk and we can say that the marginal utility of
successive glasses of water drunk over some time period will decline
steadily.^ But water has many other uses. A fairly high marginal utility will

1 Mathematicians will recognise marginal utility as the first derivative of total utility ^^^th

respect to the quantity consumed. If describes total utility as a function of the

quantities of X consumed then marginal utility is dUldX—{'{X).

much money you would have to be given to induce you
2 If you doubt this ask yourself how
to cut your consumption of water by one glass per week. The amount will not be vast. Now try

theamount necessary to induce you to cut it by a second glass, then by a third glass. In this
way you will see that the fewer glasses you are consuming already, the higher the marginal
utility of one glass of water.
 intermediate theory of demand
be attached to some minimum quantity for bathing but much more
than
this minimum will be used for frequent baths and for
having a water level
m the bath higher than is absolutely
necessary The last gallon used for
athmg is likely to have quite a low marginal utility Again some small
quantity of water is necessary for cleaning one s teeth but a vastnumber of
people leave the tap runmng while scrubbing their teeth, and they can

Units of commodityXciMisumedper
month

0 a
Units of commodity Xcwisumed per
month

Ftg 15 I Marginal and total utility curves for a household

hardly pretend that the water so consumed between wetting and rinsing
the brush has a high utility Similarly for cleaning one s car A couple of pails
of water per week for this purpose probably have high utility But the
gallons poured down the dram if the hose is left on between successive nnses
provide no more than a minor convenience When we consider all the

extravagant uses of water for the modem consumer we can hardl> doubt
that the marginal utility of the last, say, 30 per cent of all the units con
sumed is very low even though the total utibty of all the units consumed is
 THEORIES OF HOUSEHOLD BEHAVIOUR 181

infinitely high. The marginal utility cur\'e is probably of the shape shown
in Figure 15.2. (This is more than idle speculation for it leads to the
prediction that there wll be a large increase in the amount of water con-
sumed if water
provided free rather than at a modest price. This pre-
is

diction has, as we shall see later, been confirmed over and over again to the
discomfort of many water authorities.)

Fig. 13.2 Hypothetical marginal utility curve for water consumed by a typical
household.

Ca.v marginal utility ever REACH ZERO? With many commodities

there issome maximum consumption after which additional units would
confer no additional utility, and if the indhadual were forced to consume
,more they would actually reduce his total utility. Tobacco and food are
ob\dous examples. There is some maximum number of cigarettes that most
people would smoke per day even if they did not have to tvorry about the
cost. Few people would want to go to the point of chain smoking from the
second they awoke until the second they fell asleep. Long before con-
sumption reached this point additional cigarettes smoked would cease to
add to utility and would begin to subtract from it, i.e., additional cigarettes
would have a negative marginal utility or as it is sometimes called a
disutility. The same is undoubtedly true of many other commodities
such as food, alcoholic beverages and most recreations. (Although some
people might be happy to play golf from sun up to sunset seven days a week
for the rest of their lives, many would not.)

The Hypothesis of Utility Maximisation

We now introduce a basic assumption about the behaviour of the indivi-

dual members of households.
 182 THE INTERMEDIATE THEORY OF DEMAND
The members of households will so act in respect to
their consumption decisions as to gain
the largest
possible total utility. We say they seek to
maximise
their total utility.

1 his IS just another uaj of saying that the members of households

try
to make themscKcs as ncll olT as the) possibly can in the circumstances
in which themseKes *
tliey find

From this we can immediately derive one prediction that turns out to be
important at a later stage

Consumption of any free good will be pushed to the

point at which its marginal utility is zero.

A free good is one for which no pnee needs to be paid As long as a

further unit consumed has a positive marginal utility, total utility can be
increased b> consuming more of the commodu> InFigure 15 3 if cigarettes
were a good the consumer would consume Of of these per day, making
free
total utilit)from smoking as laige as possible {fg) At this point, further
cigarettes consumed have a negative marginal utility and so consumption
will not be increased beyond Of
\V’e can now restate the proposition about scarcity on page 59 m the

following terms

If all goods were free, the total amounts that the popu>
lation would want to consume would greatly exceed
the amounts that could be produced with the available
supplies of resources. Therefore marginal utilities
must remain positive for at least some goods, i.e.,
households would get additional utility from con-
suming more of them.
A SINGLE HOUSEHOLD WITH A FIXED INCOME FACING GIVEN MARKET
PRICES We can now go on to the more important point of considenng the
position when all commodities command a posiUve price How will a
household adjust Us expenditure so as to maximise the total utility of its
members when it has to pay for the goods it consumes Should it go to the
selfish and
I This IS sometimts uken to mean that individuab are assumed to be narrowly
devoid ofany altruisuc motives This is not »
If ihe individual derives utility from giving his
money away to others this can be incorporated into the analysis and we can compare for

example the marginal utility of a pound that he gives away with the marginal utility of a
pound that he spends on himself
2 The justification for personifying a household m this context is given m Chapter 6
pages 76-7
 THEORIES OF HOUSEHOLD BEHAVIOUR 183

Quantity of cigarettes consumed

per day

Fig. 15.3 Hypothetical total and marginal utility curves for cigarettes consumed
by a typical household.

point at which the marginal utility of each commodity is the same? At

this point it would value equally the last unit of each commodity con-
sumed. This would make sense only if each commodity had the same price
per unit.
When different commodities have different prices, the condition for total
utility to be maximised is that the marginal utility of each commodity
should be proportional to its price. If there are n commodities we can write
this as:

marginal of good marginal utility of good 2

utility 1 _
price of good 1 price of good 2

marginal utility of good n

price of good n

Consider the case of two commodities X

and Y when the price of is .£2 X
per unit and Y £\ per unit. Assume that total consumption of and 7 has X
been taken to the point at which the marginal utilities are the same for
 184 THE INTERMEDIATE THEORY OF DEMAND
both goods Now if one
less unit of X
were bought, two more of } could be
bought This would leave total expenditure constant but it would
raise total
utility, because the marginal utility of one
unit of Z is equal to that of one
unit of y Thus It would pay the household to reduce
its consumption of X
and increase its consumption of Y According to the hypothesis of
dimmish
ing marginal utility, the marginal utility of A' will nse and that of Y will
fall
as It consumes more Y and less X
Say the household goes to the point at
which the marginal utility of X is three times that of > Now, clearly, it
has lowered its consumption of Afand raised that of Tby too much Consider
what would happen if it now transferred £2 worth of expenditure from Y
to X This reduces its consumption of ¥ by two units and increases its
consumption of A by one unit But since the marginal utility of the unit of
A IS three times that of one unit of Y and hence l-J times that of two units
of Y Its total utility is increased by this transfer of expenditure The house
hold will thus transfer expenditure from T to A' and as it does so the
marginal utility of X will fall and that of Y will nse If it carncs on with this
until the marginal utility of AT is only twice that of Y it will be m a position
in which total utility is as high as possible In this case a very small transfer
of expenditure between the two goods leaves utility unaffected since cutting
Y consumption by two allows it to increase X consumption by one and
the utility of one X is twice that of one 1 (and hence equal to that of 2Y)
‘
so a small transfer leaves total utility unchanged

A SINGLE HOUSEHOLD WITH A FIXED INCOME FACED WITH A CHANGE IN

MARKET PRICES Now Consider what happens when market prices change
Say for example that a household starts m a posiiion in which
marginal utility of good I price of good 1
_ _ ^
marginal utility of good 2 price of good 2

i An alternative derivation of the »ame nile in jummary form u ai follows

(a) The household must so allocate m expenditure that the marginal otility of the Vast U
spent on each commodity a the same (otherwise M could raise total utility by transfemng ex
penditure to those goods where the marginal utility of a penny spent is highest)

(b)This gives us the following equihbnum condition

roarg ut of Irf spent on good 1 = marg ut of id spent on
good 2
spent on good I depends on how much you get of good
1
(c) The marginal utility of l,f
uoil of that good Similarly for good 2 We can wnte
for Id and on the marginal utility of one
this marg ut of one unit of good 1
mTi n. of 1 J spmt on good I - „r . of good 1

marg ut of one unit of good 2

m.rg ot of Id Jpont on good 2 - g„„d 2

of good 6 and a unit

which says nothing more than that sf the marginal utility of a ut .
I IS

of good I costs 2./ the marginal utility of W

spent on good I is 3
(cont on next page)
 THEORIES OF HOUSEHOLD BEHAVIOUR 185

and that the price of good 1 rises until

price of good 1

price of good 2

What will happen ? In general we expect the household to reduce its pur-
chases of 1 and to increase its purchases of all other commodities until the
ratios of marginal utilities conform to the ratios of the prices.* Will the
household cut its purchases of 1 a lot or a little? Figure 15.4 shows two
possibilities. In the case of good I' the marginjil utility curve is very' flat

Fig 15.4 Two illustrative marginal utility curves for a household.

over the relevant range and consumption of the good is cut from Oa to Ob'

before the marginal utility of the good is doubled- Up to that point there is

a clear gain in transferring expenditure from 1 to other goods. In the case

of good 1" the marginal utility curve is fairly steop over the relevant range
and only a slight reduction in purchases from Oo to Ob" suffices to double

(d) If we now substitute from (c) into (b) we get:

marg. ut. of good ! marg. ut. o f good 2

price of good 1
price of good 2

Multiplying through by (price of good l)/(marginal utility of good 2) we get:

marginal utility of good 1 price pfgood 1

marginal utility of good 2 price of good 2

which two goods. repetition of the argument will

is equation (1) in the text for the first .'V

show it to be true for all pairs of goods.

1 For purpose of this argument we assume that good I taKes up a very small percentage of
the household’s total income so that when it buys less of 1 the effects on its expenditure on all
other goods are negligible so that the marginal utilities of all goods other than 1 change in onlv
a negligible fashion.
 186 THE INTERMEDIATE THEORY OF DEMAND
the marginal utility In both cases the household reaches an
equilibnum
that satisfies relation (1), but in one case
purchases fall a lot ’vhile in the
other they
fall only a little The common sense of this
is that with good 1'
themembers of the household do not value all the units between
Ob' and
Oa much more highly than the Oath unit and, when the cost
of obtaining
units of thegood doubles, they prefer to dispense with them and buy other
goods In the case of 1" the members of the household value
most of the
units between Ob' and Oa very highly, and, when the cost of obtaining
them doubles, they continue to consume most of them, only cutting out
those units between b" and a which are the ones valued at less than twice
the value given to the Oath unit
Notice that the reactions arc quite independent of the slope of the curve
outside of the relevant range Curve (1) could be as shown by the sobd line
orIt could be as shown by the broken line and the result would be the same

As we have shown the curves, good 1 has a higher total utility than curve
1' since the marginal utility of the first few units very high indeed in the
is

case of r, but, neveiihelcss, the nse in pnee reduces the consumption of 1'

by much more than T

This leads us to the following important conclusion

The response of quantity demanded to a change sn

price (he., the elasticity of demand) depends on the
marginal utility over the relevant range and has no
necessary relation to the total utility of the good

The Derivation of the Demand Curve from Marginal Utility Theory*

For completeness we now sketch the derivation of the hypothesis of a

shall

downward sloping demand curve from utility theory The denvation from
indifference-preference or from revealcd-preference theory is more satis-
factory, and this section can be skipped if desired start by assuming We
that the commodity taking a very small part of total con-
in question is

sumers’ expenditure so that vanaUons in the amount spent on this

commodity will have only a negligible effect on the marginal utilities of all
other goods — hence on the margtna} utility of another pound of income
To a close approximation therefore we say that the marginal utility of in-
come IS constant throughout our exerasc and this allows us to use money to
measure the marginal utility of the good m question If we ask how
much
money would be given up to get one more unit of the commodity we have
a measure of the marginal utility of this good In Figure 15 5
we measure
the marginal utility of good in X
against the total quantity of A con-
sumed If we now assume that the household is faced with vanous prices of
1 This section is optional
 THEORIES OF HOUSEHOLD BEHAVIOUR 187

the good it will clearly pay it to go on buying the good as long as the
marginal utility of additional units measured in pounds exceeds the number
of pounds it must give up to obtain the good. Thus, at each price, the

quantity purchased will be that which equates price with marginal utility.
If we amount purchased at each price, we have the
plot from this the
household’s demand cur\'e which is shown in Figure I5.5(ii) and which is
clearly identical to the cur\^e in Figure 15.5(i).

Fig. 15.5 The derivation of a

household’s demand curve for
some product from the marginal i

utility curve for that product.

ii

Total and Marginal Utilities : The 'Paradox' of Value

Early economists struggling tvith the problem of what determined the

relative prices of commodities, encountered tvhat they considered to
be a
 188 THE INTERMEDIATE THEORY OF DEMAND
paradox very necessary commodities such as water were
observed to have
pnees which were low relative to the pnccs of many luxury
commodities
such as diamonds These early economists argued somewhat
as Mows
Water is necessary to our very existence while diamonds are a frivolous
luxury that could disappear from the face of the earth tomorrow
without
causing any real upset Docs it not seem odd therefore that water is so cheap
relative to diamonds^ It took 75 years before this apparent
paradox was
satisfactorily resolved and so we should not be too surprised if many present
day students policy makers and ordinary atizens still fall into the confusion
which so long ensnared economists The policy implication of so doing can
sometimes be very serious so it is worth tabng time to consider the problem
The paradox of value is more accurately described as being the problem
created by the fact that an intuitively appealing hypothesis was obviously
refuted by everyday empincal observations The hypothesis that the early
economists were implicitly working with was that the relative prices of goods
ought to conform to their relative Mai utilities i c

price of good total utility of good

I
_ 1

pnee of good 2 total utility of good 2

This hypothesis is continually refuted Clearly the total utility of water

exceeds that of diamonds (if you are in any doubt apply the test of question 1

on page 178 to these two goods) and yet diamonds are expensive relative to
water

The relation between variable market prices and marginal

UTILITIES Wehave seen above that any single consumer seeking to maxi
misc reach a position described by equation (1) on
his total utility will
page 183 and not the position described by equation (2) VVe now show that
above To
in equilibrium market prices must conform to the relation in (1)
see this start by assuming that they do not Assume for
example that we
arc m a position where for every household

marg vKAAy tX {\) ptvet of (1)

^ ^2^
marg utility of (2) pnee of (2)

try to increase
No household will be in equihbnum Every household will
bid up
Itsconsumption of (1) and lower its consumption of (2) This will
the price of( 1 ) and lower
the price of (2) The change in prices will induce
producers to produce more (1) and less (2) and as these
changed quantities
will fall and the
of the two goods are consumed the maipnal utility of (1)
marginal utility of (2) will nse
 THEORIES OF HOUSEHOLD BEHAVIOUR 189

Both of these changes (i.e., the rise in the relative price of (1) and the
fallin its marginal utility) will continue until:

marg. utility of (1) price of (1)

marg. utility of (2) price of (2)
The student should make sure that he understands this very important
argument by reversing the inequality on (3) above so that:
marg, utility of (1) price of (1)
(4)
marg, utility of (2) price of (2)

and then repeating the argument to show how the situation described by (1)
will be restored.
Thus the ‘paradox’ of value is resolved by saying that the intuitively
appealing hypothesis described by (2) is wrong in two senses. It is wrong
empirically in that it does not explain the relative prices vs'e see in the world.
It is also wrong in the sense that it is not a logically valid deduction from
the early economists’ own assumption that people try to maximise their
well being. Indeed we have already shown that (1) and not (2) is the pre-
diction that follows logically from the assumption that consumers try to
maximise their total utility.

Necessities, luxuries and total utility: The general public tends to

rank goods on a scale as luxuries or necessities. Students usually do the same
and market behaviour bearing some relation to this
often expect to find
We have already considered an example on pages 117-18
classification.
where we rejected on empirical grounds the hypothesis that there was a rela-
tion between demand elasticity and the degree of necessity of the good. If we
can give any meaning to the ranking of goods as necessaries or luxuries it is
probably in terms of a ranking of their total utilities which is determined by
asking question 1 on page 178 for each pair of goods and ranking them
accordingly. A few selected goods are shown on the scale below. Others
could easily be filled in.

‘luxuries’ ‘necessities’

diamonds transport. . .shelter. . .clothing

. . .entertainment . . . water . . . food . . . air

Total Utility
low high

We have already seen that ive expect no market behaviour to be explained by total

utilities.
 190 THE INTERMEDIATE THEORY OF DEMAND
Emotional reactions to goods Tlierc is very little doubt that the
emotional reaction of people to goods is in response to
their total utiliues
rather than to their marginal ones We often hear an
argument such as the
following ‘Water is a necessity of life of cntical importance
to rich and
poor, and it would be wrong, therefore, Co make people pay for so necessary
a commodity This is typical of the public's reactions to many goods But
’

such views often produce curious results If, for example, water is provided
free instead of at a modest cost, the new consumption that will occur will
be on account of the many uses that yield a relatively low utility {such as
letting the tap run while cleaning one’s teeth and one’s car) The relevant
question when deciding between a zero and a modest price for water is not
‘Is water so necessary that we would not want to deprive anyone of all of
It’’ but rather ‘Arc the uses of water which our policy will encourage so
necessary that we don’t want to discourage anyone from adopting them m
spite of the fact that it is costly to provide the water for these uses ’’ Clearly,
these two questions could be given different answers This leads us to a
general warning which we shall take up again in a later chapter

Extreme cautloo should be used in basing policies on

'a consideration of total utilities Very often, if not ab

ways, it is marginal utilities that are relevant, rather

than total utilities.

Attitude Surveys and Total Utilities an Example

The between total and marginal utilities is one of the most

distinction
important in of economics As an example of its pervasiveness, consider
all

a case from the apparently unrelated field of attitude surveys which are
so
the
popular these days in sociolc^y and political science These surveys take
form of asking such questions as
‘Do you prefer blondes or brunettes'^’
’’
‘Do you like the Conservatives more than the Socialists
influenced your
‘In deciding to live in area A rather than B what factors
schools,
choice’ List the following in order of importance neighbours,
amou-oJ. of Oree land.,
closeness to ywmTTmng acita, ymet, quralvvy of
play areas for children, general amenities’
‘In choosing a university what factors
were important to you’ List in
residential facili-
order of importance, environment, academic excellence,
facilities, clubs
ties, parents’ opinion, school opinion, sports
add
The reader of the Sunday papers and popular books will be
able to

many other examples to their list (aU of which are drawn from real
cases

surveys, and most of tht

rather than the author’s imagination) AH of the above
others the reader will he able to add, attempt to
measure total rather than marginal
 THEORIES OF HOUSEHOLD BEHAVIOUR 191

Utilities} There is, of course, nothing wrong with this per se. Anyone is free
to measure anything that interests him and in some cases knowledge of total
utilities may be of practical value. But in many cases actual behaviour will,

as we have already seen, be determined by marginal utilities, and if one

attempts to predict such behaviour from a (correct) knowledge of total
utilities one will be hopelessly in error.

This general warning is probably sufficient at this point since a careful

statement of what goes wrong in each case is usually fairly complex. How-
ever the barest outline of one example can be sketched in for the interested.
If it seems too tough - and it is not easy - it may be skipped by intending
economists since attitude surveys are seldom used in that subject. Intending
and political scientists should, however, stick with it until they
sociologists
really understand it.

Assume that University X is disturbed that staff is spread around over

its

many far flung villages with no tendency for an ‘academic community’ to

develop near by the University site. They employ some sociologists to study

the causes of thisand the researchers conduct the study suggested above (of
course it will be more detailed than we have suggested). Assume that they
find that most people have located away from the University in order to be
near the sea, being influenced by the proximity of swimming for themselves
and their children. The University wants to attract new staff members to
locate close to the site; will it suffice to build swimming facilities on the
University housing estate?
The surveyors have measured the relative total utilities of the quantities
of the various attributes actually consumed. The quantity consumed is of
course a function of the price. If swimming is free as it is at the seaside, the
total utility from it may
be large. But if a charge is to be levied for it (and
somehow the University must recoup the cost of providing swimming pools
on its estate) the total utility people will then get from it depends on the
shape of their marginal utility curve. If typical curves are like curve A in
Figure 15.6 then at even a modest price people would consume no swim-
ming and would not be attracted to an area which provided such facilities
at a price in excess of the marginal valuation they placed on the first unit
consumed. If the curve is like B then at a modest price total consumption
would not be much less than when the price was zero and total utility
would remain quite high.
Much more can be said but the above should be enough to show that the
knowledge provided by these surveys is little guide to the policy makers in
deciding how to attract staff back to the proximity of the University. On
the other hand, two relevant questions are ‘How much less would the price
I I am indebted to Mr G. G. Archibald for making this very penetrating point,
when we
were discussing the practical value of a particular attitude survey.
 192 he intermediate theory of demand
of housmg have lo be near
the me ,o lead vou ,o locate here rather thaa

ad.aa,ag, of the seas.de local,

or. over a nearby one)
by how much would we have to change the
> and ‘At rm,„j fj
attractions of the nearby

Ftg 15 6 Two alternative marginal

utiiu) curves for swimming for a
typical household

the 9umm«r months

location to make >ou move from

the seaside’’ Both of these are marginal
questions All too often social investigators think that by measuring total
utilities (at existing prices) they can infer something about what is needed
to change behaviour (at unspecified pnces)

The 'Law' of Diminishing Marginal Utility of Income*

So far we have been discussing the hypothesis of diminishing marginal

utilityfrom the consumption of successive units of a single commodity It is
often stated that it is a ‘law’ of economics that the utility or satisfaction
gamed from the expenditure of an extra pound’s worth of income dimi-
nishes steadil) as income rises Thus it may be stated that an extra pound

of income will increase a man’s satisfaction more when he is poor than it will
when he is rich Yi we mahe the added assumption that ah peopW art TnoTC
or less the same m their capacity for enjoyment, then it follows from the

‘law’ of diminishing marginal uuhty of income that a pound taken away

from a rich man causes him less loss of satisfaction than does a pound taken
away from a poor man Thus, if wc wish lo raise a given sum in tax revenue
and tf we wish all individuals lo have an equal loss of satisfaction, it might
be necessary to take away, say, two or three pounds from the rich man for

I This secuon is optional and can be omitted without interupung the flow of the argument
 : :

THEORIES OF HOUSEHOLD BEHAVIOUR 193

every pound that we take from the poor man. Thus the ‘law’ of diminishing
marginal utility of money plus the idea that all individuals are approxi-
mately equal in their capacity for satisfaction can be made the basis for
redistributive policies such as the progressive income tax.
If the ‘law’ is a testable proposition, then
must tell us that some things
it

can happen in the world while other things carmot happen. In order to see
if this is the case, let us imagine how we might go about refuting it. .'Assume,

for example, that we go to an individual and ask him if his satisfactions from

spending additional units of income have diminished as his income rose.

The individual might then tell us the following story

‘I suffer from stomach trouble which results in intense agony after ever)'
meal in fact ;
I positively dislike eating. Also I seldom notice my surround-
ings so I don’t much care where I live. However, my doctor tells me I

must eat well to keep my stomach functioning at all, and it is necessary to

live in a tolerably comfortable place in order to keep healthy. Thus, when
I had a very low income I spent most of it on food and shelter, and derived
very little satisfaction from so doing. After my income rose beyond a certain
level, I was able to afford a few small luxuries and these brought me real

pleasure when my income rose still further, I was able to buy large items
;

of luxury; a fast car, holidays abroad and a telescope to pursue my hobby

of amateur astronomy. Clearly, the satisfaction that I receive from the
expenditure of additional increments of income has been increasing steadily
as my income has risen.’

Now, how might a defender of the ‘law’ of diminishing utility of income

? There is one ver)' common defence which needs careful
react to this attack
attention here. Let us imagine a supporter of the ‘law’ replying as follows

is only fooling himself. If he really got more satisfaction from

‘This person
his luxuries than from food and shelter, he would have bought the luxuries
when he was poor. His behaviour shows that he values his health more than
the other satisfactions that he mentions. The fact that he chose food and
shelter in preference to these luxuries proves that he gets more utility from
the food and shelter.’

Now appears to be a very strong defence, but let us see what effect it
this
has on our ‘law’. The ‘law’ states that successive increments of income bring
progressively diminishing utility because they are used to purchase goods
which bring successively smaller amounts of utility. How do we know which
goods bring the consumer most utility? We do this by observing which goods
he purchases when his income is low. But if we state that the consumer will
choose first the goods which bring him most utility and define the goods
which bring him most utility to be those which he chooses first, we have
7
 ,

*J+ TMF INTERMEDIATF THEORY OF DEMAVD

done nollimg more than to diicoierouroivn
detinmons, i.e have uttereda
tiulologj vvhieh 15 compatible vetlh my slate of
the universe
Clearl) sucli a defence of the ‘law
reduces it to a useless taulologv
it
does not mitter what eonsumers do, no
inaltcr how masochistic, selness’or
selfish, rational or imtional they ate,
all behaviour is compatible with
this
nutolog) Clcirly, if the 'law’ is to tell us something
about the world we
need to ha\e an independent indicator of utility or
satisfaction Ifwe ha\e
this, wc can then say 'gootJ A
Rues this consumer more satisfaction than
docs good B (at equivalent prices) and wc therefore
predict that he will
buy good A when his income is low and good B only when his income
rises
In the absence of such a measure of utility we have no theory at all
Until
a measure has been put forward and until the predictions of the theory have
been tested against observed consumers’ behaviour it is c\tremcly mis
leading even to talk about the law of diminishing marginal utility of
income, in fact it is not even clear that wc are making a positive (i e
potentially testiblc) statement when wc say that the marginal utility of
money is declining ‘

I* II II not my purpoiein (hiiteccon to reject the bypoitiMii of dim nnhingmarpnalulil ly

of income All thai [ wish to do in (he test u to make the general poini that unlets tteare very
careful Me can argue in a circle thereby making ouratatemeni true by defnition which robs
ilofall empincat content (i e it then says noihingabout the world) and makes u unmterescmg
for positive economia
In fact 1 1 thai the theory it either empty or wrong hfy mam reason for
u my pnvate opinion
holding this is that consumers only have clearly specified lattes m the neighbourhood of
view
their present eonsumpiion position A man vvith an income of £500 per year has no idea what
he would end up consuming if his tncorae rose to £5 000 nOr of what value he would come to
place on the goods he has learned to consume (many of which he would have been toully un
aware of when his income was £500 because for example he read different magazines) We
learn what our tastes arc or we develop them as our consumption possibilities develop It

seems to me therefore that it is quite possible that as our ignorance recedes we will F nd our
selves encountenng increasing marginal utility of income This theory has at least two predic

Uons which conflict with those of tradiUonal theory (1) according to this theory a man may
work longer hours as hii income rises while acconl ng to traditional theory he mint work fewer

houn (assuming leisure not to be an Inferw good) (2) according to this theory consumption

patterns will not be reversible a man with an income of £500 will consume
one bundle of
and he
goods but raise his income to £5 OOO for a few years and then lower it back to £500
will consume quite a diflerenv bundle of goods because of what he learned when his income
washigh accordingtotrad tionaltheorybcmustgobacktohisonginalpatternofconsumpuon
an
The whole question of the hypothesis of the dimimshing marginal utility of income is
advanced one that cannot be gone at this point but what has been said should be enough
mW
hypoihes s
to show (1) that we cannot accept it as obviously correct and (2) that alternative
will have confl cting testable implications and a choice can be made between
the theories on

dence rather than on a pnort reasonableness

 THEORIES OF HOUSEHOLD BEHAVIOUR 195

REVEALED PREFERENCE
We now pass to a discussion of a modern theor>' of household behaviour
called revealed preference. We shall use this theory to distinguish between
what is called an income and a substitution effect of a price change, which

is a distinction we have not yet made, and then to give an alternative

derivation of the downward sloping household demand curve based not on
the assumption of diminishing marginal utility but rather on the assumption
that the household behaves consistently.

A Change in Relative Prices The Income and Substitution Effect

:

We now wish to invesdgate in more detail the consequences of a change in

relative prices. We consider a household whose initial budget line ab in is

Figure 15.7 and assume that initially it chooses the combination indicated

>
o

Commodity X

Fig 15.7 A fall in the price of commodity X.

by the point a {Ov of Y and Os of X). Now assume that the price of X falls,
themoney price of Y and the household’s income remaining constant. The
new set of possibilities open to the household is indicated by the budget line
 196 THE INTERMEDrATE THEORY OF DEMAND
ac Assume that the household now chooses the combination {Ot of y and
Oil of X] from
^
all of those available to it More of both
and Y are X
consumed
Now It IS possible,
and it turns out to be useful, to analyse this movement
from a to ^ into two constituent parts, one called
the substitution
EFFECT and the other called the income effect It will be noted
that this
fall in the price of X does have something
of the same cifect as a change in
income because it makes it possible for the household, if it so
wishes, to have
more of all commodities In Figure 15 7 the price fall makes the total shaded
are newly available, combinations indicated by points within this area
were
not available at the original set of prices Points within area indicate (!)
newly available combinations containing more Y but Jess X than was being
consumed at a Points within area (2) indicate newly available combinations
containing more X but less Y than was consumed at a Points within area
however, indicate newly available combinations containing more X and
(3),
more Y than was consumed at a
We are considering a fall in the pnee of X and the movement of the
household from a to ^ What we now wish to do is to break up this move-
ment into two parts, one doe to the pure change in relative prices and the
second due to the income effect (the household being able to have more of all
goods) To do this we can imagine ourselves reducing the household s in
come until there is no income effect of the change in pnee - e until it is
i

no longer able to consume more of all goods than it was consuming origin
ally To do this we reduce its income until al the new tet of prices, a is just
able to buy as onginai bundle of goods Graphically, reducing income wah
relative prices constant at ifia(Oc) means that the line ac slides inwards to

wards the origin, parallel to itself until a passes through the point o This
IS shown in Figure 15 8 by the broken line a c

The total effect of this price change may now be broken up into two parts
1 X becomes cheaper relative to Y butsimultaneously income is reduced
so that the household can just buy the bundle of goods that it was buying
originally The household may move from a to some other point on o f say ,

the point e The change m the quantity of A purchased (from Oq to Om) is

defined as the substitution effect, it is the effect solely of a change in
relative

pnees Generally we expect the household to buy more of the relatively

cheaper good, X, and less of the relatively more expensive good, Y
Relative prices are now held constant al their new level
and the
2
household s income is returned
Graphically the budget line is moved
outwards, parallel to itself, unul it is returned to the position ac The
move-
final position p defined as
ment from the intermediate position € to the is

the INCOME EFFECT, mn more of is bought because, with relative pnees

constant, the household s income rises

 THEORIES OF HOUSEHOLD BEHAVIOUR 197

In practice, of course, the household moves directly from a to Our

theory, however, says that demand is a function both of relative prices and
of real income. A change in the money price of one good changes both
relative prices and real income and what we have done here is to separate
out the effects of the two changes.

Commodity X

Fig 15.8 The income and the substitution efTect.

THE DIRECTION OF THE SUBSTITUTION EFFECT^

We are now in a position to derive the basic prediction of demand theory:
the substitution effect can never cause a reduction in purchases of the com-
modity whose price has fallen, or, in slightly more technical language, the
substitution effect is non-negative.

To derive this prediction we introduce the basic behavioural hypothesis:

the household always acts consistently. This makes it necessary to define
which we are attaching the word ‘consistent’. Once we
carefully the idea to
know the household’s income and the market prices facing it, we know the
combinations of goods which it is possible for the household to choose these ;

will be all the points on, or inside, its budget line. In Figure 15.9 the

1 The rest of the chapter may be omitted ivithout affecting one’s ability to follotv any of the
analysis in the rest of the book.
198 THE intermediate THEORY OF DEMAND
combmauons available to the household are those within or on the
border
of the shaded tnangle Oai (For simphcity we shall speak of the points
in
the triangle Oab )
The household must rdioose out combtuatiou out of all
the ones open to it \Vc denote the actual combination chosen by
z and the
whole of combmations rejected by S {S thus stands for all points
set the m
triangle Oab other than the point z) We now define consistent behaviour
as follows if the household chooses the combination z m preference to all the combina
tions in 5 it will never subsequently choose any combinaUon from S in a situation in

ukich z ij also available

Let us now consider a X

pnee of accompanied by a sufficient
fall in the
reduction in the household s income so that it can just purchase its onginal
combination of goods indicated by the point z In Figure 15 9 the budget

to a c Now the household can choose any point on the

line moves from ab
the segment za of this new budget hne
budget line e c Consider first

Points to the left of z were available m

the first situation they are therefore
in favour of z Since z is still
part of the set S which was originally rejected
out comistewcy assumption
available in the new situation it follows ftom
that z cannot be rejected in favour ofa point
on the segment o z On the
the segment 1« outside of the onginal set S they
other hand points on zf
 THEORIES OF HOUSEHOLD BEHAVIOUR 199

were not available in the original position and there is thus no inconsistency
in choosing any point to the right of z in preference to 2 Thus the household
.

either stays at z or it moves to the right along the segment zc' which means

that, either it consumes the same amount of X (the good whose price has
fallen), or it consumes more of it. We conclude therefore that, given our
assumptions, the substitution effect can never lead the household to buy less
of the commodity whose price has fallen; it either buys the same as, or
more than, it was buying before the price fell.

The Slope of a Household's Demand Curve

We have seen that, when the price of X falls, there is an income effect as
well as a substitution effect. If X has an income elasticity of demand
greater than zero (i.e., it is not an inferior good) the income effect will lead
to an increase in the purchases of X (the point jS will lie to the right of the
point e in Figure 15.8). Since the substitution effect cannot be negative, an
income elasticity of demand larger than zero is sufficient to ensure that more
X will be purchased when its price falls (i.e., the point ^ lies to the right of
the point a in Figure 15.8).

This means that the demand curve, which shows the

quantity of X demanded at alternative prices of X,
will slope downward to the right.

On the other hand, a negative income elasticity is not sufficient to ensure

that the household will buy less of the good when its price falls; this will

only happen if the negative income effect is large enough to outweigh the
substitution effect. In Figure 15.10 we have illustrated two cases ofa negative
income effect. The initial is ab and the household’s initial posi-
budget line
tion is at a, purchases of Om. The price of X now falls so that the
X being
budget line shifts to ac. To measure the substitution effect we reduce income
until the household is just able to buy its original combination of X and Y
indicated by the point a; this gives the budget line a'c'. Assume that it now
buys the bundle indicated by e, purchasing mn more X on account of the
substitution effect. Assume further that X is an inferior good so that the in-
come effect will be negative, and examine two possible cases. In the first
case the income effect causes the household to move from e to on the new
budget line ac. It buys np less X on account of the income effect, but this is
not sufficient to overcome the substitution effect, and the household is observed

to increase its purchases of X

by mp as a result of the fall in its price. In the second
case, the income effect is and the household moved from
stronger e to u as

a result of the rise in its income, thus reducing purchases of X by an

 200 THE INTER^fEDrATE THEORY OF
DEMAND
amount nq. The household mil be observed
to reduce its
to Oq as a result of the fall tn the price *
purchases of X from Om

We conclude, on the assumption that the household acts

therefore, that,
consistently as defined above, demand cur\e for any product will be
its

downward sloping unless the income effect is negative (the good is inferior)
and of sufficient magnitude^ to overcome the substitution effect ^

1 This case in >vh]ch a Tall in pnee Ivvven the consumption of a commodity is called (he
Giffen case after (he Victorian economist Sir Franas Giffen who is supposed to have observed
an actual example of it The significance of this Giffen case is considered at some length in the
next chapter
2 The income effect of a given pnee change depends partly on the importance of the com-
modity in the household s budget If, for example the household s income is 1,000 per annum
and it IS spending £\0e I per cent of its income) on some commodity then a 50 per cent
(i ,

fall in the pnee of the commodity is equivalent to allowing it to purchase the original bundle
and railing his income by £5 per annum, i e , by ^ of 1 per cent
3 The theory of demand given in the previous sections is based on the theory of Recratid
Preferince There is an alternative demand theory formulated in terms of indifference
CURVES The economics specialiit vvilt often encounter indifTerence curves and he will need to
study the theory based on them at some stage, but it is not necessary to do so before completing
this book For those who have studied indifferHicc theory already, it should be pointed out

that the defimtion of the substitution effect used in the indifference theory is slightly different
 THEORIES OF HOUSEHOLD BEHAVIOUR 201

The terminology of the reaction to demand is not fully standardised and

it can be rather confusing to the student. In Table 15.1 we present a table
of equivalents which should help the reader to understand overlapping
terminology.

Table 15.1

TERMINOLOGY USED TO DESCRIBE THE REACTION

OF THE Q.UANTITY DEMANDED TO A CHANGE
IN INCOME
Numerical measure Verbal description Income Type
of income elasticity of income elasticity effect of good
ey < 0 negative negative inferior
o V V positive but inelastic positive superior
1 < ey < CO positive and elastic positive superior

CONSUMER RATIONALITY
It is often said that economic theory is based on the assumption of rational
behaviour. There are three points that need to be made in this context.
First, rational behaviour is often used in a value sense to say that certain
forms of behaviour (called rational ones) are better than others (called
‘irrational’ ones). Second, if rational is meant in a positive rather than a
normative sense, then it is very difficult to establish by casual observation
that certain behaviour is not rational. Third, if by rational we mean
‘consistent’ behaviour then such an assumption is necessar)' not only for
economics but for any subject which attempts to build theories about
human behaviour. Let us consider these points one at a time.
Positive economics cannot assert that certain forms of consumer satis-
faction are rational while other forms are irrational. We cannot demon-
strate, for example, that it is sensible or rational for the consumer to derive

from the one used above. In revealed preference theory the price of X is lowered and income
isreduced until the original combination of goods can just be purchased', any increase in the purchase
of X is now ascribed to the substitution effect. In indifference theory the price of X is lowered
and the household’s income is reduced until the original level of satisfaction {or utility) can just be
attained; any increase in the purchase of X is then ascribed to the substitution effect. In re-

vealed preference theory it is shown that the substitution effect must be positive. The great
advantage of using the definition in revealed preference is that the>’ are operational. We saw
in the previous footnote that the income effect of a price change can be readily calculated in
revealed preference terms, it is quite immeasurable in indifference curve terms.

7*
 202 THE INTERMEDIATE THEORY OF DEMAND
from consuming a good in isolation and
satisfaction
silly or irrational for
him to derive satisfaction from showing ihe
good off to his neighboun and
boasting about how much he paid for it
Posiiivc economics can be used lo
analyse any form of known consumer
behaviour no matter how immoral
misguided or chauvinistic it may seem to be to the
person conducting the
analysis
by rational behaviour we mean something like choosing
If
appropriate
means to amve at given ends, then it is by no means easy to discover by
casual observation that a certain behaviour pattern is irrational
It is some
times said, for example, that it is irrational for the consumer to judge
the
quality of an article by its price Certainly the consumer who docs this can
occasionally be very greatly misled The
consumer can, however,
typical
have only very imperfect knowledge of the wide range of articles from
which he has to choose If he trusts his own amateur technical knowledge
he may sometimes go very badly astray It may well be that price is in fact
the best single indicator ofquality available to the consumer Such an index
isundoubtedly less than perfect but a consumer who used it consistently
might do better on average than a consumer who used his own judgment
based on incomplete and often mistaken technical knowledge If this were
the case, we would have to conclude that the consumer whojudged quality
by price was rational
in the sense defined, even though we might sometimes
observe him rnaking mistakes The moral to this story is that it is very
difficult,by looking at the very best means of making a few choices to
decide what is a rational means of choosing when one is faced with the need
to make a great many choices based on an imperfect knowledge of a com*
plex world
Finally, if by rational we mean consistent (i c the individual docs not
,

have contradictory preferences), then any theory of behaviour necessarily

assumes this, for, if we make inconsistent assumptions about the consumer s
preferences, we will have no theory at all but a set of contradictory pro
positions Thus, to say that we assume consistency {i e rationality) is only ,

to say that we have a theory The extent to which people do behave con-
sistently in the real world is, as we have already seen Chapter 1 (see m
page a matter than can only be established by careful, detailed
10),
‘
observation

contained in it
1 The beginner need not worry very much about this paragraph The idta
lot of silly misguided criticism of economics The
advanced
does however help to dismiss a
however that the use of the word consistent here is different from that
student should notice
III ihe revealed preference theory on page J97
The general rationality principle in this section
IS that the person should act consistently ta
reipecl la ihe saruhlt] that do in/uence Ms acUon This
that do inflo
very general sutement u given empincal content when we specify the variables
dice his action In revealed preference theory, as in other theories of demand, the
number of
 THEORIES OF HOUSEHOLD BEHAVIOUR 203

variables is severely restricted and all the consumer is assumed to care about is the quantities
of the various commodities that he consumes. The consumer is then assumed to take consistent
decisions with respect solely to the quantities of the commodity available to him. To take an
example the consistency assumption on page 197 would be refuted if the consumer cares not
only about what he gets but also about who sells it to him, and in situation (1) he took z and
rejected the other points in because he liked the man selling the bundle z and in situation (2)
.S'

he took some other point from S even though z were available because he did not like the new
salesman for the bundle z. This does not violate the general rationality principle discussed in
the text above, but it does violate the revealed preference consistency assumption because the
former amounts to assuming that nothing influences the consumer other than the quantities
he is able to consume.
 CHAPTER 16

THE THEORY OF DEMAND:

measurements and tests

In this chapter we shall consider first the question of testing the theory of
demand in a very general way Then we shall consider the problem of
trying to measure quantitatively the influence on demand of changes m
each of (he factors which affect it Finally, we go on to discuss the entical
importance to economic theory of such quantitative estimates

CAN DEMAND THEORY BE TESTED?

The mam problem encountered m answering this quesuon is to discover if
the theory makes any predictions which are precise enough to be able to
be tested To study this we shall consider the relation between demand and
each of the vanables that are supposed to influence it, but before doing
thiswe must dispose of one common but misguided means of testing the
theory

Testing by Observing the Behaviour of Isolated Households

ItIS common, particularly amongst students to cnticisc demand theory by

arguing that, from personal experience, one know that people do not always
behave m
the rauonal manner assumed by the theory Cases in which con
sumers do not behave consistently can usually be multiplied from casual
observation of the behaviour of one’s self and of one’s acquaintances Such
observations, when carefully documented, do refute the theory that all
households always behave as assumed by the theory Usually, however, m
positive economics we are not interested in the behaviour of every single
household but rather in their aggregate behaviour as shown, for example,
by a market demand curve AH of the positive predictions developed in
 THE THEORY OF DEMAND: MEASUREMENTS AND TESTS 205

Chapter 11, for example, depend on some knowledge of the shape of the
relevant market demand curve.
The existence of a downward-sloping, relatively stable, market demand
cur\'e does not require that all households behave as is assumed by the
theory all of the time. Such fully consistent behaviour on the part of every-
one at all times is sufficient but not necessary'* for a downward-sloping
market demand curve. ^ Consider two other possibilities. First, some house-
holds may always behave in a manner not assumed by the theory. House-
holds whose members are mental defectives or have serious mental
The inconsistent or erratic behaviour
disturbances are obvious possibilities.
of such households will not cause demand curves to depart from their down-
ward slope, provided these households account for a minority of total
purchasers of any product. Their erratic behaviour will be swamped by the
normal behaviour of the majority of households. Second, an occasional
irrationality or inconsistency (such as ‘buying a lot of strawberries just for
the heck of it even though their price hsis gone up’) on the part of every
household will not upset the downward slope of the market demand cur\'e

so long as these isolated inconsistencies do not occur at the same time in all
households. As long as such inconsistencies are unrelated across households,
occurringnow in one and now in another, their effect will be swamped by
thenormal behaviour of the majority of households. The conclusion reached
may be summarised as follows.
The downward slope of the demand curve requires
only that at any moment of time most households are
behaving as is assumed by the theory; this is quite
compatible with behaviour contrary to the assump-
tions of the theory on the part of some households all
of the time and on the part of all households some of
the time.^

If we cannot test the theory about the behaviour of market demand by

observing the isolated behaviour of individual households, we must now
consider what can be learned by observing such things as total demand for

a product, the price of the product, other prices, and the total income of all
consumers, which are the variables that appear in our theory of market
demand. This is best done by considering in turn the relationship between
demand and each of the influencing variables.'^
1 If you have forgotten this important distinction you should re-read section 2 on page 34.
2 We are here ignoring the possibility of'Giffen Goods’ described on page 212 of this chapter.
3 We have relied here on intuitive argument but this result can be shown rigorously by the
use of mathematics. It is desirable at this stage to review pages 10-13 on the prediction of
individual and group behaviour.
4 These hypothesised relations are described in detail in Chapter 7.
 206 THE INTERMEDIATE THEORY OF DEMAND

The Relationship between Demand and Tastes

A change tn tastes causes demand to change This proposition is
not really testable
unless we have some way of measuring a
change in tastes If we had then
we could imagine saying something like ‘Each time tastes
have changed
m favour of this commodity demand has fallen, so we must reject the view
than a change tastes m m
favour of this commodity causes an increase in
thedemand for it This we cannot do because we do not have an indepen
’

dent measure of taste changes What we usually do is to infer

taste changes
from the data for demand We make such statements as ‘In spite the of nse
in price, demand increased so there must have been a change m tastes in
favour of this commodity More generally, we are likely to account for all
’

of the changes, in demand that we can vn Itnns of piites and incomes and
then assert that the residual must be due to changes in tastes (and to errors
of measurement) This does not concern us unduly because wc are not
greatly concerned to establish precise relations between tastes and demand
and wc are prepared to take it as obvious from even the most casual obser
vation that tastes do mnuence demand
The we cannot identify those changes in demand that are due
fact that
to changes in tastes, because we cannot measure taste changes does how
ever cause us much trouble when we come to consider the relation between
demand and other factors Whenever we see something happening that docs
not agree with our theory » is always possible that a change in tastes
accounted for what we saw Say, for example, incomes and other pnees
were known to be constant, while the pnee of some commodity rose and at
the same time, more was observed to be bought This gives us observations
such as the ones illustrated in Figure 16 I(i) It may be that the demand
curve for the product has a shape similar to that shown by the broken line
D in Figure 16 l{ii) but it may also be that the nse in pnee coincided with
a change in demand curve shifted from D" xo D" in
tastes so that the

Figure 16 1 we only have two observations we will not be able to

(iii) If
distinguish between these two possibilities since we have no independent
way of telling whether or not tastes changed If however we have many
observationswc can get some idea vX where the balance of probabilities lies
between these two situations If we have 26 observations (say the price
changed each week over a period of six months) which having removed the
illustrated in Figure
of other prices and incomes,^ look like those
influence

16 2 wc have to stretch the point a great deal to avoid the conclusion

will
that the evidence conflicts with the hypothesis of a
downward sloping
demand curve
statist
This can be done through muHipIc regresston analysis or other more
sophisucated
I

cal techniques in the manner alluded to in Chapter 3

 THE THEORY OF DEMAND; MEASUREMENTS AND TESTS 207

Price

Fig. 16.1 Alternative interpretations

of two price-quantity observations:
either the demand curve slopes
upwards or it has shifted.

Of course, we can always explain away this apparent refutation by saying

that tastes must have changed in favour of this commodity each time that
its and against the commodity each time that its price fell. This
price rose
‘alibi’ away a single conflicting
can certainly be used with effect to explain
observation, but we shall be uncomfortable in using the same alibi 26 times
in six months, and we shall begin to suspect a fault in the hypothesis that
demand and price vary inversely with each other. What we now have is a
problem in statistical testing.^ We are not prepared to throw away a theory
after only one conflicting observation, but we are prepared to abandon it as
soon as the probability of making a given set of observations if the theory is
correct diminishes drastically. Thus, statistically, the theory is testable. If

I See page 46.

 intermediate theorv of demand
demand curves are observed lo slope up%vani so that
we have to make ver>
special and unlikely assumptions about how
tastes just happened to ha\c
changed each time prices changed we will prefer to reject our
theory that
demand curves slope downward Fortunately, there is as we shall
see a
great deal of evidence that demand curves do slope
downward The pre
dictions of the theory have - with one notable exception - always been
found be in agreement with the facts
to
This problem of testing a theory that contains an unmeasurable vanable
is a frequently encountered one and wc shall encounter
it again in the

theory of the supply of factors of production (see Chapter 32) and it is

worth emphasising the conclusions reached above

Fi£ 16 2 Twenty six price quantity

observations for which one
explanation that the demand
curve slopes upward is very much
more likely than the alternative
explanation that the demand
curve has shifted between each
of the 26 observations

Quantity

1 If we have only a single conflicting observation then we cannot rule

out or even regard as very unlikely the possibility that the vanable we
cannot measure changed \n such a way as to produce the apparent conflict
2 If we have a mass of evidence and we are testing statistically the
h> potheses of the relation between the things we can measure then we can
assess the probability of the weight of unfavourable evidence being pro
duced by offsetting movements in the variable we cannot measure If we
frequently we shall
find we have to invoke the excuse always or even very
legard vVi-n -aS- wJi. pjrefec tn cejedL the hypothesis about the

way m which measurable vanables are related

‘

We can strengthen our theory greatly if we make the assumption that

theories If we accept
1 You should now re read pages 16-17 and 50-3 on «mng economic
that a single observation can refute a theory then the
nonmeasurable vanable causes us unsur
it makes a single observation
refutation impossible If we reject
mountable troubles because
the idea of single observation refutations we have touch less trouble We are not upset i

be invoked to explain away sortiethitig odd but

occasionally the unmeasurable vanable has to
we regard our theory as unsausfaetory if it can only be saved by repeated use of the explanation
that the unmeasurable vanable must have changed in an offsetting manner
 :

THE THEORY OF DEMAND: MEASUREMENTS AND TESTS 209

changes in tastes are independent of changes in the other variables which

influencedemand. This assumption means that a change in tastes in favour
of a commodity (a rightward shifi in the demand cur\'e) is just as likely to
occur when prices rise as when more formal terms, shifts in
prices fall or, in

demand curves (due to taste changes) are unrelated to movements along it.
This allows us to calculate precisely the probability that offsetting changes
in tastes could actually occur a large number of times. The probability in
the previous example of 26 offsetting changes with tastes just happening to
change commodity when prices rose and away from it when
in favour of the
prices fell being one chance in about 70 million. Such odds are so low* that
we would prefer to reject the hypothesis that what \s'e observed in Figure
16.2 was due to changes in tastes, and so reject the hypothesis that the de-
mand for this particular good slopes downward.

The Influence of the Prices of Other Commodities

In Chapter 7, we made a distinction between goods that are complements

to one another and those that are demand for
substitutes.^ Consider the
commodity X. This demand will vary inversely with the price of a complemen-
tary commodity (i.e., when the price falls demand for X will rise), and will
vary directly with the price of a substitute commodity (i.e., when the price falls,
demand for X will fall). There may also be a group of commodities, varia-
tions in the price of which leave demand for X unchanged. These commodi-
ties lieon the boundary' .that divides the goods that are substitutes for X
from those that are complements to X.
These three reactions - demand for A' rises, falls, or remains unchanged
when the price of some other good varies - cover all conceivable possibilities
there is nothing else that could possibly happen. So far we merely have a
set of labels to attach to all possibilities. We do not have a testable theory
unless we have a way of deciding in advance which goods are substitutable for,
and which are complementary' to, X.
We can do this sometimes from technical knowledge alone. This is par-
ticularly important if we consider the demands for inputs in production.
Steel plates, electric welders and welder operators are complementary'.
1 If tastes changed each week there is a 50-50 chance that they changed in favour of or
against the commodity. Then there is a 50-50 chance that tastes changed in the direction to

offset the price change. In the second week there is also one chance in two. The chances that

they changed the ‘right’ way in both weeks are i.i=i and the chances that they changed the
right way for 26 successive weeks are 1/67,108,864.
In fact the odds are much lower. The quoted odds are only for taste changes in the offsetting
direction' in fact they would have had to be of sufficient magnitude to outweigh the reacdon to
the price change in each and every case.

2 See page 84.

 THE THEORY OF DEMAND; MEASUREMENTS AND TESTS 211

been observed to fall from -70 to -40 we are pretty safe in predicting that a
rise in income next year will be met by a less than proportionate rise in the

demand for that product.If, on the other hand, the income elasticity of

demand and electricity are both observed to have been above 2.0
for cars
for several years and also to be rising, it is fairly safe to predict that rises in
income in the next few years will be met by more than proportionate rises
in the demand for cars and electricity. The fact that these elasticities are
observed not to change rapidly or capriciously allows us to predict into the
near future from a knowledge of the level and direction of change of exist-
ing income elasticities.
The second fact that helps us give empirical content to our theory' is the
observation that with respect to broad categories of consumption expendi-
ture households in any one Western country behave in a fashion roughly
similar to those in other Western countries. (Indeed it is not even clear that
the qualification Western is necessary.) At low levels of income food tends
to have a fairly high income elasticity of demand but as the level of income
rises, the income elasticity of demand for food tends to fall well below unity,
so that very little of any additional amount of income gets spent on food.
The phenomenon has been observed in every growing country that has
approached the levels of income currently enjoyed by the countries of
Western Europe. Thus we can confidently predict (1) that as long as pro-
ductivity growth continues in agriculture the long-run drift from the land
will continue in Western countries (unless they have a large export market
for their agricultural goods), and (2) that when other countries of the world
succeed in achieving sustained positive rates of growth they will encounter
within a predictable time horizon the problem of a declining agricultural
sector. Households in the United States currently enjoy an average level of
income higher than that enjoyed by households anywhere else in the world.
By observing broad changes in the pattern of demand in the United States
we can get some idea what will happen to demand in other countries 20 or
30 years from now when they achieve the level of per capita income
currently being enjoyed in the United States. The most significant phe-
nomenon in the United States in the post-war period is the decline in the
income elasticity of demand for durable consumer’s goods and the rise in
the income elasticity of demand for services. This has necessitated a con-
tinued transfer of resources in the United States that is not being accomp-
lished with perfect smoothness. If there is stability in household behaviour
in thesebroad patterns across countries then the countries of Western
Europe can look fonvard to similar pressures on the pattern of resource allo-
cation within a generation or two. At a less generalised level it is possible to

hazard the guess that those responsible for providing electricity and tele-
phones in Britain would not have so frequently underestimated the rise in
 212 THE INTERMEDIA.TE THEORY OF DEMAND
the demand for then prodntts rf they had .tnd.ed
the .neome elashctte.
ol demand for these prodncts in the United States
over the last few decades
The incomes of VV estern countries are doubling
every 30 j ears or so Thus
over such penods of time changes in
income exert a major influence on
changes m demand Indeed, some hnoivledge of vshat income
elasticities aic
and are likely to be 15one of the most potent tools at the economist
s com
mand for predicting the future needs of the economy in myriads ordilTetent
aspects

Demand for a Commodity and the Commodity's Own Price

In this case we have, as we saw m the previous chapter, a much more defi
nite predictionPerhaps it is best to develop it historically Eatl^ demand
theory ignored the income effect of a price change* and thus predicted
demand curves must always slope downwards This became kno^>n
that all
as the law of demand the price of a product and the amount demanded
\ary inversely with each other
Great interest was therefore attached to an apparent refutation of this
law reported by the Victorian economist Sir Francis Giffen He is reputed
to have observed and documented that an increase in the price of potatoes
dunng an Insh potato famine led to an increase in the consumption of
potatoes If this observation is correct (i c , if Giffen really made it, about
which there now appears to be some doubt, and if there is onl> a low proba
measurement of a magnitude sufficient to account for
bility of errors in his
the result) it docs refute the hypothesis that all demand curies alwais slope
downwards Docs it refute the modem theory of demand ’ The answer is

‘No because this is just the type of rare exception to the normal case that
IS envisaged by the modern theory Pontoes accounted for a
\ery large frac
tion of the total expenditure of the households affected by the price change
If potatoes are an inferior good so that the income effect is negative it is
possible that the large negative income effect overcame the normal sub
^
stitution effect
Thus the modern theory of demand only makes an unequivocal prediction
when we have extraneous information about income elasticities of demand
Since incomes change continuously due to economic growth we are
for-

having such information about most commodities \Vhen, as

in
tunate m
is positive
the case of most commodities, we know that the income effect
(income elasticity of demand exceeds zero) we can predict in
advance that
the quanUty demanded of the commodity wil! vary inversely with its price

the income effect to be negative (i e , we know that we

are
Where we know
dealing with an inferior good) we cannot be sure of this result The only
1 See pages 195-7
2 See pages 198 9
 THE THEORY OF DEMAND; MEASUREMENTS AND TESTS 213

thing we can say is that the smaller the proportion of total expenditure
accounted for by commodity, the less important is the income effect and
this

the more likely are we to get the normal result of price and quantity varying
inversely with each- other. Indeed from the fact that only one exception of
this type has ever been seriously alleged, and that we are not even sure of

the documentation of that one, we can say that evidence suggests that even
in the case of inferior goods the downward-sloping demand cuiv'e appears
to be the more typical case. Finally, if we have no knowledge about the
income effect we can still hazard a probabalistic statement. The great weight
of existing evidence suggests that if with no prior knowledge you had to
guess whether the demand curve for some commodity X was downward or
upward sloping the former choice would be the odds-on favourite.
A very different set of possible exceptions has been suggested in some of
the theories of demand developed in the last 20 years. Several possibilities
have been suggested and they all depend on the assumption that the con-
sumer is influenced by factors other than the position of his budget line.^
We shall consider one of these possibilities as an example.
Assume that a household’s satisfaction depends not only on the quantity
of the commodity that it consumes but on the price it has to pay for the
commodity. The household may, for example, buy diamonds not because
'its members particularly like diamonds per se, but because they wish to

show off their wealth in an ostentatious but socially acceptable way. In

the words of Thorstein Veblen, ‘they indulge in conspicuous consumption,
and when their neighbours copy them they are indulging in pecuniary emu-
lation'} The household values diamonds precisely because they are expen-
sive; thus, a fall in their price might lead it to stop buying them and to
switch to a more satisfactory object of ostentatious display. If enough house-
holds acted similarly, this could lead to an upward rather than a downward-
sloping demand curve for diamonds.^
A great deal of work has been put into the statistical study of demand and
in the next section we shall consider some of the problems involved. In the
1 See note 1, page 202.
2 These phrases are drawn from a book which is worth browsing because of its many in-

sights which are still relevant to today’s world. Thorstein Veblen, The Theory of the Leisure

Class (1899) now available in paperback.

3 To the best of my knowledge, no-one - with the possible exception of Giffen - has ever
observed an upward-sloping market demand curve. A. moment’s thought about the
statistically

industrial uses of diamonds and the masses of lower-income consumers who could buy dia-
monds only if they were sufficiently inexpensive would seem to suggest that upward-sloping
demand curves for individual households are much more likely than upward-sloping market
demand curves for the same commodity. You should be reminded of the discussion on pages
202-3 about the ability of the theory of the downward-sloping demand curve to accommodate
odd behaviour on the part of a small group of households (this time the odd group is the rich

rather than the mentally defective or deranged).

214 THE INTERRfEDIATE THEORY OF DEMAND
meantime we can note that in sf»te of the possible exceptions given above
a mats of evidence has accumulated suggesting that most (virtually
all)
demand curves do in fact slope downwards For practical purposes, we can
regard the hypothesis of the downward-sloping demand curve as useful and
m conformity with the evidence
Wecan go, and economists have gone, farther than this We have a
reasonable idea of the value of demand elasticity in many cases Thus, not
only do v\c know that demand curves slope downward, we also know, vnthm
a reasonable margin of error, how steeply many of them slope downwards
Reasonably precise knowledge about demand curves is a necessity if we arc
to make real-world applications of the demand and supply theory developed
in thisbook If wc knew nothing at all empirically about these curves, the
theory would be devoid of any real world application If wc do have this
knowledge, then we can predict in advance the efTccts of changes m many
factors such as taxes, costs, the amount of competition in a particular

market, and so forth The more accurate u our knowledge of the shape
of these curves, the smaller will be the margin of error in such predictions
As an illustration of the kind of knowledge that wc do have, estimates of
the reaction of quantity demanded to price and income changes is given for
a few selected commodities m the following Table ^

Table 16 I

PRICES AND INCOME ELASTICITIES FOR SELECTED

foodstuffs in the UK
PereaUa^e of
variation accounted

for by thangts in
income, own price.

Income Price and pnees of

elastiaty elasticity related goods
Commodity

79 53
Flour
34 41 64
Home-produced beef and veal
70 47 70
Home-produced mutton and lamb 1

35 88 71
Bacon and ham
1 17 27 33
Poultry
71 69 69
Cream 1

92 9? 52
Oranges
44 74
Sugar 09
Meals away from home 2 39 Not estimated Not esumated
- -v-
SOURCE Kicnara none
Vol I
Kingdom 1920-38 (Cambridge Owversiiy Press 1954),
 THE THEORY OF DEMAND: MEASUREMENTS AND TESTS 215

MEASUREMENT OF DEMAND
The above discussion leads us directly to a consideration of the problems of
the measurement of demand. Three of the most famous names in this field
are Henry Schultz of the University of Chicago, Richard Stone of Cam-
bridge University and Herman Wold of the University of Uppsala, Sweden.'
Many difficult problems must be overcome before we can obtain even an
approximate measure of the response of demand to price changes from
empirical data. Most of these problems are of a very technical nature and
cannot be appreciated without a knowledge of statistical theory. We must,
however, mention two of the most important of these.

Everything is varying at once: When we observe changes in market

demand over time it is almost invariably the case that the various influences
that are assumed to affect demand will all be changing. What, for example,
are we to make of the observation that the quantity of butter demanded rose
by 10 per cent over a period in which household income rose by 5 per cent,
the price of butter fell by 3 per cent and the price of margarine rose by
4 per cent? How much of the change is due to income elasticity of demand,
how much to price elasticity and how much to the cross elasticity between
butter and margarine.^ The answer is that we cannot tell if all we have is
this one observation. If, however, we have a large number of observations

showing say quantity demanded, income, price of butter and price of

margarine every month for two or three years we can hope to discover the
separate influence of each of the variables. This general problem is dis-
cussed in Chapter 3 but the development and elaboration of the solution
belongs to a course on statistics. The most frequently used technique for
estimating the separate effect of each of these variables on demand is called
multiple regression analysis but in certain cases more complex techniques
become necessary.

The identification problem: Another characteristic problem arises

when we try to estimate the relation between demand for a commodity and
the price of that commodity from data for quantity purchased and price. In
this case we are interested in the reaction of the quantity demanded to only

1 H. Schultz, The Theory and Measurement of Demand (University of Chicago Press, 1938;

reprinted, 1957); R. Stone, The Measurement of Consumers' Expenditure and Behaviour in the United
Kingdom, 1920-38 (Cambridge University’ Press, 1954); H. Wold and L. Jureen, Demand
Analysis: A Study in Econometrics (John Wiley, 1953).

2 If you are not completely dear on the meaning of these three important elasticities you
should review pages 118-121.
216 THE INTERMEDIATE THEORY OF DEMAND

one of the \anables that influence it and the problem would seem relatively
simple As we shall sec, this is not the case The solution to this problem is
beyond the scope of this book but the present discussion should serve as a
warning against the very common practice of trying to infer something
about the shape of the demand curve from price and quantity data alone
The nature of the problem can be illustrated simply We start by assum
mg that all situations that we observe in the real world are equthbnum ones,
m the sense that they are produced by the intersection of demand and supply
curves If, as in Figure 16 3(i), the demand curve stays put while the supply

curve moves up and down, possibly because of crop variations in some

agncultural commodity, then the pncc-quantity observations illustrated in

Figure 16 3(ij) will be generated If we draw a line through these observed

points, we v/ill have a good approximation to the demand curve m Figure
16 3(i)

shifting supply curve and a fixed demand

curve
(i) A
(n) Observations generated by (i)
Fig IS 3

supply curve stays put while

Now assume, as ,n Figure 16 4(l), that the
to changes in "“mber
the demand curve moves about, owtng perhaps
lelatious that will
their incomes Now the
price quantity
of consumers or m curve thri^ugh
be observed are those given Ftgun: 16 4(.i) If we draw a
m
a fan- approamaltou to the supply curve m
these points, we will obtain
Figure 16 4(i)
good But now what happens if both curves shift, as m Figure
So far so
will be very unlike
16 5(i) ^ In we will obtain a sen^ of points that
this case
 THE THEORY OF DEMAND: MEASUREMENTS AND TESTS 217

either the demand or the supply curves that generated them. A few such
points are shown in Figure 16.5(ii).^
This difficult problem is not insurmountable. A proof of the solution to
the problem belongs to a textbook in econometrics and we can only assert
the solution here without attempting to prove it. The key to identifying both
the demand and supply curves in Figure 16.5 (i), given the observations in
Figure 16.5(ii), is to relate the demand and supply to other variables as
well as price - supply
one variable and demand to some other variable. For
to
example, we might relate supply of the commodity not only to the price of
the commodity but also to its costs of production, and we might relate de-

Price

Quantity

(i) A shifting demand curve and a fixed supply curve,

(ii) Observations generated by (i).
Fig 16.4

mand not only to the price of the commodity but also to consumers’ incomes.
Provided that both of these other factors, cost of production and income,
vary sufficiently, it will now be possible to determine the relation between
supply and price as well as the relation between demand and price.
In serious work concern is usually given to identifying the various rela-
tions. In some applied work, however, the problem is often ignored. When-
ever you see an argument such as the following: ‘Last year the price of
whisky rose by 10 per cent and whisky exports hardly fell at all, so we know
that the foreign elasticity of demand must be very low’, you should ask if
the author has really identified the demand curve. If the rise in price was
I A careful statement of these problems in the measurement of demand (without any hint
as to how they might be solved) can be found in E. J. Working, ‘What Do Statistical Demand
Curves Show?’, Qiiarleiiy Journal of Economics, 1927; reprinted in Readings in Price Theory

(American Economic Association, 1953).

 218 THE INTERMEDIATE THEORY OF
DEMAND
1“'.
that '.'J
“ T
the short run
wtaky ivc may actually have discuvmrf
sttppl, mn, of whisky is very inelastic (since whisky take,
several years to manufacture)
The general result to kelp
“ro'Tool'on telling us which curve
£
shifted
and
lither'thl lernd „7tr,±^v '‘f
''’V'

(i) Shifting demand and supply curves

(ii) Observations generated by (i)

Fig 165

WHY IS THE MEASUREMENT OF DEMAND IMPORTANT?

Much work has been done on the measurcnseni of demand and demand
elasticity This work is of great value because it provides our theory of price
with empirical content If wc knew notAmg about demand elasuciucs then
all of the exercises we have gone through m previous chapters would have
very little application to the real world
A somewhat difTcrcnt view of the importance of empirical measures of
demand is held by some economists We shall quote from a classic state

ment of this view by Lionel Robbins ^

‘Our a priori deductions do not provide any justification for saying that
caviar is an economic good and carnon a disutility Still less do they in
1 L Robbins An Essay on tht Nalatt and Sgnificantt aj Eanamv Scime (Macmillan 1932)

pages 98 101 Every economics specialist should read this provocative work It contains the

classic statements of many views stili held by economists II also states a view on the nature of
economic theory and its relation to empincal observations that is directly contradictory to the
one presented in this book For a view inucb closer to the one presented here however seethe
same authors The Present Position of Economics , Riouta Dt Etonomica September 1939
Many other econotmsii at the time sharedlbnviewand we have quoted from Robbins because
It IS such a clear statement of the view then commonly held and not unknown even today
For

a similar statement see lor example LvonMises Human Act on Hodge 1949 Chapter 2
 THE THEORY OF DEMAND: MEASUREMENTS AND TESTS 219

form US concerning the intensity of the demand for caviar or the demand
tobe rid of carrion. But is it not desirable to transcend such limita-
. . .

tions ? Ought we not to be in a position to give numerical values to the

scales of valuation, to establish quantitative laws of demand and supply ?
... No doubt such knowledge would be useful. But a moment’s reflection
should make it plain that we are here entering into a field of investigation
where there is no reason to suppose that uniformities are to be discovered.
‘A simple illustration should make this clear. Suppose we are con-
fronted with an order fixing the price of herrings at a point below the
price hitherto ruling in the market. Suppose we are in a position to say,
“According to the researches of Blank {1907-1908) the elasticity of de-
mand common herring {clupea harengus) is 1-3; the present price-fixing
for
order therefore may be expected to leave an excess of demand over supply
of two million barrels”.
‘But can we hope an enviable position ? Let us assume
to obtain such
that in 1907-8 Blank had succeeded in ascertaining that, with a given
price change in that year, the elasticity of demand was -3. What reason 1

is there to suppose that he was unearthing a constant law ? [The demand

for herrings depends on many things: fashion, theological views, the

availability of other foods, the distribution of income, the current state
of the art of cooking, etc.] Is it possible reasonably to suppose that coef-
ficients derived from the observation of a particular herring market at a
particular time and place have any permanent significance - save as
Economic History?’
The above argument runs somewhat as follows ‘I can think of no reasons :

why the relationship in question (e.g., the relation between demand and
price) should be a stable one I can in fact think of several reasons why it
;

should not be stable; I conclude, therefore, that in the real world the rela-
tionship will not be stable, and attempts to observe whether or not it is stable
can be ruled out on grounds as a waste of time.’ This argument is
a priori
rather a curious one, and it appears, although it may not have been the
author’s intention, that it could have been used to stop at an early stage the
investigations that produced observations of practically every stable relation

we now know.^
1 For example;
(1) Consider all the factors that influence a falling body [append a list of 20 or so factors
which do, or could be imagined to, influence such a body] ;
clearly, since most of these are
variable, we will find no constant, stable relations in the realm of falling bodies.

(2) Observe in detail the complex, varied and variable behaviour of four or five gas mole-
cules; clearly there can be no stable laws about the behaviour of gas (i.e., Boyle s law is a priori
impossible).

(3) Consider all of the factors which cause a person to make a mistake when we ask him to
measure a given distance [list a multitude of factors influencing human behaviour, ending with
 INTERMEDIATE THEORY OF DEMAND
The fat cnticism of the passage quoted u, thus,
that « pnm arguments
although they may strongly suggest the
hypothesis that certam relauoushtps
will not be stable ones, can never
establish this Only empirical
observations
can demonstrate such proposmons about the real
world Even if the o/mri
arguments turn out to be correct most of the time, there
always exists the
possibility that m a few cases they will be wrong Only cmpincal observation
IS capable of discovering the case in which the a prion argument is wrong
The second major cntiosm of the passage is that it is of critical importance
to economic theory to know just how vanable any given relationship is If, for
example, tastes are so vanable that demand curves shift about violently
from day to day, then of the comparative static-equilibrium analysis of
all
the previous chapters would be useless, for only by accident would any
market be near its equilibnum and this would only occur momcntanly If,
on the other hand, tastes and other factors change extremely slowly, then
we might do very well to r^rd the relation between demand and pnee as
constant for purposes of all predictions up
50 years Even if
of, say, to 20 or
^vc could shotv, on a prion grounds, that every relation between two or more
variables used m
economic theory was necessarily not a stable one, it would
be critical for purposes of theory to know the quantitative amount of the
lack of stability Only empirical observations can show this and such obser-
vations are thus important for economic theory as well as for economic
history
Let us consider an example of this point The relation between the de-
mand and their price might have been so variable
for herrings that the

elasticity of demand was 1 37 m 1903, 01 m

1905, 8 73 in 1906 1 00 in
1907 and 41 2 in 1908 If demand for all goods varied such a capricious m
way, pnee theory would be of very little use, for we would be unable to
predict the effects of the sort of shift in costs, taxes etc that we have been ,

considering If, on the other hand, all the measures of the elasticity of de
mand for herrings over a period of 20 years lay between 1 2 and 1 45, then

we could predict the effects of vanous changes in the herring market with a
close degree of accuracy and with a high degree of confidence
We would
be astounded, and indeed we would sospecx a fraud, if a large number
of

measures of the elasticity of hemng demand, made several places andm

can be no stable reiauons
free will jtaicd in iCalics or capitals ai preference] Clearly there
regarding human errors The normal curve of error is « prim impossible
might on consumption goods
affect expenditure
(4) Consider the number of
factor* lhal
your list until you run out of
im order to make the argument seem really conclusive carry on
might affect con
paper - you will have no trouble doing this for the number of factors which
consump
sumption IS infinite] Clearly ii wiU be imp<»sibIetoiind any stable relation between
tion expenditure and a few variables such as the level and distnbuuon of
income Such stable
statistical relations as have actually been found are « prim impossible
 THE THEORY OF DEMAND: MEASUREMENTS AND TESTS 221

over a large number of years, all produced the value of, say, 1-347. What we

want to know, however, is how much spatial and temporal variation there is
in the demand for herrings. Only empirical observations can settle this

question.
Finally, even if we find substandal variations in our relations, we want
to know if these variations appear capricious or if they display a systematic
pattern that might lead us to expect that herring demand is related to other
factors. We might, for example, find a strong but sometimes interrupted
tendency for the elasticity of demand of herrings to fall over time. We might
then find that this systematic variation in price elasticity could be accounted
for by income variations (as the population gets richer its demand for
herrings and less affected by price variations and so the demand be-
is less

comes more and more inelastic). We might now find that a high proportion
of the changes in herring demand could be accounted for by assuming a
stable relation between demand on the one hand and price and income on the
other. In general, what looks like a very unstable relation between two
variables may turn out to be only part of a highly stable relation between
three or more variables.
All of this leads us to the following conclusion : the theory of demand and
price can have few applications world without some empirical
to the real
observations of quantitative magnitudes. Empirical measurements are criti-
cal to economic theory'. Without some knowledge of the stability or instability
of a particular relation (e.g., the relation between price and quantity), we
cannot use economic theory to make useful predictions about the real world.

Modern Measures of Demand

A great deal of work has been done in recent years on the theory and
measurement of demand.^ Particular interest has attached to the demand
for consumers’ durables, such as cars, radios, refrigerators, television sets
and houses. Demands goods are particularly interesting
for these types of
because they constitute a large fraction of total demand and because such
demands can be quite variable from one year to the next. Since the com-
modity is durable, it can always be made to ‘make do’ for another year, thus
purchases can be postponed with greater ease than can purchases of non-
durables such as food and services. If enough households decide simul-
taneously to postpone purchases of durables for even six months, this
decision can have a major effect on the economy'.

1 See, for a survey of recent literature, Robert Ferber, ‘Research on Household Behavior ,

American Economic Review, March 1962.

 PART 4

THE INTERMEDIATE
THEORY OF SUPPLY
CHAPTER 1 7

BACKGROUND TO THE
THEORY OF SUPPLY

In Chapter 8, we hypothesised that the supply of a commodity depends

upon many factors, including the goals of firms, technology, the price of
the commodity, prices of other commodities and prices of factors of produc-
tion. In the elementary' theory of price outlined in Part II, we assumed the

existence of a supply curve relating the quantity of goods that firms wished
to sell to the price of the commodity. We must now take a more detailed
look into the theory' of supply to find out how the decisions of individual
producers determine the quantities of goods actually supplied to the market.
This requires attention to producers' goals, to the choices they have, to the
information they' get, to the signals to which they respond, and to the insti-

tutions in which they operate.

A theory' of supply, or, as often called, a theory' of production, is
it is

needed in order to answer a whole host of important questions If consumers

:

want more of a commodity how can they get it? How quickly'? How ex-

pensively? How certainly? How will a subsidy, or a tax, or a new la^v, or

an invention affect does the kind of market in which firms
supply? How
operate affect the price and quantity' of goods available to consumers ? Will
large firms be more efficient than small ones ? Will they become monopolies ?
Why What are take-overs and do they affect the w'ork-
do firms combine?
ings of the economy ? What are the causes and consequences of advertising?
Will the existence of labour unions, minimum-wage laws and anti-monopoly
laws make much difference to what is produced and how it is produced ?
This is questions that we cannot answer wth-
but a sample of the many
out a theory of production. In an introductory book, we cannot answer
them all, but we can outline the basic theory' of producdon, use it to ex-
amine some of these questions, and, what is at least as important, we can
see what other information would be required to permit complete analy
ses

of these questions.
8
 226 THE INTERMEDIATE THEORY OF SUPPLY
THE NATURE OF THE FIRM
In Chapter 6 we discussed ihe assumptions made in economic theors about
the major decision takers in the economy We have already discussed the
theory of household bchavnour m some detail and we now come to the
second important group of decision takers firms
It will be recalled from Chapter 6 that
the firm was defined as the unit
that lakes decisions with respect to the production and
sale of commodities
The concept of the firm covers a wide variety of actual business under-
takings At one extreme, there is the owner-operated firm, such as the
family-operated grocery store or the small manufacturing company, in
which all important decisions arc made by a single person At the other
extreme, there is the large joint-siock company, such as the Bntish Motor
Corporation (BMC) or Imperial Chemical Industries (ICI), in which de
cisions arc made by many difTercnt persons To uke an example, when
BMC decided to introduce a new model car, someone made the decision
to do so Someone decided to call It the Mini Minor Somconcdecidedhow
and where to produce It Someone decided how to promote its sales In the
face of very high sales someone decided m what directions to develop from
the onginal model (a high-powered version called the Mini Cooper and a
senes of more luxurious v ersions on the same pnnciple as the original Mini
Minor)
How then can we regard the firm as a single decision taking unit even
though we know, as the above example shows, that decision taking may be
very decentralised wuhm a large firm=* We do this by assuming that the
efTorts of difTerent decision taken arc unified by the fact that they share
common goals In the above example the common goal is the manufacture
of successful cars and other products that earn profits for the Bntish Motor
Corporation This assumption is cntica! to the whole traditional theory of

the firm and we may state il formally as follows

Economic theory assumes that the same principles

underlie each decision made within a firm and that
the decision is uninfluenced by who makes it. It is thus
assumed that we can abstract from the peculiarities
of
of the persons making the decisions and the kinds
organisation in -which they work. We thus assume that
unit
the firm can be regarded as a decision-making
having obj’ectives of it own and making decisions de-
signed to achieve these oiqectvves

proprietor, a plant
Whether a decision is made by a small independent
the
manager, or a Board of Directors, that person is, as far as theory goes,
 BACKGROUND TO THE THEORY OF SUPPLY 227

firm for the purposes of that decision. This is a truly heroic assumption. It
amounts to saying that for purposes of predicting their behaviour, at least
in those aspects that interest us, we can treat the farm, the corner grocery,
the large department store, the small engineering firm, the giant" chemical
combine, the large shipyard, the giant oil firm and that largest of all busi-
ness organisations the General Motors Corporation of .America, all under
the umbrella of a single theory of the behaviour of the firm. Even if this
turns out to be only partially correct it will represent an enormously valuable
simplification which will show' the vast power of theory in revealing some
unity of behaviour where to the casual observer there is only a bewildering
diversity.^
Criticisms of the theory of the firm because of its neglecting to identify
either decision-takers or the institutional structure w'ithin which decisions
are made are discussed in Chapter 29. Some competing hypotheses about
actual business behaviour are also discussed at that time. The final test of
whether or not such factors can be legitimately ignored is an empirical one;
if the theory that we develop by ignoring these factors is successful in pre-

dicting the outcome of the kind of events in which we are interested, then
we can conclude that we w'ere correct in assuming that these factors could
be safely ignored.

THE MOTIVATION OF THE FIRM

We assume that the firm makes decisions in such a w'ay that its profits will

be as large as possible. In technical language, we assume that the firm

maximises its profits.

Profits are defined as the difference between the value of the sales of the
firm and the costs to the firm of producing what is sold. This definition leads
us to look carefully at sales and at costs. In the elementary theory' of the
firm we assume that the rate of production is kept exactly equal to the rate
of sales. Thus, when considering the sales revenue of the firm, we can treat
the value of production and the value of sales interchangeably since they
are identical.^ The definition and measurement of cost is a difficult matter
and it is discussed in Chapter 19.
1 The student must not be surprised if at first encounter the theory seems rather abstract
and out of touch with reality. Because it does generalise over such a wide variety of behaviour
it must ignore those features with which we are most familiar and
which in our eyes distinguish
the fanner and the grocer from the British Motor Corporation. Any theory that generalises
over a wide variety of apparently diverse behaviour necessarily has this characteristic, because
it ignores those factors that are most obvious to us and which
create in our minds the appear-

ance of diversity.
2 This simplification, which neglects such complications as inventories and goods in process,
is appropriate to an introductory book. It can be removed
without seriously affecting any of
the conclusions reached in the present treatment.
 228 THE INTERMEDIATE THEORY OF SUPPLY
The assumption of profit maxinusation enables us to
predict the be
haviour of the firm in regard to the vanous choices
open to it We do this by
studying the effect that making each of the
choices would have on the firm’s
profits We then predict that the firm will
select from the alternatives open
to It the one that produces the lai^est profits
At this point the reader is cnutled to ask if wc are justified in going ahead
to build an elaborate theory based on such a crude
assumption about the
motives of the businessmen It is well known that some
businessmen are
inspired some of the time by motives other than an overwhelming desiit
to make as much money as they possibly can Cases in which businessmen
have gone after political influence, and others in which decisions have been
influenced by philanthropic motives, are not difficult to document Should
we not, therefore, say that the assumption that businessmen seek to maxi
mise profits is refuted by empirical evidence^ Should we not say that many
motives affect business decisions and a much richer theory than simple pro
fit maximisation is necessary to handle so complex a problem as business
decisions ’
The real world is complex A theory picks on certain factors and deals
with them on the assumption that they are important ones, and that the
ones Ignored are the relatively less important ones If it is true that the key
facton have been included, then the theory will be successful in predicting
what will happen m
the real world underspecified circumstances It follows
that it is not an important criticism to point out that a theory ignores some
factors knoivn to be present in the world, this tells us nothing more than

that the theory is a theory and not just a photographic reproduction of

reality in complexity If the theory has ignored some important fac-
its full

tors, predictions will be contradictol by the evidence, at least in

then Its

those situations in which the factor ignored is quantitatively important

How does all of this relate to profil-maximising theory^ First, this theory
does not say that profitis the only factor that ever influences the business

man believed only that profits are an important consideration, impor-

It is

tant enough so that a theory that assumes profit maximisation to be the

sole motive of the businessman will produce predictions that are substan
tially correct It follows from this that to point out that businessmen arc
sometimes motivated by considerations other than profits does not con-
stitute a relevant criticism of the theory It may well be
that profit-

maximising theory is substantially wrong but, if so, the way to demonstrate

this is to show that the predictions that follow from the theory
arc refuted

by the facts This, of course, requires that wc have mastered the theory
first

shall therefore press on to discover the implications of the

assumption
We
that businessmen seek only to maianiisc profits, when we have
completed

this task we shall ask how the theory nu^t be tested against cmpincal evi-
 BACKGROUND TO THE THEORY OF SUPPLY 229

dence. Finally, in Chapter 29 we shall consider several alternative theories

that have been put forward.

Nationalised industries: In the United Kingdom and many other

Western countries we must allow for the fact that a significant part of total
production is in the hands of industries that are owned by the state. Usually
these industries are run by boards that are appointed by the state but given
considerable autonomy within the firamework of broad directives on tvhat
goals to pursue. If the nationalised industries seek to maximise their profits
then their behaviour will be indistinguishable from that of private firms. If,

however, they are given other goals, such as always to try just to cover all
their costs so as tomake neither profits nor losses, then their behaviour will
differ from that of the private sector.* Once we know the objectives of these
industries we will be able to predict their response to changes in market
signals. In the next few chapters we shall confine ourselves to the behaviour
of privately-owned firms, but in Chapter 38 we shall return to the question
of nationally ovvmed ones.

CHOICES OPEN TO THE FIRM

In the theory of production, as in the real world, the firm must decide what
to produce, how to produce it, and in what quantities to produce it.
Initially, we shall assume that the firm produces a well-defined single
product. The choice of the method to be used in making this product is the
heart of the theory of production, and the choice of the quantity in which
it is to be produced is an integral part of the theory of price. Both are dis-

cussed at length in subsequent chapters. The remainder of this chapter is

designed to prepare for this discussion, as well as to lay the groundwork

for the discussion of costs.

Factors of Production

Production is something like a sausage machine: certain factors, such as raw

materials and the and labour, are fed in at one end and
services of capital
some product emerges at the other. The materials and factor services that
are fed in at the one end and used in the process of production are often
called INPUTS and the products that emerge at the other end are often
called OUTPUTS. One way of looking at the process is to regard the factors

I This will not always produce behaviour which difTers from that of profit-maximising
firms. If, as is often the case, the industry is currently making losses they will be trying to re-

duce these and will select the most profitable of any alternative open to them. The profit-
maximising thcoiy tvill thus be able to predict much of their behaviour.
 230 THE INTERMEDIATE THEORY OF SUPPLY
as being combined to produce the output Another way, equally
useful «
to regard the materials and factor services as being used up, or sacnficed
in order to gam the output
’

Eachdistinct input into the production process can

be regarded as a
facioT of production There are literally hundreds
of factors of production
entenng into the output of a specific good Among the factors
entering into
car production are, to name only a few, sheet steel, rubber, spark plugs,
electncity, night watchmen, accountants, fork-lift operators, managen and
painters We
can group these inputs into four broad classes {1} those that
are inputs to the car manufacturer but outputs to some other manufacturer,
such as spark plugs, electncity and sheet steel, (2) those that are provided
directly by nature, such as land and raw matenals,
(3) those that arc pro-
vided directly by households, such as labour, and (4) those that arc pro-
vided by the machines used in the manufacture of cars
This last class of inputs - machines - is one of the distinguishing features
of modern as opposed to pnnimvc production Instead of making all con-
sumers’ goods directly with the aid only of such simple tools as nature pro-
vides, productive effort goes into the manufacture of tools, machines, and
other goods that are not desired m themselves but only as an aid to making
further goods Such tools are called capital goods, which we define as
man-made aids to further production Tlie use of capital goods renders our
production processes roundabout Instead of making what we want directly,
we engage in a roundabout process, first making the goods that we need in
order to help us make what we finally want
many cases, production is very roundabout indeed For example, a
In
worker may be employed m a factory making machines that arc used in
mining coal, the coal may be burned by a power plant to make electncity
the electricity may provide power for a factory that makes machine tools,
the tools may be used to make a tractor, the tractor may be used by a potato
farmer to help in the production of potatoes, and the potatoes may be eaten
by a consumer Such roundabout production is worthwhile if the farmer
using his tractor can produce more potatoes than could be produced if
all

to apply
the factors of production involved in this roundabout chain were
tools as
themselves directly to the production of potatoes (using only such
were provided by nature)
speak
For many purposes, including ease of exposition, it is convenient to
produc
of certain broad types of factors the services of which contnbute to
used, in-
tion A popular classification is land, labour and capital Land, so
products, labour includes all human sersices,
and
cludes all pnmary
capital includes all man-made aids to further production The reader

land labour,
I In the example of the car manufacturer we found four classes of inputs
•apital and other producu that were the car industry* inputs but the
outputs of other rms
 BACKGROUND TO THE THEORY OF SUPPLY 231

should never lose sight of the fact that these broad designations may cover
ver>' different things and that, for many problems, attention to a more de-
tailed list of factors is essential.

TECHNICAL AND ECONOMIC EFFICIENCY

In general, there is more than one way to produce a given product. Indeed,

if thiswere not the case there would be no need for firms to face the decision
of how to produce. It is possible to produce agricultural commodities by
farming a small quantity of land very- intensively, combining a great deal of
labour and capital with each acre of land; it is also possible to produce the
same commodities by farming a great deal of land very extensively, using
only a small amount of labour and capital per acre of land.
Consider the printing of this book. It would have been possible to set the
type for this book in at least four ways: by using a linotype machine, by
using a monotype machine, by photocomposition or by assembling the
individual letters and spaces by hand. Which process is best ? If the output
is the same, the best process is the one that uses the fewest inputs or, in other

words, the one that is technically most efficient. Technical, or technological,

efficiency measures use of inputs in physical terms; economic efficiency measures
use in terms of costs.
A simple illustration should make the distinction clear. Suppose we

Table 17.1

KNOWN WAYS TO PRODUCE 100 UNITS

OF OUTPUT PER MONTH

Quantity of inputs required

Capital Labour

Method A 6 200
Method B 10 250
Method C 10 150
Method D 40 50

we trace these ‘inputs’ back to the firm that proraded them, we will find that they were pro-
duced with the four types of inputs. Eventually, however, if we continue to trace all products
back to their sources, we will find that all production in the economy can be accounted for by
the services of three kinds of inputs: land, labour and capital. The other products appear only
because the stages of production are broken up between different firms so that at any one stage
a firm is using as inputs goods produced by other firms.
 232 THE INTERMEDIATE TBEOET OF SUPPLY
assume that a product uses only tsvo fictots of production,
ivhith Me shaft
call capital and labour, and the stale of technology
is such that onr engineers
lell ns there are only four knoun ways to
produce the desired output, which
100 units per month Which method shall we me’ The
IS
methods are
summarised in Table 17 1

Method B IS Uchnologtrally tneffttenl because it uses more of both resources

than method A,' .and is thus cleariyuTstefuJ Amonj methods A CandD,
method A is more eflicient m tis use of capital but most labour using
Method D conser\cs labour, but uses much more capital Method C is inter
mediate between them* On an oscrall basis, all three are ftchnologually
fffictenl because no one uses less of all resources than the others *
Economic efUcicnc) involves choosing from among the technologically
efficient combinations the one that represents the least sacrifice for the firm
The sacrifice of method A is whatever it costs to purchase the services of 6
units of capital plus the cost of hiring the services of 200 units of labour, and

Table 172
ECONOMIC EFFICIENCY DEPENDS ON FACTOR PRICES

fnctor pnees ptr unit Total test sjjatlors

Capital Labour hfethoi A MilhodC 1 Method D

1
1 ^ 1

Case I £X £3 cm 1
C9i0 CiMO
Case 11 20 5 1,120 950 1 050
Case 111 15 5 1090 900 850
1

similarly for methods C and D

The economically most efficient method is
assumed sets of
the one that costs the least Table 17 2 deals with three
will depend on factor prices
factor prices It shows that economic efficiency
Given the quantity of inputs required (sec Tabic 17 1), method A is most
efficient with case I prices^ method C is most
efficient with case 11 prices

and method D
is most efficjcni with case III prices

1 It might also br so classified because it w« the same amount of capital and more labour
han method C ^ t
method technologically most effecient because
D
2 The student who is tempted to consider
think twite
i uses only 90 units of all resource* ahouW
background engineering or physics will note that t^echnical eflicienc)
3 Students with a i«
I Itself a broader concept than engineering
elTcicno An engineer may judge the elTciency o
the fuel that is actually
.steam engine according to the percenUge of the potential energy in
oi
concerns the quantities of-i// facton
.reduced as power by the eitgine Technical cfTacncy
iroduction used while engineering efllcietiey concentrates on just one
 :

BACKGROUND TO THE THEORY OF SUPPLY 233

The Principle of Substitution and Factor Proportions in Production

A SINGLE FIRM : Econoitiic efficiency as we have defined it makes the choice
of a method of production depend upon the relative prices of factors. In
the example just used, when capital is very expensive relative to labour
(case I), the efficient firm uses a production method (A) that uses very little
capital and a great deal of labour (see 17.1). If the price of capital relative
to labour decreases, as happens when we move from case I to case II, and
from case II to case III, it ‘pays’ to shift to methods of production that use
less labour and more capital.*

The principle of substitution says that (for a given set

of technical possibilities) efficient production will
substitute cheaper factors for more expensive ones.

This principle is a logical proposition. It tells us how the economically

most method of producdon necessarily changes as relative factor
efficient
prices change. If we now add to this our behavioural assumption that firms
are profit maximisers and so will choose the method of producing any given
output that involves the least cost, we get the following prediction about
substitution in the real world
The methods of production will tend to change if the
relative prices of factors change. Relatively more of
the cheaper factor will be used and relatively less of
more expensive one.
the
The whole economy: The relative prices of factors of production in an
economy will tend to reflect the relative scarcity of these factors in relation
to the demand for them.
In a country with a great deal of land and a small population, for example,
there will be a large supply of land relative to the demand for it, and its
price will be low while, because labour is in short supply, the wage rate
may be high. In such circumstances firms producing agricultural goods will
tend to make lavish use of (cheap) land and to economise on (expensive)
labour; a production process will be adopted that is land extensive and
labour intensive. On the other hand, a small country' with a large popula-
tion will be one in which the demand for land will be high relative to its
supply. Thus, land will be very expensive and firms producing agricultural
goods will tend to economise on land by using a great deal of labour per
unit of land. In this case a productive process will tend to grow up that is
land intensive and labour extensive.
Thus we see that relative factor prices will reflect the relative scarcities
(in relation to demand) of different factors of production: abundant factors

1 In case I, one unit of capital is equivalent in cost to 16§ (='^) units oflabourj in case II,

the ratio is 4 (
= -x)> aud in case III, 3 (=f).
 234 THE INTERMEDIATE THEORY OF SUPPLY
will have pnces that are lew
relative to the pnces offactota
that ate scarce
l-irms seeking their own pnvate
profit will be led to use much of the
facton
with which the whole country is plentifully
endowed, and to economise on
the factors which are in scarce supply
m the whole country This general
theory of factor pncing and factor
combinations by firms, helps to explain
the widely differing productive processes
to be found in such differently
endowed countries as say, the United Kingdom, Denmark, India,
Australia
and the United States *
This discussion provides an example of what we mean when we
say that
thepnee system is an automauc control system No single firm needs to be
aware of national factor surpluses and scarcities Prices determined on the
competitive market tend to reflect these and individual firms which ncier
look beyond their own private profit are nonetheless led to economise on
factors that are scarce to the nation as a whole
The above suggests that we should not be surprised to discover that
methods of producing the same commodity differ m different countnes In
the United States, where labour is highly skilled and very expensive, a steel
company may use very elaborate machinery to economise on labour In
China, where labour is abundant and capital very scarce, a much less

mechanised method of production may be appropriate The Western cn

gineer who feels that Chinese arc way behind Westerners because they are
using methods discarded in the West as inefficient long ago may be missing
the truth about economic efficiency in use of resources The suggestion, often
made, that to aid underdeveloped countnes we need merely to export
Western ‘know-how may be incomplete
It appears that the price system does lead profit maximising firms to take

account of the nation’s relative factor scarcities when deciding which of the
possible methods of production to adopt One must avoid jumping to the
conclusion that whatever productive processes are produced by the market
are always the best possible ones and that they should never be interfered
with There is, however, a strong common sense appeal the idea that any m
take account
society interested in getting the most out of its resources should
processes
of the relative scarcity of its resources in decidmg what productive
the extent that the rclauvc pnces of factors of production
do
to adopt To
seek-
approximate relative scaratics of these factors, pnvate firms
reflect the
on factors that are
ing their own pnvate profit will be led to economise
generally scarce ^ , ,

1 This IS probably a reasonable statement »hen we thmb of such exogtnoosly dtlerntned

of whic
faclors as land and labour It is much less certain when we come to capital the supply
u endogenously determined
that factor pnces do not
2 The reason for being cautious m the statements of this paragraph is

necessarily perfectly reflect scarcity to society pnvate cost is not always social cost , current scar
apters
aty IS not always future scarcity, etc Some of the issuei involved are discussed in later c
CHAPTER 18

THE ORGANISATION
OF PRODUCTION

We saw in Chapter 17 that in standard theory vve abstractfrom the insti-

seUmg in -which decisions ate made. We assume that a\\ firms,
whatever their structure, make individual decisions in such a way as to
maximise their profits. Of all the assumptions of standard theory, the one
that the structure of the firm can be ignored is the one that has come under
most frequent and most sustained attack. Once we have developed the
standard theory, we shall devote considerable attention to such criticisms.
In order to understand what we are abstracting from, and in order to con-
sider the criticisms of our abstractions, it is necessary at this stage that the
student should understand something of the nature of the modern business
organisation.^

PROPRIETORSHIPS, PARTNERSHIPS AND JOINT-STOCK

COMPANIES
There are three major forms of private business organisation, the single
proprietorship, the partnership and the joint-stock company, which in
America is called the corporation. In the single proprietorship, there is a
single owner who is personally responsible for everything done by the busi-
ness. In the partnership, there are two or more joint owners each of whom
is personally responsible for everything done by the business. In the joint-

stock company, the firm is regarded in law as having an entity of its own
and the owners are not each personally responsible for everything that is

done by the firm.

1 This chapter also introduces much of the elementary terminology of business organisation
and Without a knowledge of the meaning of these terms, the reader tvill be unable
finance. to

strike any contact between what he has learned from this book and what he has read in

magazines or newspapers about the world of finance, industry and trade.

 the intermediate theorv of supply
In most Western countncs it is
necessary to add two other forms of
organising production, nationalised indusines and
goods and services pro
vided without direct charge by the state
Nationalised industries arc owned
by the state and are usually under the direction of a more
or less indepen
dent board appointed by the state Although their ownership diffen
the
organisation and legal status of these firms is very similar to
that of the
joint stock company and they are perforce very large firms Their activity
issimilar also in the sense that they gam their revenue by
selling to the
public the product that they produce In the United Kingdom about
18 per
cent of total production
is accounted for by these firms, the most important

of which in 1966 were, railroads, coal, airlines, electricity, gas, postal and
telephonic communication
The final form of production is quite dissimilar from all those considered
abo\e because revenue is obtained from general taxation and the product is
provided free to the public Important examples found in all countries arc
defence, roads and education In the UK
we rfJust also add the National
Health Service which provides medical and hospital services free to the
general public In countncs without nationalised medical services hospitals
and doctors behave just as do other firms, in that they purchase factors of
production on the open market and gam revenue by selling their products
to those peoplewho wish to, and can afford to, purchase them In the next
few chapters we shall confine our attention to the privately owned sector
of the economy, but we shall return to the publicly owned sector in
Chapter 38

The Proprietorship and the Partnership Advantsgesand Disadvantages

of the single proptietonlup is that the owner can
The major advantage
readily maintain full control over the firm He »s the Boss The disadvan-

tages are, first, that the size of the firm is limited by the capital the owner

can personally raise, and, second, that he is personally responsible m law

for all debts of the firm

disadvantage of the
The partnership overcomes to some extent the first

proprietorship but not the second Ten partners may be

able to finance a
subject to
much bigger enterprise than could one owner, but they are still
Each partner is fully liable for a o
what IS called unlimited liability
amount of money
the debts of the firm This liability is independent of the
a particular proprietor may have invest^ the firm m Thus, if a tent

that matter) when

partner makes X'.OOO available (or ^ClOO, or nothing, for
he joins a firm that subsequently becomes bankrupt with debts of^l ,

other nine partners is fully liable for

then this individual, together with the
 THE ORGANISATION OF PRODUCTION 237

the 00,000. If none of the other partners should have saleable personal
assets, whereas the tenth partner has a house, a car. furniture, and some
investments, he may lose all of his possessions so that the debts of the partner-
ship can be cleared. Obviously, a man with substantial personal assets will
be unwilling to enter a partnership unless he has eomplete trust in all the
other partners and a full knowledge of all the obligations of the firm. As a
direct consequence of unlimited liability, it is difficult to raise money from
many persons through a partnership because of the need for all of the part-
ners to have complete confidence and trust in each partner and to have a
full knowledge of all the firm’s business. .An investor may be willing to risk

;^1,000, but not be willing to jeopardise his entire fortune; if, however, he
joins a partnership in order to do the former, he cannot avoid doing the
latter.

The Joint-Stock Company Advantages and Disadvantages

;

The joint-stock company is regarded in law as an entity separate from the

individuals who own it. It can enter into contracts, it can sue and be sued,
it can own property, it can contract debts, and it can generally incur obli-

gations that are the legal obligations of the company, but not of its owners.
This means that the company can enter into contracts in its own right and
that its liability to adhere to such contracts can be enforced by sueing the
company, but not by sueing the owners.*
The company issues shares that are purchased by the general public. The
company obtains the money paid for the shares, and the shareholders be-
come its owners. The shareholders are entitled to share in the profits of the
company. When such profits are paid out, they are called dividends. The
shareholders are also entitled to split up the assets of the company should
it be liquidated.
This method of finance means that the owners of the firm cannot all be
the managers. The line of control is as follows. The shareholders (and there
may be board of directors. The board of
tens of thousands of them) elect a
directorsis supposed to act as a cabinet. It sets broad issues of policy and

appoints senior managers. The managers are supposed to act as civil ser-
vants and to carry out the wishes of the directors, translating the broad lines
of policy laid the directors into a series of detailed decisions.
down by
The most important aspect of a joint stock company from the point of
view of its owners is that the owners have limited liability. Should the
company go bankrupt, the personal liability of any shareholder is limited

1 The right to be sued may not seem to be an advantage. It is an advantage

because it

makes it possible for others to enter into enforceable contracts with the company.
 '

238 THE INTERMEDIATE THEORY OF

SUPPLY
to the amount of money that he has actually invested
m the firm bv purchas
mg shares Since the shareholdcn and
Its
the corporation are receded
as
separate entities, the shareholder is not
liable to lose his personal assets
order to meet the firm’s obligauons The
m
great advantage of the joint-stock
hmited-hability compan> is that it can raise capital
from a very large num-
ber of individuals each of whom gets a share in the
firm’s profits, but -who,
beyond nsking the loss of the amount actually invested,
has no liabiliUes'
Thus each investor knows his maximum risk exactly, and he may sit back
and collect his dividends without knowing anything about the policy or
operation of the firm that he, along with many others, actually
owns
The disadvantages (from the point of view of the investor) are, fint, that
he may have little say in the management of the firm (for example, if the
owners of a majority of the shares decide that the corporation should not
pay dividends, an individual investor cannot compel them to pay him his
share of the earnings), and, second that the corporation is subject to in-
come taxation, as are the investor’s dividends
The joint stock form of business organisation is employed wherever very
large enterpnses are found The reason is that it has decisive advantages
over any other form in being able to raise the large sums of capital required
for major enterpnses Histoncally, wherever and whenever large accumu-
lations of capital m a single firm were required the limited-liability, joint-
stock company developed It is not usually found in service industries or
agriculture because these do not need large quantities of capital to function

FINANCING THE MODERN FIRM

The money that a firm raises to carry on its business is called its capital This
money may be borrowed from outsiders or subscribed by the owners of the
firm ^Vc therefore speak of the firm’s own capital and its borrowed capital
The plant and equipment (and all the man made aids to further produc-
tion) owned by the firm are referred to as its capital equipment or capital goods

This use of the same word, capital, to refer to both an amount of

money and
a (quantity of goods can be confusing, and the terms money capital and
real

capital is used
capital are sometimes used instead Most often the single term
interchangeably It is usually clear from the context whether a sum of

money or a stock of equipment is being referred to

Whether taxation of
1 Debate over this ‘double taxation ofincome has existed for decades
nature of the coiise
corporate income is fair whether there arc ofSening pnvileges and the
shall not discuss
quences of double taxation are mafleis for a course in public finance and we
them in this book .

the money capital

2 The two uses are not totally independent of each other since much of
for production
raised by a firm will be used to purchase the capital goods that the firm requires
 THE ORGANISATION OF PRODUCTION 239

Firms raise capital in several The most important of these are

ways.
(1) selling shares, stocks or they are variously called) either by
equities (as

private or public sale; (2) borrowing by the sale of bonds or debentures',

(3) borrowing from banks or other financial institutions; and (4) re-
investing the firm's profits. All four methods are open to the joint-stock
company; only the last two are open to the single proprietorship or the
partnership.

Eq^uity capital: Capital raised by the sale of stocks is often called

EQ^uiTY CAPITAL. The pcrsons who buy these stocks are called the stock-
holders or theshareholders of the firm. They are the owners of the
firm; they have made their money available to the firm, and they risk losing
it in return for a share in the firm’s profits.
Stocks in a firm often proliferate into a bewildering number of types.
Basically, however, there isco.mmon stock and preferred stock. Com-
mon slock usually carries voting rights and has a residual claim on profits;
have been met, including those of holders of preferred
after all other claims
shares, the remaining profits, if any, belong to the common-stock holders.
The dividends earned by the stockholder may be zero or positive. There is
no upper limit to the possible profits that may be earned by the company
and hence to possible dividends that could be paid out to holders of com-
mon stock.
Preferred stocks be voting or nonvoting, cumulative or noncumu-
may
lative. Basically, the difference between preferred stock and common stock

is that preferred stocks cany' with them a right to a preference over common

stocks to any profits that may be available after other obligations have been
met, but also a stated maximum to the rate of dividends that will be paid
per shilling originally invested.
If the firm earns no profits it has no obligation to pay out diridends to its

shareholders.

Debenture capital; Capital raised by the sale of bonds is often called

debenture capital. The persons who purchase the bonds are called the
firm’s BONDHOLDERS. They are creditors, not owners, of the firm. They
have loaned money to the firm in return for a promise to pay a stated sum
of money each y'ear by way of interest on the loan, and also to repay the
loan at some stated time in the future (say, five, ten, or twenty years hence).
s part ivhether or not
This promise to pay is a legal obligation on the firm
profitshave been made. If these payments cannot be met, then the bond-
holders can force the firm into liquidation. Should this happen, the
bondholders have a claim to the firm’s assets prior to that of any of the
shareholders. Only if the bondholders and all other creditors have been
 240 THE INTERMEDIATE THEORY OF SUPPLY

=*““"pt to recover anything for

them-

The disadvantage of raising capital through the sale

of bonds is that
interest payments must be met whether or not there ate
any profits Many
a firm that would have survived a temporary
crisis had all its capital hcen
equity capital has been forced into hquidation
because it could not meet
Its contractual obligations
to pay interest to its bondholders

Capital borrowed from banks Many firms borrow money from banks
to finance some of their operations This money is
usually borrowed for
much shorter periods of time than is debenture capital Also bank loans
contribute a relatively small fraction of the total capital of firms For some
firms, however, bank borrowing constitutes the most important source of
capital The very small single propnetorship that strikes on a good idea and
begins to groiv may have very few other methods of raising money open to
It Great industrial giants with vast issues of stocks and bonds often began
in someone’s back shed particularly when a whole new industry is just
beginning to develop For the first few embryo years, bank borrowing may
be critical to the growth of such firms Most large firms also make some use
of bank borrowing as a source of funds There is the advantage that extra
money can be raised through this method much more quickly (often noth-

ing more than a ’phone call to the bank manager is needed), than by using
most other forms of borrowing

Capital from re-invested profits Avery important means of obtain

mg funds through the rc-mvesiing, or ploughing back of the firm s own

IS

profits This method of obtaining funds is used by firms of all sizes For
young growing firms it is one of the most important sources of finance
of the easiest ways for the controllers of large firms to raise money
is
One
to retain some of the firm’s own profits rather than to
pay them out as
shareholder does not wish his profits to be
dividends to shareholders If the
' In many cases, firms
re invested there is very httle llial he can do about it
year,
pay out a standard dividend to their common stock holders year after
holding bac^ any profits m excess ot t’liis amount,
an6p*^yrng*r«iCTrf(St«>'‘^«^
necessary to make such
reserves when current profits fall short of the amount
payments Common stocks on which a standard payment is made year in
the original
and year out in spite of fluctuating profits are a very far cry from
idea of a common stock, which yielded little or nothing in
bad times and
very high returns in good times
re
I He can of course sell his share* »nd »/the company managen have re invested the
extra earning
tamed profits wisely the market vatuc of ihc share* should rise to reflect the
power of the company
 THE ORGANISATION OF PRODUCTION 241

The Capital Market and the Stock Market

Broadly speaking the capital market is the place where new capital is raised

and the stock market is the place where existing stocks and bonds are traded.
If a firm requires new capital it will obtain the money by selling bonds (i.e.,
borrow the money from the public) or by selling equities (i.e., by increasing
the number of people who have a .share in the ownership of the firm). These
new issues will be sold on the capital market.
If a member of the public buys a bond from a firm, he cannot get his
money back from the firm until the expiration of a stated period of time; if,
for example, a ten-year bond bought then the bond will be redeemed by
is

the company (i.e., be paid back) ten years after the bond was
the loan \vill

issued. If the individual wishes to obtain his money back at an earlier date,
the only thing he can do is to persuade someone else to take over the loan
and sell the bond to him. If an individual buys shares issued by a company,
he hands over money to the company and becomes one of its owners. Once
this has been done, the individual is unable to demand his money back

from the firm (except in the unlikely event that the firm itself is liquidated).
If he wishes to withdraw his capital from the company the only thing he
can do is to persuade someone else to take over his piece of the firm and sell
his shares to that person.

The market where existing issues of stocks and bonds are sold by one
individual to another is called the stock market. The trading of shares on
the stock market indicates a transfer of the existing ownership of the com-
pany; it does not indicate that the company is raising new money from the
public. The company is none the less interested in the price of its shares on
the stock market, because this price indicates the degree of confidence that
investorshave in the future earning power of the company. If the company
isnot expected to do well in the future, then investors will not be willing to
pay large sums of money for a share in the ownership of the company the ;

price of the company’s shares slump. If the company is expected to do

will
very well in the future, then a share in the ownership of the firm will appear
to be a good prospect; investors will be willing to pay a high price for shares
in the company and
;
the prices of its shares will be high.

THE WIDESPREAD OWNERSHIP OF THE MODERN JOINT-

STOCK COMPANY
In some large companies a family or other small group provided most of
the original capital and the firm grew without substantial sales of equities
to the public. British examples are Ferrantis, the electronics firm, and
Pilkington Bros, until recently one of the largest private companies in the
world, which was founded in 1828. It is still controlled by the Pilkington
 the intermediate theory of supply
lam.ly Moms
and Hillman motors started and once
owned by Lord
Niidield and Lord Rooles arc now public
companies with many owners
Bowateis, wilh assets of around £200 million,
is still controlled by
Sir Eric
Bowatcr In Marks &
Spencer, The Home of Fraser and Great
Univenal
Stores, the founding families still keep control
by means of issuing nonvoting
shares On the other hand, there are many companies
where no one group
IS predommant and which are owned
by lens of thousands of shareholders
the most prominent example in the UK being ICI The charactenstic pat
tern of much of company ownership is that thousands of shareholders own
minute fractions of the total stock, while
dominant groups (often including
other companies) hold from 3 percent to 20 per cent of the voting stock

The Consequences of Widespread Ownership

The days of the single propnetor who was both the owner and the manager
of a company arc gone forever m major areas of the business world Divem
fication of ownership is a major charactenstic of modern business Does it
matter ’ In the remainder of the chapter we consider three hypotheses that
hive been advanced about the consequences of the diversification of
corporate o'^ncrship

1 The Hypothesis of Minority Control Because of the widespread distribution of

shares, the on ners of a minority of the stock are usually able to control a majority of the
voting shares and thus to exercise effective control over the decisions of the company

Let us sec how Each share of common stock

these results might occur
has one vote in a company individual or group controlling 51 per cent
Any
of the stock clearly controls a majonty of the votes But suppose one group

owns only 30 per cent of the stock and the remaining 70 per cent is distn
buted so widely that few of the dispersed group even bother to vote, in this
event 30 per cent may be the ovenvhelming majonty of the shares actually
voted How large a percentage is actually required to control the majonty
‘

depends on the pattern of ownership and on whether there has been a major
effort to collect proxies A colourful, but rare, pbenomentm m
the ‘proxy fight in which competing factions of stockholden
(or
history is

management) attempt to collect the voting rights of the dispersed and

fraction of
generally disinterested stockholders In general, a very small
domi-
shares,sometimes as small as 5 per cent, actively voted may exercise
nant influence at meetings of stockholders
Another aspect of minority control is made possible through the device
1 Shares must be voted at ihe annual meetmg of stockholden They may be voted in person

or by assigning a proxy U> someone who wdl be aiiendwig

 THE ORGANISATION OF PRODUCTION 243

known as a holding company. Suppose, in a certain company, call it A, owner-

ship of 20 per cent of the stock would give dominant control. A new com-
pany, which we shall call B, is now formed. B purchases 20 per cent of the
stock in A. Company B can now control A. But no more than 51 per cent of
the stock of B is required to control the stock in A. Indeed, if 20 per cent
ownership of B is sufficient to control B's affairs, an amount of money equal
only to 4 per cent of the value of ^’s stock (20 per cent of 20 per cent) is
required for a group to gain control of B and thus of A. Now suppose a new
company, C, is formed to purchase 20 per cent of Company B. . .

2 The Hypothesis of the Separation of Ownershipfrom Control. Because of diversified

ownership and the difficulty of controlling the managers through the Board of Directors,

the managers rather than the stockholders exercise effective control over the decisions of
the company.
The argument offered in support of this hypothesis is as follows. In order
to conduct the complicated business of running a large firm, a full-time pro-
fessional management group must be given broad powers of decision.
Although managerial decisions can be reviewed from time to time, they
cannot be supervised in detail. If the managerial group behaves badly it
may later be removed and replaced, but this is a drastic action and a dis-
ruptive one, and it is infrequently employed. Within very wide limits then,
effective control of the company’s activities does reside with the managers,
who need not even be stockholders. Although the managers are legally the
employees of the stockholders, they are able to remain largely unaffected
by them.

3 The Hypothesis of Intercorporate Control Groups. Whole sectors of the economy

are effectively controlled by small groups of people through the mechanism of inter-
locking directorships.

If eachmember of a small group holds directorships in several companies,

the group can control the boards of directors of many different companies
without being so obvious as to have the identical set of persons on each and
every board. By controlling the boards of directors, this group can exert
effective and relatively unostentatious control over the companies them-
selves.

This hypothesis requires not only that interlocking directorships should

be common, but also that boards of directors should be able to control the
policies of corporations in ways that would not be approved of by managers
or by stockholders. This hypothesis is in conflict with the second hypothesis,
logic, but it is limited
pyramiding of control via holding companies has no limits in
2 This
be mentioned that holding companies serve
in both law and in practicability. It should also
many purposes other than the rather suspect one described here.
 the intermediate theory of supply
for one cannot hold simuliancously that
managers take the effective de
cis.ons ignonng the interests
ofshatthoJders and directors and that direc
tors take the effective decisions, ignonng the interests of
manageis and
shareholders One can
hold cither the second or the third
hypothesis but
not both simultaneously

The Empirical Validity of the Hypotheses

These hypotheses have received a vast amount of empirical
attention since
they were first posed three decades ago ‘ The hypothesis of minonty
control
is widely accepted u is usually observed to be
, the case that the owners of a
minority of the shares in a company arc able to exercise effective control over
the election of a board of directors The hypothesis of the separation of
ownership from control is widely but not universally held The great battles
that arc required when shareholden occasionally try to oppose the policies
of managers or directon show just how hard it is for the diffused set of
owners to control directors and managers and this suggests to most
observers that there is a significant degree of control not in shareholders
hands The third hypothesis of intercorporate control groups has some ad
hcrents but is not generally accepted There is formidable evidence that
interlocking directorships exist on a large scale, ^ but there does not seem to
be any evidence that the common directors exert any significant influence
altenng the firm’s behaviour from what it would be if no such interlocking
existed

The Economic Significance of these Hypotheses

The significance of these hypotheses for economics depends on whether if

behaviour of the firm would be significantly different from what it

true, the
would be if the hypotheses were false We shall consider them one at a time

1 As far as the behaviour of the firm is concerned the hypothesis of

minority control only important if the stockholders arc able to exert a

is

significant influence on the firm s behaviour, and if the controlling

minonty
have interests and motives different from the holders of the majonty of the
the firm
firm’s stocks If all stockholders are mainly interested in having
SMtspeeiaHylhcpionwnivgttudyhyAckilfA BerJe Jr and G C Means Tht
Modern
1

romn and PrtvaU Properly (Macmilbn 1937) and R A Gordon B

srntst leadershp ,n the
Corp
(University of California Press 1945 rcpnnted
paperbackm 1961)
because they represent
2 Some individuals are dirccwn of many compatnes Thra may be
names who
banks or other investment groups that finance companies or because they are big
e
lend authority to the company or because they are merely highly valued for their a
vice
will have
mere fact that individuals hold many direcwrships means that many companies
may
at least one director in common with many other companie*
 THE ORGANISATION OF PRODUCTION 245

maximise profits, then it does not matter, as far as economics is concerned,

its

which of stockholders actually influences the firm’s policy. There is no

set
accepted evidence to show that controlling groups of stockholders do in fact
commonly seek objectives different from those sought by the holders of the
majority of the firm’s stock.

2 For the hypothesis of the separation of ownership and control to be

important it is necessary that the managers should be able to exert effective

control over business decisions and that they should wish to act differently
from the way and directors wish them to act. If the mana-
the stockholders
gers are motivated by a desire to maximise the firm’s profits - either be-
cause it is in their own interests to do so or because they voluntarily choose
to reflect the stockholders’ interests - then it does not matter that they have
effective control over decisions. If the managers wish to pursue goals other
than profit maximisation, then the behaviour of the firm will be different
according to whether the managers or the owners exercise effective control.
In Chapter 29 we shall consider a theory that takes the separation of owner-
ship and control as its starting point and that then proceeds on the assump-
tion that managers are motivated by desires other than to maximise the
profits of the firm. This is a genuine competing hypothesis and the only
way to choose between it and the hypothesis that the managers seek to
maximise the firm’s profits (either because they want to or because they are
forced to by the owners) is to confront the predictions of each theory with
factual observations and see which better predicts the observed phenomena.^

3 Since the empirical validity of the hypothesis of control through inter-

locking directorships is not generally accepted we do not need to be de-
tained long in speculating what would be if it were true. It is
its significance
perhaps just worth mentioning that the hypothesis was often advanced by
Marxists and it was presumed by them to provide significant evidence of
the domination of the whole of the means of production by a small group
of grasping, monopoly capitalists.

OTHER QUESTIONS
The economist is interested in more than the effect of the corporate form
on the decision making process of the firm. Does the corporate form con-

1 Perhaps a word or two more about the interests of managers is in order here. Stockholders

elect directors Directors are supposed to represent stockholders

who appoint managers.
the managers merely carry out. In fact, the
links
interests and to determine broad policies that
are typically weak enough so that top management often does truly control the
desdny of the
company over long periods of time. As long as directors have confidence in the managerial
elect and re-
group, they accept and ratify their proposals, and stockholders characteristically
elect directors who are proposed to them. Although the members of the top
management group
246 THE INTERMEDIATE THEORY OF SUPPLY

tnbute to more cfTicient production by fostenng the establishment of larger

firms than could otherwise be organised^ Does the company, insofar as it
encourages reinvestment of earnings, affect (for better, for worse) the alio
cation of resources between investment and consumption’ Docs the com
pany change the nature of market processes either as a seller of goods or a
buyer of factors’ Docs the company ccntndize economic power’ Some of
these questions are discussed m
deUtil later in this book

nred not be stockholden they onen do hold mable amounts of the stock in the company that

they manage this stock often being acquired as a direct result of bonuses or compensation for

their services
CHAPTER 19

THE MEASUREMENT
OF COSTS

The basic hypothesis in the theory of supply is that, in making decisions

about production, firms seek to maximise tbeir profits. The profit from pro-
duction consists of the difference between the value of the outputs and the
value of the inputs. The value of output is the revenue the firm gets from
selling its product^ while the value of input is the cost of these inputs. This
may be written

TT = R-C

where tt is profits, R is revenues and C is costs. Thus a theory of the

behaviour of profits may be broken down into a theory of the behaviour of

revehues and costs. In the next two chapters we consider the theory of costs
and in subsequent chapters consider theories of revenue.
we
A given output produced by a given technique, say 8,000 private cars
produced each week by BMC
with its present production metliods, will
have a given set of inputs associated with it, so many man hours of various
types of labourers, supervisors, managers and technicians, so many
materials, so much light and other services and so many hours of the time
of various machines. If we wish to know the cost of this diverse set of factor
separate
inputs we must value each in money terms and then add up the
costs of each factor to obtain the total cost to of BMC
producing 8,000
able put a value
private cars per week. In order to do this we need to be to

on a unit of each of the separate inputs used. It is to tliis problem of valuing

this
the inputs used in the productive process that we address ourselves in
chapter.
that the firm sells all of
1 Remember the simplifying assumption introduced on page 227,
its output.
 248 THE INTERMEDIATE THEORY OF SUPPLY

HOW DO WE MEASURE THE COST OF INPUTS?

Tht assignmcm of monetary values to phystcal quanttt.es of tnpms
may be
very easy in some cases and very hard in otheis
Furthermore, dilTercnt
people or d.lTerenl groups may assign diflerent
values to the same input
Economists might want to discuss production
behaviour of firms for a
variety of reasons

1 to predict how the firm’s behaviour wiH respond to specified changes

in the conditions it faces,
2 to help the firm make the best decisions it can in achieving its goals
3 to evaluate how well firms use scarce resources

The same measure of cost need not be correct for all of these purposes
For example, if the firmhappens to be misinformed about the value of some
resource, it will behave according to that misinformation In predicting the
firm’s behaviour, the economist should use the information the firm actually
uses, e\cn if he knows it to be based on misinformation But in helping the
firm to achieve its goals, the economist should substitute the correct
information
Economists know exactly how to define costs in order to solve problems of
she hvnd tsStd in (2) and {Z) above If we assume shai businessmen use she
same concept of costs, the economists definition will be appropriate for

problems of type (1) as well We shall make this assumption for the moment
The consequences of the assumption being m error are discussed in
Chapter 29 below
Although the details of economic ‘costing’ vary, they arc governed by a
common principle that is sometimes called user cost but is more commonly
called opportunity cost

The a particular venture is

cost of using something in
the benefit foregone (or opportunity lost) by not using
it in its best alternative use.*

firm must decide

In principle, measuring opportunity cost is easy The
monetary value
what factors of prcducSion st has used and assign to each a
equal to what it has sacrificed order to have the use of the factor When
m
some tough
we come to apply this principle to specific cases, however,
problems arise
firm hires or buys, such
In the case of those factors of production that the
spends £5 on
as labour and materials, there is no problem If the firm
could have purchased
hiring some factor, it sacrifices all the other things it
said the idea of
1 way
Th.j u just another putimg the point made on page 60 when we
of
opportunity cost was to measure the cost of something m terms of foregone alternatives
 THE MEASUREMENT OF' COSTS 249

with the £5', thus the price paid for the factor measures the opportunity
cost of employing it. The price per unit is clearly known and, since pay-
ments are made explicitly, the cost will be deducted from its revenues in
assessing the profitability of a given line of production.
In the case of factors of production that the firm itself owns, the cost must
also be assessed but, since no payment is made to anyone outside the firm,
these costs are not so obvious. Such costs are called imputed costs. If the
most profitable lines are to be discovered, the cost of these
must be factors
reckoned at their current market prices. That such amounts do represent
can be seen by the fact that the firm could earn a revenue
costs to the firm

by leasing someone else such factors as it owns. If, for example, a firm
to
uses £10,000 worth of its own money that could have been loaned out to
someone else at 5 per cent, yielding £500 per year, then this £500 should
be deducted from the firm’s revenue as the cost of funds used. If, to continue
the example, the firm makes only £400 over all other costs, then we should
say not that the firm made £400 but £100, for if it closed
rather that it lost

down completely and merely loaned out its capital to someone else, it could
have earned £500. Similarly, if the owner of a firm also acts as its manager,
‘

he should impute as costs to the firm a management fee equal to what he

could obtain if he leased his services to some other firm. If the firm uses its

own it should charge as a cost a fee reflecting the

plant and equipment,
market value of these services. (After all, the firm could cease operations
and rent its plant and equipment to someone else.)
The most difficult problem of imputed costs concerns the evaluation of
the service of risk-taking. Business enterprise is often a risky affair. Some
enterprises are more risky than others and someone must take the risk in

each case. The borne by the owners of the firm who, if the
risk is in fact
enterprise fails, may lose the money they have invested in the firm. The
owners will not take these risks unless they receive a remuneration in return.
They must expect to receive a return in excess of what they could have
obtained by investing their money in a virtually riskless manner, say by
buying a Government bond. In the sense in which we have used the term,
risk-taking is a factor of production. a service that must be provided if
It is

the firm is to carry on production and it must be paid for by the firm.

If a firm does not yield a return sufficient to compensate for the risks in-

volved, the firm -will not be able to persuade people to contribute money to

it in return for a part-ownership in the firm.

Charging the market value of the capital is appropriate if the firm can obtain all the funds
1

that it requires by borrowing. If, as is sometimes the case, this is not so, then the
firm will place

a high value on the funds that it does have. Such a firm must cost its capital by looking at
the

inability to raise all the capital it wants

other ventures it might have undertaken, since its

means that it will be unable to do all the things it would like to do.
250 THE INTERMEDIATE THEORY OF SUPPLY

Raw matenals, supplies, labour, use of buildings and

machinery, use of
capital, services of managers, and risk-taking, plus many
others areVactors
of production that are used in producing the firm’s
output If the> ha\e
alternative productive uses (and they usually do), the firm
must charge
Itself for their use,
otherwise it cannot determine whether or not it is making
the most profitable use of the resources it controls or purchases

Two Common Errors in Calculating Of^rtunity Cost

The most frequent sources of error in calculating opportunity cost are to

assume that the cost of something is always related to what has to be paid
out currently before the factor can be used and to assume that the oppor-
tunity cost of a factor owned by the finn is related to what the factor cost
when It was originally purchased by the firm

The opportunity cost of a factor owned by a firm does not

E qUAL what the FIRM MUST CURRENTLY PAY OUT TO SECURE THE USE
OF THE FACTOR Usually this current payment for a factor owned by the
firm IS zero while the opportunity cost is correctly measured by the market
price of the services of the factor If the firm uses its own money capital this

IS not free because the firm is sacrificing what it could have earned by lend-
ing the money to someone else We have already discussed this problem
above but n is worth listing n again here as one of the two most prevalent
errors in calculating opportunity cost

The opportunity cost of a factor owned bv the firm bears no

NECESSARY RELATION TO THE PRICE THAT WAS PAID WHEN THE FACTOR
WAS ORIGINALLY PURCHASED Probably the most prevalent source of error
in calculating opportunity cost is to take historical cost into consideration
This is the error that seems to die hardest and which still pervades much of
Let us consider in lerms of a simple -vnd extreme
business practice it

example where the pnnciplc should be obvious

ago
Assume that a firm has a set of machines that it purchased some time
years so the firm
for £100,000 These machines have an expected life of ten
per year
calculates the ‘depreciation cost of these machines at £10,000
product an
Further assume that the machines can be used to make one
s plant t cy
nothing else Since they are installed lO one part of the firm
value is negligi
cannot be leased to any other firm and their scrap
^
cost o a
Assume that if the machines are used to produce this product the
produced can
other factors utilised will amount to £25.000 while the goods
be sold for £29,000 u total
t

Now if we add in the dcpreaauon ‘costs’ of running the machine the

, i
 THE MEASUREMENT OF COSTS 251

cost of operation comes to ;(;35,000 and with revenue at ;^;29,000 this makes
an annual loss of ,{^6,000 per year. Surely the goods should not be made!
The fallacy in this argument lies in adding in a charge based on the
historical cost of the machines as one of the costs of current operation. The
machines have no alternative uses whatsoever. Clearly their opportunity
we have defined it, is zero. The total costs of producing this line of
cost, as

goods is thus only ^^25,000 per year (assuming all other costs have been

correctly assessed) and the line of production shows an annual profit of

;()4,000, not a loss of /)6,000.
Which calculation is correct? If the firm decides this line of production
is unprofitable and does not continue it, it will have no money to pay out
and no revenue received on this account. If the firm takes the economist’s
advice and pursues the line of production it will pay out ^025,000 and
receive ^29,000 thus making it ^4,000 per year richer than if it had not
done so. Clearly this line of production is profitable. Realising this depends
on appreciating that the amount the firm happened to have paid out for
the machines in the past is of no relevance whatsoever in deciding the
correct use of the machines once it has installed them on the premises.
Put in the terms of a maxim of behaviour we can say: ‘byegones are
byegones’, and they should have no influence on deciding what is currently
the most profitable thing to do. This is a very important proposition and we
shall return to itagain and again, and in the meantime we may note that
one of the great deficiencies of current practices amongst accountants is the
frequent failure to observe this rule and the giving of advice on current pro-
duction decision that takes past expenditure into account.* To reiterate our
general rule:

The opportunity cost of using any factor is what is

currently foregone by using it. With factors currently
obtained from outside the firm this cost is measured
by the price currently paid for their services, with
factors already owned by the firm this is measured by
the amount for which the factors could be hired out
or sold to another firm at present.
ignored
1 a very important principle that extends well beyond economics and is often
This is

in these other areas as well.In many poker games, for example, the cards are dealt a
round at
a time and betting proceeds after each players’ hand has been augmented by one card. Players
who bet heavily on early rounds because their hands looked promising often stay in through
later rounds on indifferent hands because they ‘already have such a stake in the pot
The pro- .

fessional player knows that after each round of cards his bet should be made on the proba-

bilities that the hand he currently holds winner when all the cards have been
will turn into a
constitutes a
dealt. If the probabilities look poor after the fourth card has been dealt (five
complete hand), he should abandon the hand and refuse to bet further whether he has put 5s

or £5 into the pot already. amateur who bases his current decisions on what he has put
The
 '

252 THE INTERMEDIATE THEORY OF SUPPLY

Payments to Capital Costs or Profits ?

We have defined be the excess of revenues over opportunity costs

profits to
whatever the source of such excess This is a technical and specialised
definition of a word that is m
everyday use It is, thercfoie, a potential
source of confusion to the student who runs into other uses of the same word
The businessman defines profits as the excess of revenues over the costs
with which his accountant provides him The major differences are that
since the accountant does not include as costs charges for nsk-taking and
use of the owners own capital, these items arc recorded by the businessman
as part of his profits When the businessman says he needs profits of such
and such an amount merely to make it worth while staying in business at
all, he IS making sense within his definition For his profits must be large

enough to pay for those factors of production that he uses, but are not costed
by his accountants
The income tax authocuics have yet another defimtton of profits, which
IS implicit m the thousands of rules as to what may be (and what may not

be) included as a deduction from revenue in arriving at taxable income In

some cases, the taxing authoniies allow more for cost than the accountant
recommends, in other cases they allow less
Economists themselves often use the term profits in different senses We
have included in business costs the opportunity cost of capital and ofnsk
taking, 1 e what could be earned by supplying the capital to other users
,

plus a margin for any extra risk in this particular industry Clearly, if the
enterprise cannot yield revenues equal to these costs it will not be continued
and the capital will be transferred to other more remunerative uses In some
treatments this opportunity cost of capital is called normal profits The differ-
ence between the two treatments is purely semantic m one case we say that
the enterprise will not be continued unless revenues cover ail costs, m the
other wc say the enterprise will not be continued unless it covers costs plus

normal These arc two ways of saying the same thing for a particu-
profits
lar line of endeavour to be worth continuing it must cover the opportunity
could
cost of all the factors involved (i e , they must earn at least what they
cost of
earn mtheir best alternative use), whether we call the opportunity
capital a cost or a normal profit is onlya matter of words
be clear about different meanings of the term
profits
It IS important to

loser if he plays in rational

com
into the pot in earlier rounds of betting will be a consistent
to take account of them in
pany In poker war and economics byegones are byegoncs and
current decisions is to court disaster'
ere
1 In the first edition I used the term nonnal profits I have changed terminology
because it emphasises both the cost aspect of these profits
and the similarity with other costs
Equivalents between the two approaches are as follows
taking Super normal profits —
pix*
Normal profits = opportunity cost eif capital and
 THE MEASUREMENT OF COSTS 253

not only to avoid fruitless semantic arguments but also because a theory
that predicts that certain behaviour is a function of profits defined in one
way will not necessarily predict
behaviour accurately given some otlier
definition.For example, if the economist predicts that new firms Mill enter
an industr>' whenever there are business profits, his prediction will fre-
quendy be wrong if he is working from the businessman’s definition of
profits. Our definition of profits is for manypurposes the most useful, but
the student who washes to apply it to business behaviour or to tax policy
must be prepared to make the appropriate adjustments.

SOCIAL AND PRIVATE COSTS

In this chapter we have discussed the cost calculation of individual firms.
The firm assesses the opportunity cost of factors according to the market
market prices of the resources used reflect
prices of these factors. If the
adequately the alternative uses of the resources from the whole society’s
point of view, then the firms decisions will be based on a consideration of
the value of the alternative uses of these resources in the whole society. There
are many reasons why this may not be so and this leads us to distinguish

two concepts of cost.

Private cost: Measures the value of such alternative uses of the resources
that are available to the firm and is based on the market value of the factors
currently purchased by the firm and the price that could be obtained by
selling to outsiders the services of factors owned by the firm.

Social cost: Measures the value of such alternative uses of resources that
are available to the whole society.
Let us consider a number of cases where these two costs do not coincide.
1 A factory located on a river discharges its waste into the river and the
resulting pollution destroys fish liv'ing downstream and forces several com-
munides both to install costly water purification plants before water from
the river can be used and to build swimming pools since the river is no
longer safe for swimming. The private cost to the factory^ of using the river
as a waste disposal unit is zero since there is no alternative use for the river

to the factory and no price has to be paid for the privilege. The social cost
consists of the foregone fish, the resources used in the purification plants and
the pools (plus the costs incurred by a host of other consequences
swimming
of water pollution) The social cost of using the river to dispose of waste
.

may well be greatly in excess of the social cost of having the factory s
waste disposed of in a chemical treatment plant. Yet there MU be no
incentive for the private firm to adopt what is socially the least costly

method since the private costs of the two alternative methods do not reflect

the social costs.

 254 THE INTERMEDIATE TIIEORV OF SUPPLY
2 An industnal complex discharges smoke into the atmosphere, pnvate
cost zero, social cost, when done on a large scale prodigious TTic total
social cost to Britain of the millions of tom of industrial waste poured into
the atmosphere over llit last 150 years staggers the imagination It is to be
measured in such d:\erse terms as extra resources devoted to washing
and
cleaning shirts and sheets svhich soil more rapidly, the cost of sandblasting
cathedrals to remove the soot and gnme, extra doctor and hospital services
to treat the chronic diseases caused by such air pollution (to say nothing
of the human cost m terms of suffering and earlier death for which no
money price can be calculated), extra services of many sorts to deal with
the smog blankets that paralyse whole cities penodtcally The list can be
extended almost indefinitely
3 The nationalised telephone authority in a mythical country called
Bntonia initiates a nationwide automatic dialling system (called DTS)
with inadequate and inefficicm equipment As a result, millions of hours arc
spent by subscribers, many of whom are the employees of business firms
trying to complete calls which cither go astray or encounter fully engaged
signals on the overloaded lines The telephone company finds that it has
saved thousands of houn of telephone operators’ time, the cost of which
exceeds the cost of the automatic system Pnvate cost to the telephone
authority of the automatic system is less than that of the old long*distance
calls*through-ihe-operaior system Tlie social cost, however, is vastly in

excess of private cost because of the man hours wasted by subscribers trying
(with much less specialised efficiency than the operators could bring to
bear) to complete their own calls in the face of a bad system One harassed
Bntonian manager was overheard to say ‘Before I ask my secretary to get
me someone in another city on the ’phone 1 have to ask my self if she really
’
has 15 minutes to spare for the purpose
4 A pnvate timber firm cuts a forest which had an alternative of pro-
viding a nature sanctuary and a nitiona! recreation area The citizens
valued the park highly but there was no way for the firm to bnng this
sold
alternative use into its calculations since public parks are not bought or
on any market
nafcmri
5 Another timber firm cuts a forest and thereby destroys a
watershed bringing drought and destruction on neighbounng farms Social
cost exceeds the pnvate cost by all the foregone agricultural output, plus
resources subsequently used to provide a stable water supply to the farms
We shall return to these and further examples when we discuss micro
economic policy Chapter 38 In the meantime this should emphasise the
m
and the potential
possibility of divergencies between pnvate and social costs
attention
senousness of such divergencies In the meantime we shall focus
on the decisions of individual firms responding to their own pnvate
costs
APPENDIX TO CHAPTER 19

BALANCE SHEETS, INCOME

STATEMENTS AND COSTS OF
PRODUCTION: THE VIEWS
OF THE ECONOMIST AND
THE ACCOUNTANT^

Accounting isa major branch of AN EXAMPLE

study in and of itself.Many students of
economics will want to study account- Mr James Maykby, the
Late in 1965.
ing at some stage in their careers. It is
Deputy Chairman of the Acme .Artifi-
not possible to give a short course in cial Flower Company (at a salary- of
accounting in this appendix, but rather ;025.000 per year), decided he w'ould
to acquaint the reader with the kind of go into business for himself. He quit
summary statements that are used by
his job and organised the Maykby Leaf
both economists and accountants. Bal- Company. He purchased suitable plant

ance sheets report the picture of a firm and equipment for ^80,000 and sup-
at a moment in time. They balance in the plied himself withsome raw materials
sense that they show the assets (or and supplies. By 31 December 1965,
valuable things) oiv-ned by the firm on he was in a position to start manu-
one side and the claims against those facturing. The funds for his enterprise

assets on the other side. Income state-
ments do not refer to a moment in time
but to a period of time (like a year), and
report in summary fashion the flows of
resources through the firm in the course
of its operations. Balance sheets thus
were ^40,000 raised as a bank loan on
the factory' (on which he is obligated to
pay interest of £2,400 per year) and
£55,000 of his own funds, which had
previously been im'ested in common
stocks. He also owed £5.000 to certain

measure a stock, income statements a firms that had provided him with
flow. supplies.

In order to illustrate what balance Maykby. ivho is a trained account-

sheets and income statements are we ant, drew up a statement of his

shall treat the same example from two

points of view: that of the accountant 1 This .AppendLx can be omitted sv-ithout

and that of the economist. interrupting the flow of the argument.

 1

256 THE INTERMEDIATE TMEORy'oF SUpfLY

Table 19 A
MAYKBY LEAF COMPANY, ACCOUNTANT’S BALANCE
SHEET, 31 DECEMBER 1965

Assets Liabilities and Equity

Cash m bank £ 5,000 Owed to suppliers
1 5,000
Plant and equipment 80,000 Bank loan
Raw matenals and Equity 55 000
supplies 15,000
Total liabilities and
Total assets £100,000 equity £100,000

Table 19 A2
MAYKBY LEAF COMPANY, ACCOUNTANT’S BALANCE
SHEET, 31 DECEMBER 1966
Cash
1 in bank (see Owed to supplicn of
Exhibit 1) £ 32,600 factors (see Exhibit 4) £ 10 000
Plant and equipment (see Bank loan 40,000
Exhibit 2) 73,000 Equity (see Exhibit 5) 70,600
Raw materials and supplies
(see Exhibit 3) 15,000 Total liabilities and
equit) £120,600
Total assets £120,600

company’s position as of 31 December amount of ;^60,000, of which it still

^
1965 (sec Table 19 Al, above) owed £10,000 at the end of the year
Mr Maykby showed this balance 2 The firm manufactured artificial
sheet to his brother in-/aw, an econo* leaves and flowers whose sale \alue was
mist, and was very pleased and sur- £100,000 At year’s end it had sold all
prised' to find that he agreed that this of these, and still had on hand £15,000
was a fair and accurate statement of worth of raw matcnals
the position of the company as it 3 The firm paid off the £5 000
of
prepared to start operation owed to suppliers at the beginning
During 1966 the company had a the year
busy year, hiring factors, producing 4 At the very end of 1966, the com-
and goods, and so on The
selling pan) purchased a new machine for
following points summarise these acti- £5,000 and paid cash for u
vities of the 12-month period 5 It paid the bank £2,400 interest
The firm hired labour and pur- on the loan
chased additional raw materials in the 6 Mr Maykby paid himscli

2 In example we will treat all purchased

I He usually finds that he and his brother in this

ia'v disagree about everything and hired facton in a single category

 BALANCE SH’EETS AND INCOME STATEMENTS 257

Table 19.A3
EXHIBITS TO BALANCE SHEET OF 31 DECEMBER 1966

Exhibit 1 . Cash
Balance, 1 Januan' 1966 £ 5,000
+ Deposits
Proceeds of sales of goods 100,000 ;^105,000

—Payments
Payments to suppliers (1965 bills) 5,000
Payments for labour and additional raw
materials 50.000
of Mr Maykby
Salar)' 10.000
Purchase of new machine 5,000
Interest payment to bank 2,400 72,400

Balance, 31 December 1966 £ 32,600

Exhibit 2. Plant and Equipment

Balance, 1 January 1966 r 80,000
+New machine purchased 5,000 £ 85,000
— Depreciation charged 12,000

Balance, 31 December 1966 £ 73,000

Exhibit
3. Raw Materials and Supplies
On hand 1 January 1966 £ 15,000
Purchases in 1966 60,000 £ 75,000

Used for production during 1966 60,000

15,000
On hand 31 December 1966

Exhibit 4. Owed to Suppliers

Balance, 1 January 1966 £ 5,000
New purchases, 1966 60,000 £ 65,000

Paid on old accounts 5,000

50,000 55,000
Paid on new accounts
Balance, 31 December 1966 £ 10,000

Exhibit 5. Equity
Original investment £ 55,000
Plus income earned during year (see
15,600
income statement)
Balance, 31 December 1966 £ 70,600

9
258 THE INTERMEDIATE THEORY OF SUPPLY
;^10,000, ‘instead of salary' (See machine, which is an exchange of
Table 19 A3 assets - cash for plant and equipment -
and which will be entered as a cost in
the income statements of some future
AN ACCOUNTANT'S BALANCE
periods as depreciation is charged, and
SHEET AND INCOME STATE- the pa>mcnt of past debts, wbch
MENT entered the income statements m the
Taking account of all these things and penod m which the things purchased
also recognising that he had deprecia- were used in production ^
tion on his plant and equipment,’ he Second, note that the net profit from
spent Ne'v Year s Day, 1967, preparing operations increased the owner’s
three financial reports (See Tables equity since it was not ‘paid out’ to
19 A2, 19 A3 and 19 A4 ) him A loss would have decreased his
These accounts rellect the operations equity
of the firm as described abo\e The Third, note that the income state

Table 19 A4
MAYKBY LEAF COMPANY. ACCOUNTANT’S INCOME
STATEMENT FOR THE YEAR 1966
Sales £100 000

Costs of operations
Hired services and raw materials used £ 60,000
Depreciation 12,000
Mr Maykby 10,000
Interest 2,400 84,400

Profit £ 15,600

book-keeping procedure by which ment, covering a year’s operations

these various activities are made to provides a link between the opening
yield both the )
ear-end balance sheet balance sheet (the assets and claims
and the income statement need not against assets at the beginning of the
concern you at this time, but you should year) and the closing balance sheet
notice several things Fourth, note that every change in a

Tirsl, note that some transactions balance s'hee t belw cen tw o datt* t.'aTi bt.
afiect the balance sheet but do not accounted for by events that occurred
enter into the current income state- during the year (See the exhibits to the
ment For example, the purchase of a balance sheet, Table 19 A3 )

1 The tax people told him he could charge 15
per cent of the cost of his equipment as depreoa
tion during 1966 and he decided to use ttus
amountmhisO'vnbooksaswtU No depTetiaUOB
was charged on the new machine
difficulty
2 Beginning students often have
and la
with the distinction between cosh fiovvs
item the
come flows If jou do analyse item by
Table
Himes in Exhibit 1 (Table 19 43) and in
19 A4 the income statement
 BALANCE SHEETS AND INCOME statements 259 I

After studying these records, Mr since that is what he could have earned
Maykby feels that it has been a good outside.
The company has money in the
year. 2 Maykby should have charged the
bank, has shown a profit, and it was
it company for the use of the ^^55,000 of
able to sell the goods it produced. He his funds. He computes that had I

is bothered, however, by the fact that Maykby left these funds in the stock
he and his wife have felt poorer than in market he would have earned £5,500
past years. Probably the cost of living in dividends and capital gains.
has gone up! 3 He tells Maykby that the de-
preciation figure is arbitrary'. The
plant and equipment purchased for
AN ECONOMIST'S BALANCE
;{^80,000 a year before now has a
SHEET AND INCOME STATE- market value of only ;^62,000. (Assume
MENT he is correct about this fact.)
When Maykby’s brother-in-law reviews The brother-in-law prepared two
the 31 December 1966 balance sheet revised statements. (See Tables I9.A5
and the 1966 income statement he and 19.A6 and the exhibit, Table
criticises them in three respects. He 19.A7.)
says: It is not hard for Mr Maykby to
1 Mr Maykby should have charged understand the difference between the
the company ^25,000 for his services. accounting profit of ;^1 5,600 and the

Table 19.A5
MAYKBY LEAF COMPANY, ECONOMIST’S INCOME
STATEMENT FOR THE YEAR 1966
Sales £100,000
Cost of operations
Hired services and raw materials £ 60,000
Depreciation^ 18,000
Interest to bank^ 2,400
Imputed cost of capital 5,500
Services of Maykby 25,000 £110,900
Loss £ 10,900

Table 19.A6
MAYKBY LEAF COMPANY, ECONOMIST’S BALANCE
SHEET, SrDECEMBER 1966
Cash £ 32,600 Owed to suppliers £ 10,000
Plant and equipment 67.000 Bank loan 40,000
Raw materials, etc. 15.000 Equity (see Exhibit A) 64,600

£114,600 £114,600

1 Market value on January 1 less market value on December 31.

2 Because the bank loan is secured by the factory, its opportunity cost seems to the economis'
properly measured by the interest payment.
 THE INTERMEDIATE THEORY OF SUPPLY

Exhibit
EXHIBIT TO BALANCE SHEET,
A Equity of Mr
Original investment
New investment by Mr
Maykby

Maykby
Table 19 A7
31 DECEMBER
— 1966

£ 55 000

Salary not collected

Return on capital not collected
£ 15,000
sisoo

Less loss from operations

£ 75,500

10,900

£ 64,600

Table 19 AS
MAYKBY'S SITUATION BEFORE AND AFTER
As stcond As
Vsee Pusident OantT-Manager
Acme Flouer of
Company Maykby Company Difference

(1) m (2)-(l)

Salary paid £25,000 £10,000 -£15,000

Earnings on capital,
invested m stocks 5,500 - 5,500

Assets owned 55.000 64,600 (Equity + 9,600

(Stocks) in Maykby Co
Net change -£10,900

reported economist’s loss of £10,900 brother-in-law prepared the report

The difference of £26,500 is made up shown m Table 19 A8
as follows Although Maykby spent the after-
Extra salary
noon muttering to himself and telling
£15,000
his wife that his brother in law was not
Imputed cost of capital 5,500
only totally lacking in any business
Extra depreciation 6,000
sense but unpleasant as v\cll he was
observed next morning at the public
£26,500
bbrary asking the librarian whether
What Maykby does not understand there was a good ‘teach-> ourself book
IS in what sense he lost £10,900 dunng on economics (Her answer is not
.the year In order to explain this his recorded
CHAPTER 20

THE VARIATION OF
COST WITH OUTPUT

In Chapter 19 we saw' how to evaluate the cost per unit of each of the types
of factors used in the productive process. In this chapter w'e w'ish to study
how the cost of production varies as the level of output varies. At a formal
level we do this in three short steps: first, we define a relation between
inputs used and output obtained ; second, we state a simple rule for choosing
the method to be used in producing each level of output and third we see
how cost varies as output varies.

THE PRODUCTION FUNCTION

The relation between inputs and outputs is summarised in what is called
the PRODUCTION FUNCTION. This is a technological relation show'ing /or a
given slate of technological knowledge how much can be produced with given
amounts of inputs. We may write this function as

where Q is the quantity produced per period of time and where/j, . .

.,f„
are the physical quantities of m different factors used. If there is an im-
provement in our technical knowledge we will generally be able to obtain

more output for the same quantity of inputs so that the function itself will

change. We may summarise this important concept as follows:

The production function shows for a given state of

technological knowledge the relation between physical
quantities of inputs and the physical quantity of out-
put achieved per period of time/
1 Some readers may find it helpful if we illustrate the more general discussion of the text

with a numerical example. Consider the production function Q — -\/fif2 which reads Q equals
262 THE INTERMEDIATE THEORY OF SUPPLY
many ways of achieving the same total
Since there are output we need
some method of choosing between them (See page 229 if you have for-
gotten this important point ) The hypothesis of profit maximisation pros ides
a simple rule for this choice any firm that is trying to maximise its profits
should select the method of produang a given output that produces U
at
the lowest possible cost This is an implication of the hypothesis of profit
maximisation and we term it the implication of cost minimisation It
can be stated formally as follows

For any specific output the firm chooses the least

costly way of achieving that output from the alterna-
tives open to it.

If there is a known stable required output rate and if the costs of factors
are known, this is all there is to it, and for the moment it is convenient to
'
assume that this is so

the square root of/( umes^^ and where Q is output of some commodity measured m thousands
of tons per month and where /, and /j are the inputs of the only two factors used to make this

commodity, measured in each case as so much per month $ay we wish an output of 6 000 tom
per month This can be obtained using 36 units of/, and 1 unit of/t but as the following table
shows It can also be obtained by using many other combinations of the two fbctori

If the state of technology chaeges so that wc became more efficient in our use ot both taevsss

the production function might for example change to

«= llv^
month
in which case each of the sets of input* in the table above produce 6 600 tons per
instead of 6 000
AnumencalcxamplchasaIreadybecngiveni5Tablel72 page232 Tabngthenumen
1*

cal m note 1 on page 261 yireieetl«itatp,»p,=£200 the lowest cost factor comb.oa
example
should cost the
tion in the tableu 6/, and 6/^ which gives a cost of £2,VXi per month (You
so ) Checking for a
other methods of producuon shown m ihe table to aanafy yourself that this w
 THE VARIATION OF COST WITH OUTPUT 263

THE COST FUNCTION

Once \ve have decided how to value the inputs used and have a criterion
for selecting the method of producing each
output we can pass level of
easily enough from the production function to the cost funcdon. We dis-
cover for given factor prices the lowest-cost method of producing each level
of output and by relating the actual cost to the level of output we obtain a
cost function which we can define as follows ‘

The shows the cost associated with each

cost function
on the assumption that the lowest-cost
level of output
method has always been chosen from the ones avail-
able.
smaller variation than is shown in the table: if /, is increased to 6-1 units yi falls to 5.9016

(approx.) in order to keep output constant. This raises total costs to ;^2,400 6s 2d (approx.).
If /i is lowered to 5-9, goes up to 6.1017 (approx.) and total costs rise to £2,400 6s lOd
(approx.).
The price of/j now rises to £450 per month, constant at ;{(200 a similar set of calculations
shows that the lowest cost method of production shifts to 4fi and 9_^ making total costs £3,600
per month. Notice that the cheaper factor has now been substituted for the more expensive one.
1* In the numerical example it can easily be shown that the least-cost factor proportion de-

pends only on relative factor prices and does not shift as total outpbt varies. When the tivo

factors areboth priced at £250, the least-cost factor proportions are /, —/2 . Total costs of
fnoduction are given by
rC = 250(/,+/2), (I)

but keeping to least-cost proportions (i.e., always having/i=/2 ) this gives

rC= 500/,. (2)

Output is given by

<2 = y/AA, (3)

which sticking to optimal proportions can be rewritten

Substituting (4) into (2) gives us our cost function

TV = 500Q. (5)

Thus total costs are a linear function of output. To get average costs per unit of output we
divide total costs by total output:

TC 500?
6)
500. (

Thus average costs per unit of output do not change as output varies.

The student can easily repeat these calculations for the case when the price of/, rises to /450
and the factors are used in the proportions

^=1
/a 9
 264 THE INTERMEDIATE THEORY OF SUPPLY
So far we have been talking at a vcr) formal level and we must now
see
ss hat is actually im olved in varying alt Gictor inputs as output
is vancd

REAL CHOICES OPEN TO THE FIRM

Let us consider some of the real decisions that must be taken by firms
If the rate of sales has fallen off should production be reduced
cotres
pondmgly or should it be held at its old rate with the unsold amounts
being stored up against an antiapatcd nse in sales in the future’ If pro
duction IS to be reduced how can this be accomplished most cheaply’
Should one whole plant be closed down or should all the plants be operated
on short time’’ If demand increases sharply and unexpectedly how can
more production be squeezed out of the existing facilities ’ It will take time
to build new plants and in the meantime something must be done to meet
the new demand All of these matters concern how best to use the existing
plant and equipment They concern time periods too short to build new
plants or to install more equipment The
made can be impledecisions
mented quickly overtime can be increased tomorrow and new worken can
be added to production as soon as they can be hired and trained
More ueighty decisions must be made when the managers consider a
longer time horizon Should the firm adopt a highly automated process that
will greatly reduce its wage bill even though it will have to borrow large
sums of money to buy the equipment’ Should u carry on instead with the
same kind of techniques it is now using’ Should it build new plants in an
area where labour is plentiful possibly even moving abroad to locations
where labour is both plentiful and cheap (but possibly not so efficient) and
adopt techniques that replace some of the more costly mechanised processes
with labour’ Should it tell the engineers who are designing its plant to
worry only about making costs as low as possible when the plant is working
to capacity or should it tell them to worry also about making the plant as
flexible as possible so that output can be \ancd over a wide range at a
reasonable level of costs’ These matters concern what the firm should do
when It IS changing or replacing its plant and equipment Decisions on them
may take a long time to implement
In the above cases the managers are making decisions from
known
possibilities But large firms also have staffs working on research and
money
development The managers of these firms must decide how much
to make available to their research and development staff in the
hope that
doing
they will come up with some wholly new and more efficient ways of
things But they must first decide in what areas the payoff for
new develop
1 If ihe firm chooses to close down one whole phut and th s plant is tl e major source of
employment in a town it w U cm cally afifect the fete of the people I ving in that town
 THE VARIATION OF COST WITH OUTPUT 265

merits will be largest and then direct their research staff to work in these
areas. If, example, a shortage of a particular labour skill or raw material
for
is anticipated, the research staff can be told to tr>' to find ways of econo-

mising on that labour skill or raw material. It is difficult, of course, 'to

anticipate particular labour shortages far in advance, but material shortages
can more often be foreseen.
What all this adds up to is that the firm is constantly making decisions;
some apply to today, some
tomorrow, and some to the far distant future.
to
These decisions are made in an uncertain world, and the firm can never
be sure that, when a particular decision finally goes into effect, the circum-
stances that gave rise to it will still make it an appropriate one. When

making such major concern of the firm is how it can raise the
decisions, a
probability that the anticipated range of output will, when the time comes,
be produced at the lowest possible cost.
These decisions are complicated by the fact that some inputs, such as
labour and raw materials can be varied on quite short notice, possibly a
matter of days or weeks at most, while others can only be varied with a
long time lag - it may take years to design, build and bring into operation
a new, fully-equipped plant.

Time Horizons for Decision-Making

In order to reduce these many manageable proportions,

decisions to
economists organise them into the three groups that we have already
distinguished. We think of the firm as making decisions about (1 j how best
to employ its e.xisting plant and equipment; (2j what new plant and equip-

ment and production processes to select, given the framework of known

technical possibilities; and (3) what to do about encouraging the invention
of new techniques. Corresponding to these three groups of decisions we
define three time periods; (1) the short run, which is a period of time over
which decisions are restricted by the fact that certain factors of production -
usually plant and equipment — cannot be increased in supply; (2) the long
run, which is a period of time over which all factors of production can be
varied but in which decisions are restricted by the existing state of knowledge
of technical possibilities; (3) the very long run, which is a period of time
over which decisions made now about such things as research and develop-
ment can change the supply of available techniques. These three time
periods are theoretical constructions. They abstract from the more compli-
cated nature of real decisions and focus only on the key factor that restricts
the range of choice in each set of decisions. Whether or not this theoretical
abstraction is useful depends on whether or not it can be used to develop
a theory that successfully predicts what choices firms will actually make in

9*
 266 THE INTERMEDIATE THEORY OF SUPPLY
vanous real situations We must now consider these three periods in more
detail

The SHORT RUN IS defined as the penod of time over

which the inputs of
some factors cannot be varied The most usual meaning is that a firm «
committed to paying for the use of a spiecified quantity of fixed factors
whether or not it needs them, and that it cannot get the use of more of
it

this factor it has on hand

than This is the meaning we shall use The
factors that can be \aned m
the short run are called variable factors
The factor that is fixed in the short run is usually an clement of capital
(such as plant and equipment), but it might be land, or the services of
management, or even the suppl> of skilled salaried labour What matten is
that at least one significant factor should be fixed
The short run does not correspond to a fixed number of months or ^ can
In some industries it may extend over many years, m others it may be only
a matter of months or even weeks In the electric power industry, for
example, where it takes three or more years to acquire and install a steam-
turbine generator, an unforeseen increase m demand will involve a long
penod during which the extra demand must be met as best as can be i^ith
the existing capital equipment The other side of this com is that, because
this equipment has a very long life, a decrease m demand leaves the firm

committed to all the costs of this equipment that do not vary with output
In the hydroelectnc industry, the short run may extend over decades At
the other end of the scale, a machine shop can acquire new equipment (or
sell existingequipment) in a matter of a few weeks, and thus the short run
IS correspondingly short If there is an increase in demand, it will ha\e to be
met With the existing stock of capital Ibr only a short time, after which it
will be possible to adjust the stock of equipment to the level made desirable
*
by the higher demand

The LONG RUN IS defined as the period long enough for the inputs of all
factors of production to be vaned, but not so long that the basic technology
it docs not corr«pond to a specific
period of
of production changes Again,
time, but varies among industries
that
The special importance of the long run in production theory is it

corresponds to the situation facing the firm when it is planning to go into

business , or to expand substantially the scale of its operations , or to branch

out into new products or new areas, or to modernise, replace, or reorganise

Its method of production

example, the
I The length of the short run is not fixed solely by technological factors For
pnce the rm is
speed of debvery and installation of new etpupment can be influenced by the
willing to pay
 THE VARIATION OF COST WITH OUTPUT 267

The planning decisions of the firm characteristically are made ^\'ith fixed
technical possibilities but with freedom to choose whatever factor pro-
portions seem to be most efficient. Once planning decisions are carried out
- once a plant is equipment purchased and installed, and so on - the
built,

firm has fixed factors and makes operating decisions in what w'e have called
the short run.

Unlike the short and the long run, the very long run is concerned with
situations in which the technological possibilities open to the firm are subject
to change. An important characteristic of industrial socieW over the recent
centuries has been the change in technology' that leads to new and improved
products and new and improved methods of production. These changes
may be by what the firm itself does, particularly by its programme
affected
of research and development. While decisions with respect to research and
development may affect the techniques of production used (and hence the
costs of production), they are of a different kind from short-run and long-
run decisions. They are discussed in Chapter 21. For the remainder of this
chapter we ignore them and treat choice of techniques of production under
the assumption of a constant, known technology.
We now return to our discussion on page 261 and ask how the firm’s costs
will vary' wth output over the short and the long run periods. Our very'
brief earlier discussion concerned the long run since we were implicitly
assuming that all factor inputs could be varied. We shall take up the dis-

cussion from there.

LONG-RUN COST BEHAVIOUR: FIXED FACTOR PRICES

We now assume explicitly that sufficient time is allowed to elapse for the
firm to make all of the adjustments that it requires. If the rate of output
is low, the firm will have a small plant and will have its whole production
process organised accordingly. If the rate of production is high, then the
firm will have a large plant and will have all its organisation geared to a
large scale of production: the firm may have many separate plants pro-
ducing its main product and may have subcontracted the production of some
of the components of its product out to specialised firms.
Whatever the
actual organisation, there will be some best method of organising pro-
duction for each given level of output and some corresponding level of costs.
Costs could be higher if the best technique of production were not used,
but they could never be lower than the costs associated with the best avail-
able technique.
Figure 20.1 is a graph of a long-run cost curve. The rate of output is
 268 THE INTERMEDIATE THEORY OF SUPPLY
measured on one axis and the cost per unit of output (i e total
cost divided
,

by the number of units) is measured on the other In order to obtain

a long
run average-cost curve, \ve plot, against any given rate of
output, say OB
the average cost that results from dvc selectitm of the lowest cost
method of
production In this case the lowest attainable cost is Ob We do the same
for each rate of output and, by joining the points, obtain the cune
illustrated
This curve is derived from the production function You should be
its meaning There arc several technically feasible
absolutely clear as to
ways of producing output 06, for the given factor prices there is one least
cost technique The cost per unit using that technique is Levels of cost
below Ob are not technically possible at that output levels above Ob (such

as OA ), while possible, arc not economically efficient The same argument

can be applied to points on the curve above A, C, D etc The curve is the
boundary between what is feasible and what is not feasible given the pro
duction function and the costs of the factors of production
If at any moment in time, the firm must change its outpul from 0-1 to
cost
OB, It will not find its costs changing from Oa to Ob The long-run
curve shows costs when all factor inputs have been fully adjusted To
move
and
from one point on this curve to another the firm must change its plant
the
equipment m
order to achieve the lowest pt»sible cost of production at
new level of output This is not something that can quickly be accomplished
cosu of pro
1* From the table of note I page 261 you can calculate » number of
different

ducing 6 000 tons of that good a rnonth when both (acton cost ^250 per unit
We have alrea y
circumstances an
noted that IOj per unit is (he lowest average cost that is atumable in these
that It IS attainable with only one factor combinauon and 6^
 —

THE VARIATION OF COST WITH OUTPUT 269

The Shape of the Long-Run Cost Curve

The long-run cost curve in Figure 20.1 had to be given some shape in order
to draw it. Figure 20.2 shows three stylised long-run cost curves for three
different firms. Firm A is what is called a falling cost firm. An expan-
sion in production will, once sufficient time has elapsed for all desired
adjustments to be made in the
techniques of production, result in a reduction
in costs per unit of output. Since money costs of factors are' assumed
constant, the in costs per unit must be because output increases faster
fall

than inputs as the scale of the firm’s production expands. Such a firm is
often said to enjoy increasing returns to sc.ale. Firm B is what is
called a rising-cost fir.m. An expansion in production will, even after

c
3
O
CL
-
to
o
t)

Average cost

Output per period

Fig 20.2
(i) (Firm A) ; Falling costs; increasing returns.
(ii) (Firm B) : Rising costs; decreasing returns,
(iii) (Firm C); Constant costs; constant returns.

sufficient time has elapsed for all adjustments to be made, be accompanied

by a rise in average costs per unit of output. Since costs per unit of inputs
are assumed to be constant, this rise in costs must be the result of an
expansion in output less than in proportion to the expansion in inputs.
Such a firm is often said to suffer decreasing returns to scale. Firm
C is a constant-cost firm. Its average costs per unit of output do not
change as the scale of output changes. This means that output must be
increasing exactly as fast as inputs are increased; the firm is said to be
encountering constant returns to scale.*

I* The production function of our example displays constant return

to scale. have We
already seen in note page 261, that average costs are constant when both inputs can be
1,

varied. The same phenomenon can be observed by' noting that doubling both inputs exactly
doubles the output: ,—
270 THE INTERMEDIATE THEORY OF SUPPLY
It IS common in the literature of economics to refer to these three siiua
tions, increasing constant or decreasing long run returns, as the lais oj
returns One will
often find writers speaking of a particular industry obeying
the Jaw of increasing long-run returns Such a usage is misleading and
is

based on a misunderstanding of the concept of an hypothesis iVhat we

have done is to label all possible cases Such a division of the possibiliues
into three groups may or may not prove useful, but such a classificatory
scheme covering all conceivable cases cannot be regarded as a theory
(much less a ‘law )

An investigation of the possibility of obtaining hypotheses about the

behaviour of long run costs is a dilTicuU task and it is relegated to the
appendix to this chapter In the meantime we can summarise the discussion
b> saying that a theory that predicts the existence of falling, long run costs

under certain circumstances and constant long run costs under other
circumstances is very easily developed Furthermore, considerable cmpmcal
evidence of the existence of both of these phenomena (firms A and C in
Figure 20 2) has been provided b> numerous empincal studies of the be
haviour of costs On the other hand a satisfactory theory that predicts nsing
long-run costs (firm B m
Figure 20 2) is very difficult to develop Further
more, the empincal evidence for the existence of this phenomenon ts rather
shaky and the whole existence of nsmg long run costs within the range of
'
our present observations is thus open to question

SHORT-RUN COST BEHAVIOUR FIXED FACTOR PRICES

Assume, first, that the firm has selected some size of plant and equipment
and IS producing at the appropnatc point on its long run cost curve Say
of
that the firm in Figure 20 I is producing output OB at an average cost
0l> per umt Now assume that the firm wishes to change its rate
of output

from OB before sufficient lime elapses to change its plant and equipment

Now multiplying both inputs by the same constant A gives

C = =» y/iVJr = ^y/7i^
production function with falling cos« »
which IS A ames the original output An nample of »
e -/i/i

and one with rising costs is

1 An excellent review of the empincsJ evidence on the behaviour of costs

Analjisu McCraw Hill e
evidence can be found in J Johnston, Slalu/ieal Cast
{
much new
York 1960)
 THE VARIATION OF COST WITH OUTPUT 271

The firm now

in a short-run situation with plant and equipment as the
is

fixed factors, and labour, management, materials, etc., as the variable

factors. What will happen to costs as output is varied above or below OB
while the size of plant remains unchanged at the size most appropriate to
producing OB? The first thing we can say is that the firm can never achieve
a lower level of cost than that shown by the long-run cost cun^e. On the
other hand, it may well do worse. The output OA can, for example, be
produced at cost Oa when the size of plant and organisation of producdon
is adjusted to that output. If OA is produced
irith a plant adapted to output
OB, costs may be higher, say Oa'. Similarly, if output OC is produced not
by a plant most appropriate to that level, but by ovenvorking a plant
designed to produce at the rate OB, then costs may be Oc' rather than Oc.^

1* In our example our production function is Q=y//ifz- When production was 6,000 tons
per month and fi and both cost £200 per unit, the least-cost was achieved by using 6/i and
S/'j making total cost £2,400. If output must be raised to 12,000 tons with /, 6xed, then 24 units
of/j would have to be used [12 = ,y(24) (6)] making total costs £6,000 or 1 Or a ton. When/,
can be varied as well tbe least cost will be achieved by using 12/, and 12/2 [12 = .y(12)(12)]
making a total cost of £4,800 or 8s a ton.

The short-run average-cost curve when /, is 6xed at 6 and/j is the variable factor both costing

£250 a unit can be derived as follows:

TC = 1,500-1-250/2 (>)

Q = V¥2
= ¥2
(
2)

substituting (2) into (1) gives

3)
TC = 1,5004-4119^ (

The short run average cost function is

(
4)

show that short-run total costs are not a linear function

Plotting a few points for (3) and (4) will
 272 THE INTERMEDIATE THEORY OF SUPPLY
We nowintroduce an empirical hypothesis about costs that
is sufficient
to guarantee the above results We assume that for each
level of production
there some best proportion m which to use the various factors
is
This pro-
portion which IS often referred to as the optimal factor
combivatiom
IS best in the sense that it a lower cost per unit of product than
results in
does any other factor proportion We abo assume that the further
assay the
actual factor proportion is from the proportion optimal for that
level of
output the higher the costs per unit The optimal factor proportion need
not be the same for each level of output Indeed when output is small the
optimal factor combination may be a lot of labour and not very much
capital when output is high the optimal combination may be a lot of
capital with relatively little labour
All that IS necessary is that for each level of output there should be an
optimal factor combination and that as input proportions depart from this
combination (for a given level of output) average costs per unit should
increase
immediately that each point on the long run cost curve must
It follows

represent an optimal combination of factors If it did not then the same

output could be produced more cheaply by using the optimal combination
and thus the original point could not have been on the long run least cost
curve It also follows that short run costs will be above the costs indicated
by the long run curve except at the point representing the output for which
*
the fixed factor was designed
Figure 20 S illustrates a firm wvth plant and equipment designed for out
put OB What happens to costs when output is varied in the short run ’ As
output is varied from OB the factor combination must vary Since some
factors arc held constant the variation of output brought about by varying
some inputs necessarily causes factor proportions to vary As output departs
from OB factor proportions depart from the optimal level Thus short run
costs will rise above the long run costs as shown in the figure All this means

IS that the firm sper unit of output will be higher if it is forced to

costs
produce some output with a plant designed for a different level of output
than they are if the firm can produce the output with a plant optimally
curve is
designed for that level of output Note that the short run cost
of
of output and short run average costs arc not constant indeed they fall up to an output
is expanded
6 000 tons where the optimal factor comb naoonisacheved after -which as output
further average costs nse
oos
1 If output IS to be increased no more of the fixed factor is available so factor propoft
constant facwr
must be vaned If output is to be decreased it would be possible to maintain
output by
proporuons by not using all of available supply of the fixed factor thus reducing
inputs by 10 per cent keep ng factor proportions constant
say 10 per cent by reducing all
fxed factor is
For redueuons in output we thus have to make an added assumpt on that the

indivisible as an integrated factory tn ghtbe *o that all of it must always be

used
 THE VARIATION OF COST WITH OUTPUT 273

tangent to (touches) the long-run cur\'e at the level of output for which the
fixed factor is appropriate and lies above it for all other levels of output.
If we now repeat the whole process, assuming that the firm is fully ad-
justed to some different level of output and that it then has to vary output

Fig 20.3 The relation between the

long-run and one short-run average
cost curve.

° Output per period

in the short run, we shall generate a new short-run cost curve. If we repeat
this process for ever}' level of output, we shall find that ever}' point on the
long-run cost curve has associated with it a short-run cost curve that touches
it at that point and that shows how costs vary as output is varied, holding
some factors fixed at the level most appropriate to that output. This is illus-
trated in Figure 20.4.^

Fig 20.4 The relation between the long-run and a series of short-run average
cost curves.

1* Each short-run curve touches the long-run curve at one point and lies above it every-
else. We say the short-run curves are tangent to the long-run one. This has an
where important,

though subtle, consequence. Two curves that are tangent at a point have the same slope at
it is tangent to SRAC, then SRAC must
LRAC also be de-
that point. If is decreasing where
creasing. Thus the point where SRAC— LRAC need not be the minimum point oiSRAC. Unless
LRAC is horizontal at the point of tangency, it will not be tangent at the minimum point of

SRAC. The economic sense of this rests upon the subtle distinction between the most efficient
way to utilise a given plant, and the most efficient way to produce the amount of output
required. It is the second that interests us as economists. If bigger plants can achieve
lower costs
from
per unit, we \vill gain by building a bigger plant and underutilising it whenever the gains
 274 THE INTERMEDIATE THEORY OF SUPPLY
The Shape of the Short-Run Cost Curve
The short-run cost curves that we have derived are U shaped or saucer
shaped This shape is a consequence of the hypothesis
that we introductd
on page 272 This hypothesis n often called the law
of variable proportions,
although, since U is an empirically testable assertion
about the world it
would be better described as the hypothesu of variable proportions
Short run
variations in output are necessanly brought about
by varying factor pro
portions This hypothesis states that, for any level of output,
an opumal
factor proportion exists, and, the mote this proportion is
departed from
the higher average costs will be per unit ‘

VARIATIONS IN FACTOR PRICES

A nse m any factor will shift all the cost curves both short
the price of
and long run - upward If, other things being equal, one factor gets more
expensive, then costs of all products that use anything at all of this factor
must nse
A rise in the price of one factor will also alter the optimal or least cost

factor proportions that underlie the long run curve Ifone factors gets more

using the bigger plant arc big enough lo offset the costs of being inefficient m our use of the
plant If there are rome gams from building bigger plants (if li^dCis declining) there u always
tome underutilisation that tsjusuded The student who is doubtful should ask himself whether
he will leam more from a book whose every word he undmunds or the book thai leachts h «i
the most They may be different The student who u suIJ troubled may take tome comfort
from the fact that less than 35 years ago one of the world s leading economic theonsU Jacob

Viner instructed hil draughuman to draw the LRAC through the minimum points of the
curves but so as never to he above the curves He reported of his draughtsman
He is a mathematician however not an economwi and he saw some mathematical objection
to this procedure which I could not succeed in understanding I could not penuade him to
disregard his Kruples as a craRsman and to follow tny rnsttuciions absurd though they might
be At the time Viner had no idea that it was his instructions and not the draughtsman s
scruples that were absurd
1 The hypothesis li often stated as liic hypothesis ^
imuashing tttumj This is exactly the same
relation looked at jn another way The hypothesis of diminishing returns states that if more and
total
more of a variable factor used in combination with a fixed {actor the amounts added to
is
The
production by successive increases in the variable factor will eventually begin to decline
and more with a given amount
reason for this u that as the firm uses more ofa vanable factor
means
of a fixed factor it increasingly departs from the optimal combination of factors This
output or
that equal additions of the vanable factor bring smaller and smaller increments of
what the same thing the cost of produang each additional unit of output must
be rising
15
diminishing
The U shaped short run cost curve can be denved either way A denvation from
returns was given in the first edition and I have gone over to die variable
proportions argument

because it seems to me to be both simple and mtuttivdy more appealing

 THE VARIATION OF COST WITH OUTPUT 275

expensive, then this may

be partially compensated for by using somewhat
less of this factor and more of others that can be substituted for it. How much
substitution occurs will depend on how much the price has risen and on
the extent to which it is technically possible to substitute other factors. In
general, however, we can say that a rise in the price of one factor will alter
the opdmal factor combination for each level of output and thus cause each
point on the long-run curve to represent a lower proportion of this factor
and a higher proportion of its substitutes.

AVERAGE, TOTAL AND MARGINAL COSTS

So far we have confined

ourselves to average costs when we talked about the
costs ofproducing various levels of output. We can, however, look at costs
in three distinct but related ways. We can look at ( 1 ) the total cost of pro-
ducing some level of output, (2) the average cost, which is total cost divided
by the number of units produced or (3) the marginal cost, which is the
difference between the total cost of producing some rate of output and the
total cost of producing at a rate of one less unit per period. We must now

consider these important concepts in detail.

1 T OTAL COST means just what it says the total cost of producing any
:

given level of output. This total cost is conveniently divided into two parts,
total fixed costs and total variable costs. Fixed costs are those costs that
do not vary with output; they will be the same if output is one unit or one
million units. The costs of any fixed factor will be a fixed cost: since the
factor cannot be varied the costs associated with it cannot be varied. Fixed
costs are also often referred to as overhead costs or unavoidable costs. Variable
COSTS are all those costs that vary directly with output, rising as more is
produced and falling as less is produced. In our previous example, since
labour was the variable factor of production, the wage bill would be a
variable cost. Variable costs are often referred to as direct costs.^

2 Average cost is the total cost of producing any given output divided
by the number of units produced so as to give the cost per unit. Average

1 In symbols we may write:

TC = TFC+ TVC
TFC= k
TVC = w,Xi,
where 7U is total cost, TFC is total fixed cost, TFC is total variable cost, A" is some constant
amount, X, is the quantity of the variable factor used and w, is the price per unit of this factor.

Since fixed costs are constant and variable costs necessarily rise as output rises, total costs must
rise with output or, to put the point more formally, TC is a function of total product and it

varies directly with it: TC=i(Q).

 276 THE INTERMEDIATE THEORY OF SUPPLY
totil cost may be divided into average fixed costs and average
variable costs
m just the same way as total costs were divided ‘

We may note that while average vanabU costs may nse or fall as pro
duction IS increased (depending upon whether output nses more rapidly or
more slowly than total variable costs) it is clear that average fixed costs
decline continuously as output increases A doubling of output always leads
to a halving of fixed costs per unit of output This is a process popularly
knotvn as spreading one s overheads
3 Marginal (or incremental) cost is the increase in total cost
resulting from raising the rate of production by one unit ^ Since fixed costs
do not vary with output marginal fixed costs are always zero Therefore
marginal costs are necessarily marginal variable costs and a change in
fixed costs vvill leave marginal costs unaffected As a simple example note

that the cost of producing a few more potatoes by farming a given amount
of land more intensively is the same whatever the rent paid for the fixed

amount of land ^
These three measures of cost are merely different ways of looking at a
single phenomenon, and they arc mathematically interrelated * Sometimes
It is convenient to use one and sometimes another
I In symbols
TV nV+TFC
ATC «

rtc
A(C «

TFC X

2 In symbols
MC, = Tv.~rc. ,

the toul oat of producing n unila minui the

1 c the marginal cost of the nth unit of output is

total cost of producing it— t (t e one le») units of output If we are producing a number o
to eac
identicaluniO of output we cannot of course ast-nbe a separate (and different) cost
unit When we speak therefore of the marginal cost of the nth unit we
mean nothing more
than the change in total costs when the rate of production is increased from
«— 1 units tonuu ts
per period of time
3 Only ch<rstmp}<nC Iwt W> prow this imporunt theorem
MC. = TC,-TC, , = (TH:.+7H;.)-(TVC. r + TTC. )

but TTC- - ^
therefore A/C, = (TlC.+Ay-fTTC. +JC) * TfC.- rn?. ,

Hence marginal costs are independent the sire of fixed costs

* IC ' 9 J-rc
^
MC - where J means the change in
 :

THE VARIATION OF COST WITH OUTPUT Til

The RELATION OF MARGINAL AND AVERAGE COSTS; Although we have

studied only average costs in this chapter, we have noted that it is possible
to move from one concept of cost to another. Figure 20.5 shows the relation
of averageand marginal costs for the short-run costs of a firm.
The between these two cun'es is a mathematical one and not a
relation
matter of economics. Although it is a purely formal relation, it is very im-
portant that you understand how these curves are related, since, when \ve
come to use our theory' of costs, we shall want to be able to use whichever of
average or marginal costs happens to be more convenient in the particular
circumstances. The average cost cuiv'e slopes downward as long as the

Fig 20.5 The relation between

average and marginal cost curves.

marginal cost curve is below it; it makes no difference whether the marginal
cost curve is itself sloping upward or downward.
When AC is declining, MC is below it. The common-sense interpretation
is of an additional unit of output lowers the average cost,
that, if the cost
it must itself be less than the average cost. When average cost is rising,

marginal cost is above it. This is because, if an additional unit is to raise

average costs, its cost must be more than average cost. It follows from these
two propositions that the marginal cost cuiA'e must cut the average cost

The term marginal cost should not be used unless the change is very, very small. For finite

changes, the term incremental cost is preferred. the changes are infinitesimal, the
Where
required to express the concept accurately. In the calculus we
differential calculus define
is

marginal cost as follows

dTC
MC =
read: ‘The derivative of total cost vdth respect to output’ and which means the
rate
which is
is familiar to the
at whichcost is changing as output changes. The concept of a derivative
student in such an everyday notion as speed : to say one drives at 80 miles per hour is to say
that the rate at which distance is covered (miles) with respect to dme (hours) is 80, or

d Distance
Speed =
d Time
278 THE INTERMEDIATE THEORY OF SUPPLY

cur\c at the lowest point on that curve This is a virtually self-evident

proposition, and a valuable one to remember ’

I * The propositions of thu paragraph arc easily proved using calculus

Let Csaverage cost, and Q=quariiity of output.

then AfC®
dQ
A
^dQ
—+C
dC/dQ ts the slope of the average cost curve Since quantity of output (Q) is positive,

0 means A/C <C (I)

d?

d<?
0 means StC >C m
dC
0 means t/C sC 13)
dQ
Condition (3) is, of coune, only a necessary condiuon of minimum C, but n u enough to prove
that marginal cost a equal to average cost at itt minimum point
APPENDIX TO CHAPTER 20

LONG-RUN VARIATIONS
IN OUTPUT^

Short-run variations in output neces- factor proportions may permit still

sarily involve changing factor proportions greater gains in efficiency.

{progressively more of the variable fac- In Chapter 20 the sources of variation
tors being applied to the fixed factor), of cost with output were characterised
but long-run cost variations can occur simply as ‘economies or diseconomies
^vith or without changes in factor pro- of scale’. In this appendix we shall use
portions. If 100 units of output can be a more specialised vocabulary. We shall
produced with 10 units of 2 and 150 X use the phrase returns to cost to describe
units of Xi (in a two-factor situation), the overall effect of all sources of long-
how might 200 units be produced ? One run variation of costs with output. Re-
\vay to produce them is to increase the turns to cost are divided into three
quantities of both factors, keeping pro- parts:
and follow the identical
portions the same, 1 Returns to scale, in which factors
techniques that were used to produce are kept in constant proportion but
100 units. This is called replication. The varied in level;
firm can build two identical 1 00-unit 2 Returns to substitution, in which
plants and operate them both. These factor proportions are varied at the
two plants would use 20 units of Xj and same time the quantities of all factors
300 units of A'l to produce 200 units of are varied; and
output. But the firm may be able to do 3 Pecuniary returns to cost, in which
better than this; perhaps a specially induced factor price variations are in-
designed 200-unit plant need not be cluded.
twice as big and not use twice as much The first two of these components
of factor Xi to achieve twice the out- are purely physical phenomena and are
put.^ Furthermore, perhaps a change in defined by the production function. In
discussing them, we assume that factor
prices are fixed and are independent of
1 This appendix is rather advanced and should
the quantity of the factor used by the
normally be omitted by beginners.
firm.
2 This would arise if the increase in size
allowed the use of more specialised (and hence
more productive) capital and labour, so that
RETURNS TO SCALE
doubling all inputs and reorganising the produc- Suppose we can produce a specified
tion process more than doubled the output. output Q* by using specified quantities
280 THE INTERMEDIATE THEORY OF SUPPLY
of a senes of inputs /Vf, Af, , AJ, the quantity of every factor used by
where the asteruk denotes the fact that ‘
the sm
multiple A
these are specific quantities In other This hypothesis is suggested by the
words, start from the specified value of definition of the long run it u anoiher
,
the production function way of saying that all facton that
affect output can be identified and
Q* = Q[XUXS, ,X*) (1)
varied by the firm In a world of idenii
Now consider multiplying each of the cal chickens (and identical chicken
inputs by a fixed amount, A In general, feed, etc ), if one hen can produce one
output can be expected to change and egg Hi one day, two hens can produce
we denote the proportion by which Q* two eggs m
one day, if they arc pro
does change by a Thus we may wnte vided with twice as much chicken feed

(and twice as much of every other fac

aQ* Q(XX*,XX^, ,XXZ), (2) tor) Although there have been at
and we list the three possibilities as tempts to show that the hypothesis ij
incon'cct, the evidence suggested in
follows
each case appears, on close examina
A a < tion to be very suspect \ common
{decreasing returns to scale)
kind of attempted refutation it of ihit

a A = sort A firm with ten identical plants

(constant returns to scale)
may produce less than ten times as
much output as a firm with one plant
because the complexities of manage
a > A
argument
ment increase This is a false
(increasing returns to scale)
which implies one of the follovving
Equation (1) says that the total output (1) that the
factor management is
originally produced depends on the fixed (2) that management of a given
exact quantities of the factors actually quality IS limited and the firm is sub
used Equation (2) descnbes a situa- stitutmg an inferior factor, or (3) that
management problems of the large
tion in which the initial quantities of the
the factors are all multiplied by the operation are more than ten limes as
same amount (e g they are ail doubled
,
complex (and thus require more than
small
or halved for A = 2 and A= 5, respec- ten Ufflcs the management) as the
tively), while the total product changes operation The first two merely imply
manage-
by some proportion, a In the next lines that the factor high quality
we attach a term to each of the logic ment IS fixed and so are not refuta

ally possible outcomes of this event tions the third implies that the firms
In order to go beyond a mere label- are nOt profit maximiscrs since they
ten identi-
ling of all possible cases it is necessary could alovays replicate by building
units
to offer hypotheses that say something cal independently functioning
replication it
about behaviour We start with two FroJii the hypothesis of
follows at once that there
can never be
technological hypotheses
Tke Hypothesis of Replication if all
inifgril mulople A
we
identifiable factors of production ate free to 1 When wc »"
»
be a whole number or ai it
be varied, it is aliiays possible to inceease nwan th« A w to

output by an integral multiple A by increasing more usiiaUy exiled an integer

 LONG-RUN variations IN OUTPUT 281

decreasing returns to scale,, for the hypo- very’low (as in the case of the whole
thesis says that replication always per- chicken) or quite substantial (as in the
mits at least constant returns to scale.' case of a blast furnace) matters a good
The hypothesis of replication does not deal and is often used to define the
rule out increasing returns to scale. Al- point of minimal efficient scale of a pro-
though it says that one can always cess of production. Below this point
double output by doubling all inputs, proportionate reductions in factor
it does not say that one cannot do better. quantities will lead tomore than pro-
Whether there can or cannot be in- portionate decreases in output.
creasing returns to scale depends upon The major source
of imperfect dmsi-
whether factors are perfectly divisible, bility has to do not with the physical
as the following hypothesis points out: characteristics of factors but with
specialisation in their use. Consider the
The Hypothesis of Perfect Divisibility . butchering of livestock. To turn one
If all factors are free to be varied, it is steer into marketable cuts of beef one
always possible to decrease output by any man with one knife must do about 100
multiple A (0<A<1) by decreasing the distinct tasks. One hundred men with
quantity of everyfactor used by the multiple X. 100 identical knives could butcher 100
steers by each butchering one steer en-
This
a testable hypothesis about
is tirely of course, the process
(this is,

the possibility of dividing input into of replication). Or, each of 100 men
smaller and smaller units and about the could become veiy' expert in one
effects of doing so. In some cases, the operation by specialising in this opera-
;

hypothesis is obviously wrong. Half a tion and performing it over and over
hen may very well produce no eggs be- again on each steer, the 100 men could
cause one hen is an indivisible quantity butcher 100 steers in less time than one
of hens. Whether a loom or a tractor or man could butcher one steer.' Where
a blast furnace or an assembly line is does indivisibility come in? One could
perfectly divisible is an empirical hire 100 specialists to butcher one steer
if one could hire human sert'ices by
matter, not a logical one. the
The hypothesis of perfect divisibility minutes for the same rates as by the day,
is frequently empirical
rejected on week, or year: this is, of course, im-
grounds. For most operations there is a possible. It is not possible to duplicate
minimum level of operation that is this efficient organisation of production
required to make certain factors effi- on a small scale because it requires a
ciently usable. Whether this level is rate of output sufficiently large that a
man can work all day at a very'
1 Since replication is limited to integral in- specialised part of the whole process.
creases in output, this statement is correct for If factors were perfectly divasible,
large enough increases in output, but may be in-
increasing returns to scale would be
correct for smaller increases. That one need
is,
impossible.^ Taken in conjunction with
never do worse than constant returns to scale if
one doubles output; but fora 50 percent increase
in output, replication
by itself does not provide a 1 Since output is always ‘p^t unit of time’,
technique of production. This is not an import- this is an increase in output.

ant qualification in the context in which we wish 2 Suppose the reverse. Suppose production at
to talk about returns, and we shall ignore it. a rate of 100 units per week appeared to use less
 )

282 THE INTERMEDIATE THEORY OF SUPPLY

the hypothesis of replication, perfect among processes is the source - as he
divisibility implies constant returns to shall see - of substantial returns to
scale over all possible ranges of output substitution
If we do not ha\e perfect divisibility If returns to scale were the onl)
of factors but we do have the possibihty source of variation ofcostwith long
run
of replication returns to scale may in \ariations of output we would expect
crease or be constant long run costs to exhibit a pattern such
When considering a single process of as that exhibited by the cunes m
production* it is typically found that, Figure 20 6 Curve xA defines the
owing to factor indivisibilities, there is \anauon of average cost with output of
some minimal efficient level of opera- some process of production as the
tion that IS required m order to achieve quantities of factors of production are

the efficient use of the process
some processes this level is very low, for
others it is more substantial (This
variation of minimal efficient scales
than twice as much of all inputs than were re
quwed produce al a rate of 50 umu per week
to
One available technique of producing 50 units
would be to use the technique used to produce
too units at half the level of use of each factor
Nothing excludes this possibility and thus the
For increased in constant proportions Out
put Oa IS the minimal efficient scale of
production of that process Curve zB
refers to a different process and has a
minimal efficient scale of production
Ob Each of these curves exhibits in
creasing itvoms sx'ak. Vq vW
of minimal efficient scale, and constant
returns to scale thereafter

initial situation could not have been a lechiio

logically efficient use of factors »f factors were RETURNS TO SUBSTITUTION
pcrfeclly divisible owing to
Long-run costs may vary
1 What we now label as a proetss of production a*
is the case that in practice corresponds to the
changes m
factor proportions as wcl
sec'i
fixed factor proportions required lo discuss re to changes in scale This is easily
scale
turns to scale in Figure 20 6, which pTCscntcd
 LONG-RUN VARIATIONS IN OUTPUT 283

cost curves for two different processes.* four hours to produce four shirts. This
If the firm anticipated production at method is more capital-using than the
levelsbelow Os (and was limited in its previous method in which each girl has
choices to thesetwo processes) it would ,
as capital one needle now one girl has
;

use process A. If output was to be at a a sewing machine. The manufacturer

rate in excess of Os, the firm would has thus changed factor proportions
prefer process B. Over the whole range (and technologies) he encounters in-
;

of output from zero to Ob the firm creasing returns to substitution. If he

would achieve lowest costs by first washes to raise output to five, six, seven,
using process A and then shifting from and eight shirts, he hires his one girl for
process A to process B at the point five, six, seven and eight hours without
where the curves cross. (Output Os.) buying a second machine. This in-
The cuiA'e xyB exhibits a pattern of sewing machine leads
divisibility of the
variation different from either process him to encounter increasing returns to
curve. This pattern can occur whenever cost in the four-to-eight shirts per day
different processes have different mini- range. If he wishes to raise output to
mal efficient scales owing to different 16, 24 or 32 shirts per day, he buys tw'o,
effective indivisibilities.^ three, or four sewing machines, each
The following example may help to with an operator working eight hours.
illustrate the interaction
of substitution He is now encountering constant re-
TOth replication and indivisibiities a : turns to scale by the process of replica-
man wishes to produce one shirt per tion. Once his output reaches 50 shirts
day. He hires one girl with a needle and per day, it pays him to buy a factory
thread to do the job. Now assume he w'ith machines for all the processes and
wishes to raise his output to two, and to utilise it at 60 per cent capacity. His

then to three, shirts per day. He hires
two, then three, each with
girls,
needles and each producing one shirt
per day. In other words, he replicates,
keeping factor proportions constant; he
encounters constant returns to scale.
If he wishes to produce four shirts per
day, he finds it worthwhile to buy a
sewing machine and hire one girl for
1 Each process by definition used constant-
factor proportions, but process A and process
must use factors in different proportions ;
if
did not, they would be the same processes and
their cost curves would coincide.
2 To return to the livestock example, our 100
men might be equipped with more specialised
equipment than 100 identical knives, in which
case a different process, use of an electric bone
saw, say, would be required. An electric bone
saw, however, might well require a larger level
of output to be efficiently used.
ratio of labour to capital has now
changed and he has encountered a fac-
tor-substitution effect. As his demand
expands, he encounters increasing re-
turns to cost as the factory approaches
100 per cent capacity. After this, in-
creases in output are obtained by
building duplicate factories so that he
isback in the range of constant returns.
Even the most casual observation of
the differences in production technique
B used in large- and small-sized plants
they supports the existence of the differences
in factor proportions that w'e are calling
differences in processes. The reason is
that large specialised machinery and
equipment, which uses a different rado
of labour to capital than does smaller,
less specialised equipment, is useful
only when the volume of output that
the firm can sell justifies its use. Con-
284 THE INTERMEDIATE THEORV OF SUPPLY
sider a fe%\ examples the use of the turns to substitution will
result (ac
assembly-ime technique, body-stamp- cording to the hypothesis ofreplication]
ing machinery and muittple-bonng when output exceeds the mmimal
engine block machines in car pro- efficient point of the lowest cost
process
duction would be profitable only if Decreasing returns to substitution are
individual operations are to be repeated not possible for a profit maximising
thousands of times This substitution of firm
mechanised equipment for labour or
for simpler tools represents factor sub
stitution The use of elaborate harvest- EMPIRICAL EVIDENCE ON RE
ing equipment (which combines many turns to scale and sub
individual tasks that might be done by STITUTION
hand and by tractor) is profitable on a Important empirical questions concern
big farm but not on a ftw acres The where and when increasing ittums
continuous strip rolling mill in sieel exist how large these returns are and
production provides a highly efficient over what ranges of output they occur
technique of producing flat steel but Many empirical studies have been
only if the volume of production is such made that attempt to measure the frac

as to utilise the capital equipment on a tjon of the total industry s market that
nearly full-time basis a single plant requires in order to
Typically the substitution involved achieve approximately minimum efli

IS of capital for labour and of complex cient size * By way of illustration v.e

machines for simpler ones Electronic
devices can handle a very lai^e volume
of operations very quickly but unless
the level of production requires very
large numbers of operations it does not
make sense to use these techniques
Increasing returns to substitution are
possible if individual processes have
points of minimal efficient scale at
' cite the findings of a study by Joe S
Bam made for industry in the United
States This study presents the follow
ing classification of a number of US
lodusiries according to the size of plant
^
required to achieve moderately eff cient
production
I Industries with very important
’ type
plant scale economies cars,

different levels of output ^ Constant re- ivnters

4 Except for in between levels of output See

1 Automation is a contcmporar> examplf of footnote page 271 The hypothesis of rtpl ca
1

(ion states that factor substitution never mtssur}

this kind of substitution \V« shall discuss Jt in is

Chapter 21 m the long run the firm can expand by repi ca

computer (ion Factor substitution is used if but
only if R
2 Even a moderate sized elecironic
vkhich costs m
excess of £l million can perform promises greater efficiency
plant
as many computations per hour a* 100 000 5 The concept of minima! optimal
ncs the
women with desk calculators The subsutuoon scale as «t is sometimes called comb
and returns to sub
of eleclionic for manual cowputatKHt may make effects of returns to scale

I fe insurance company
sense for a large but not stitution
(Harvard
for the local grocer or bookie
6 Bam /i-imm toSea Compel I or,

3 It can be proved that if every process ex Utuvctsity Press 1^56) Chapter 3

of mar
hibited perfect divisibility there would exist eon 7 Industries where at least 10 per cent
achieve efficient
ket capaaty required to
slant returns to substitution at well as to scale is

This IS left as an exercise for the student production

 LONG-RUN VARIATIONS IN OUTPUT 285

2 Industries wath moderately im- Such ‘external’ returns to cost have

portant economies:* ce-
plant-scale been sugge'sted as possible in either
ment, farm machiner>', rayon, steel. direction.
3 Industries with relatively un-
important plant-scale economies can- The Possibility of Increasing Pecuniary
ned fruits and vegetables, cigarettes, Returns. This possibility can arise when
flour, liquor, meat packing, petroleum one industry, call it industry A, uses the out-
refining, soap, shoes and tyres and put of another industry, call it B, as its own
tubes. input. If the scale of industry A expands, it

Further sources of increasing returns will demand more of Ws product and B will
may occur for the firm beyond the level expand as well. If Pi is subject to increasing
of the plant. For example, a large returns to scale, its costs will fall and the
enough firm may find it efficient to price of its product will fall as well. Thus
establish a nation-wide organisation for the price of A’s input will fall so that A is

the sale or service of its product. subject to increasing pecuniary returns.

Evidence on these matters is import-
ant because, as we shall see later, the
nature of costs in an industry' has very
important effects on such things as how
easy it is for new firms to enter the
industry, whether or not there is a
tendency toward monopolisation in the
industry, and how easy it is for the in-
dustry to adapt rapidly to changes in
demand.
PECUNIARY RETURNS TO
COST
As a firm or an industry' expands its
production, it requires increasing sup-
plies of the factors it purchases. It will
require increases in the production of
the supplying industries, which must
turn adapt to the changes in demand
they face. These changes may lead to
changes in the prices the supplying in-
dustries charge; or, from the point of
view of the purchasing firm, changes in
the costs of its factors.
At least superficially this is an appeal-
ing and plausible possibility. Consider,
for example, the great increase in
3
large-scale agricultural production,
which has vastly increased the demand
for farm machinery. If the production
of farm machinery is subject to in-
creasing returns, the cost of production
per machine will have decreased and
the price paid by farmers may well have
been reduced because of this.
Empirical evidence that this actually
occurs (as distinct from the possibility
that it might occur) is difficult to find.
The reason is that it requires long-run
reorganisations of production in supply-
ing industries, which take years to
occur. Moreover, over such long periods
in many other influences also are opera-
ting to change factor prices.^
The Hypothesis of Eventually Decreasing
Pecuniary Returns [Rising Supply Price).
Increases in the volume of purchases from
any industry will eventually cause scarcity of

1 Industries where 5 per cent to 10 per cent Note that it is only the induced changes in
of market capacity is required to achieve efficient factor prices that we include as pecuniary re-
production. turns to cost. Other changes in such prices,
2 Industries where less than 5 per cent of whether random or whether caused by factors
market capacity is required to achieve efficient other than growth in demand for the industry
production. under study, are not included.
286 THE INTERMEDIATE THEORY OF SUPPL'
those factors that are heavily used by the ally requiring increasing pecuniary re
supplying industry and thus atU lead ta turns It can never be disproved
increases in supply prices empirically for any case that does not
Whether or not this is an empirically show increasing supply pnees the pro
testable hypothesis is a matter of some ponents of the hypothesis will argue
debate among economists It is possible v/e are not yet at the point where
to find some cases that do confirm the decreasing returns occur
prediction of rising supply prices for
expanding industries For other in- This proposition is the only pro
dustries no such effect is visible, but position in the theory of long run
does not prove that nsing supply prices costs that permits increasing long
will not occur tvttilually If the hypo run costs as a function of increases
thesis IS interpreted to mean there is the ‘
in production
possibility of decreasing pecuniary re
turns It may be regarded as demon I Note that It reiK upon the argymert that,

strable, and actually demonstrated If white the f rm « free to vaiy tt* ose of &e»w
the hypothesis 13 interpreted as eventu soc ety II lim led in iti lupply ofiome fseton
CHAPTER 21

THE VERY LONG RUN

Fifty years ago roads and railways were built by gangs of tvorkers using
buckets and spades and horses pulling scrapers. Today, bulldozers, steam-
shovels, giant lorries and many other pieces of specialised equipment have
completely banished the work horse from construction sites and have almost
completely displaced the worker with his shovel. A modern fully automated
factor)' with a few white coated attendants taking readings and making

occasional adjustments bears little relation to the nois>, clanking factory 60

years ago with hordes of unskilled machine attendants pulling levers, turn-
ing screws and generally playing an important but ven' unskilled part in the
process.
A 1967 Austin car is a ver)- different product from 1907 Austin. Nylon
stockings are ver)' different from silk stockings. The fully automatic washing
machine is a long way from the best method of scrubbing board and soap
that was available 60 years ago.
Today’s managers and ivorkers are a very far cry from their counterparts
of 60 years ago. They have been at school much longer, so that even the
will
unskilled workers will be literateand numerate and their manager may have
have some smattering of knowledge of modern scientific methods of business
control; they will all be in better health, better clothed and better housed
than their counterparts in their grandfather’s generation.
In the \'er)' long run there are three sets of changes that tend to dominate
the production function.
First, the techniques available for producing existing products change.
Over the span of 70 years which encompasses a normal life time these

changes can often be dramatic. Second, totally new products become avail-
able. Seventy years ago it would have seemed inconceivable that many
families would be allocating their incomes between such products as wash-
ing machines, refrigerators, dish Tvashers, cars and holidays in Spain. Third,
the nature of inputs, particularly labour services, changes due to changes in
288 THE INTERMEDIATE THEORY OF SUPPLY
standards of such things as health and education Thus new
products are
invented and old ones are produced more efTcctuel) both because
of new
techniques and because the quaht) of a given input improves
In this chapter we shall concentrate on changes in the techniques
of
production

PRODUCTIVITY
There is no doubt whatever that over the last 20, 50 or 100 > ears, the stan
dard of living of the typical family has increased enormously Our great
grandfathers would have regarded our standard of living as something
quite out of their world And indeed it is This is true of most industrialised
countries in the world A substantial part of tins increase is due to the inven-
tion of new, improved vvajs of making existing products This causes an
increase in output per unit of input cmplo>ed and it is referred to as an
increase in productiv it\
The magnitude of increases in productivity deserves some attention The
apparently modest rate of increase of output per man hour of labour of 2 0
per cent per year leads to a doubling of output per man hour every 35
years Produciivuy mthe United Kingdom has increased at approximately
'

this rate over the last 100 vears

Since the Second World U'ar, productivity growth in the has been UK
closer to 3 per cent per year (vvhich doubles output every 23i years) while
in other countries u has been much higher In japan, for example, it has
been increasing at over 7 per cent per year, a rate which doubles output per
man hour approximately every ten years

Productivity Changes and the Theory of Supply

There no question that productivity changes affect the supplies of com-

is
changes
modities What IS m doubt is the extent to which productivity
themselves are a response to economic signah and incentives and the
extent

creative activity o
to which they just happen because of the spontaneous
scientists, inventors and other researchers
'OsingteiAnncaVniTi^’Ci'i^'’'^'^
endogenous or
say we are in doubt whether productivity changes are
exogenous ^ .

which wi grea
This IS a very important question, the answer to
y

adjust to various
influence our judgment of the ability of the economy to
impossible, but it n
1 Measurement of output in a world of changing produces is not out
techmqi les for measuring
complex Economic statisucians have developed some ingenious
put changes but we do not discus them m
this volurue

2 See page 35 for a definition erf these terms

 THE VERY LONG RUN 289

disturbances that impingeon it. We shall introduce the problem in this

chapter and return to Chapter 56 when we discuss in more detail the
it in
problems caused by economic growth.

Sources of Increases in Productivity

Long-run increases in efficiency arc due to, and can be divided into, scale
effects, increases in the quality of the inputs, changes in the known tech-
niques of production and improvements in product. Mere population
growth, other things equal, will permit higher productivity if most products
are subject to increasing returns to scale. Substitution of more and more
capital for labour as the level of production expands is likely to lead to
greater productivity. Better raw materials, better trained or educated
labour, or better machines will increase productivity even if no changes in
factor quantities or proportions take place. Better organisation of produc-
tion can alone account for increases in productivity. New ideas can raise
efficiency by being applied to new products: imagination can design a better
mousetrap, with no change in the quantity, quality or proportions of
factors.

INVENTIONS AND INNOVATIONS

An INVENTION is defined as the discovery of something new. such as a new
production technique or a new product. An innovation is defined as the
introduction of an invention into use.
It is generally accepted that innovation responds to economic incen-
tives. If we are in the depths of a depression with household income barely
sufficient tobuy the essentials of life, firms are unlikely to introduce a new
luxur)' household product even though the product has been fully perfected
by its inventor. New products and methods will not be introduced unless it
appears profitable to do so, and a change in economic incentives can change
the apparent profitibilities of various possible innovations.
But innovation can only occur w'hen there has already been an invention.
If there is a dramatic rise in labour costs, firms may now decide to take up
some labour-saving process that hitherto had been ignored since its inven-
tion, but they cannot do so if the invention has not yet been made. If we are

concerned with the response of the economy to such economic sig^nals as

changes in the relative prices both of consumers’ goods and of inputs, then
we need to know the extent to which invention responds to such incentives.
A number of hypotheses about the sources ofinv'ention have been put for-
ward and we shall list the most important of these.
1 Invention is a random process: Some men are by nature both
curious and clever. Thousands of attempts will be made to invent better
10
290 TUB INTERMFDIATE THEORY OF SUPPLY

v\ays of doing things Many men will fail, and they remain nameless, but a
fc%v succeed These arc the Hardwicks, Stepliensons, SS'hillcys, Edisons and
Fords The successful inventions become a pool of potential innovations,
and, when the climate is right, tlic) are introduced into production

2 INVFNTIONISARTSPONSETOTHEINSTITUTIOVALTRAMEWORK Things
like the patent laws, tlie ta^ structure, or the organisation of business enter
pnse stimulate or rctird the process of invention Imention, in this view,
is not exogenous to socielv, bm U does not respond primarily to economic

variables and it certainly is exogenous to the individual firm

3 Invention anO issovATtos are the product of the inherent

I one AND MOMfNTini OF sciFNCE Scicncc has a logic and momentum
of Its own There was a time for llie discovery of the steam engine, the aero
plane hybrid corn, the electron tube, and the rocket Particular men arc

the msiruinents, not the causes of scientific discovery Had Edison nc'cr
been born we would still have had at aliout the same time, both the hgl'^

bulb and the gramophone \ccording to tins view, the present is the elec-

tronic age and automation ts the industrial form of introduction ofelectronic

devices into production

4 Nlcessitv is Tttr MoriitR of invention Ignorance is only skm deep

With enough expenditure of funds men can do anv thing - split the atonit
conquer cancer fly to Mars or cultivate the desert Tlje pace and rate of
iiueniion depends upon how many resources are devoted to solving prob-
lems, and resources are devoted according to needs A firm for which a cer-
IS becoming scarce wiU discover ways to economise on using
>t,
tain factor
or will develop a more plentiful substitute for it (For example, the scarcity
of high-grade iron ore led lo the development of ways of using low-grade
ores) In this viev' automation is a response to expensive labour The
impetus for invention may thus come from within the firm But it may also
come from without governments may set pnonties and do the research
that leads to major discoveries and innovations Atomic energy, for example,
vs the result of the United Kingdom s and later of the United States’ desire

to develop superw capons dunng the Second IVorld War The space And

missile programmes are responsible for some current discoveries that have
industrial applications as well

5 Profits are the spur The profit motive not only leads men and firms
to sei 2 e the best known methods, but to develop new ways to meet both old
and new needs and vvants But profit opportunities are connected to
vitally
the economic climate and to economic circumstances
 THE VERY LONG RUN 291

Which of these hypotheses is correct? Very possibly all of them. They are
not necessarily in conflict with each other so that the evidence for one cannot
be regarded as disproving the others. The fact that some firms spend millions
of pounds on research and development to overcome specific problems or
invent new products does not change the fact that many discoveries are
made in University laboratories by men who have little knowledge of, or
interest in, current prices and scarcities. The fact that many patentable dis-
coveries are given to the world does not prove that others are not motivated
by the prospect of huge personal gain. And so on.
We have asked whether the direction of invention responds to economic
incentives. We are also critically interested in the pure volume of invention.
Even if its direction does not exactly respond to current market signals it is
still a potent source of increases in living standards. Inventions that reduce
the quantities of all inputs required to gain a given output are absolutely

efficient and will raise living standards even if they save just as much on
plentiful as on scarce factors of production.
Variations in efficiency that are possible by choice among known tech-
niques and among alternative levels of output are limited in scope. We can,
of course, never do better than a 100 per cent utilisation of what is currently
known. Improvements by invention are potentially limitless. For this reason,
the long-run struggle of men to get more from the limited resources of the
world is critically linked to discover}'.

PAST AND FUTURE PRODUCTIVITY GROWTH

Economics used to be known as the dismal science, but this designation has
long since ceased to be applied. Why? Economics was thought dismal be-
cause its predictions were dismal. The basic prediction about the very long
run put forward by classical economists was that the population of the
world would continue to expand and that the pressure of more and more
persons on the limited resources of the world would cause a decline in out-
put per head.^ The history of the world would be one of more and more
people living less and less well, with the surplus of persons that could not be
supported at all dying off by plague, famine and pestilence.

This prediction has proved dramatically correct in some of the countries

1 This was due to the operation of the ‘law of diminishing returns’ or ‘variable proportions’.
The supply of land is fixed and as more and more, people crowded on the fixed supply the

optimum factor proportions would be departed from more and more and successive equal

additions to population would bring about smaller and smaller increases in total output. Also
with the stock of knowledge constant the most productive ideas would be totally exploited and
successive increases to the capital stock would be forced to exploit successively less productive
ideas and technical possibilities.
292 THE INTERMEDIATE THEORY OF SUPPLY

now called underde\ eloped It has, hcwvever, been falsified in most of the
countnes of Europe, temperate North Amenca, Australia, New Zealand
and some other industrial countries such as Japan Why’ On the one hand,
because of voluntary restnction on population growth due to the wide
acceptance of birth control, the birth rate was lowered so that m spite of
dramatic reductions m
the death rate due to advances in medicine and
h> gicne, the population did not expand as rapidly as was foreseen by econo-
mists writing before the wide availability of birth-control techniques ‘
Second, pure knowledge and applied techniques based on it have expanded
so rapidly during the last 150 years that man’s ability to squeeze more out
of his limited resources has expanded faster than the population has ex-
panded The race is a close one and should the rate at which the fronuers of
knowledge are being pushed back fall below the rate of population increase,
economics may once again become the dismal science In the meantime,
output per person and standards of living have risen greatly in the last cen-
tury and economics has had the task of explaining why things have got
better and better rather than of explaining why things have got worse and
worse
The lot of those underdeveloped economies for which the dismal predic-
tions of the classicaleconomics proved correct is m
many ways hard to
understand If the key to the performance of Europe and Amenca is inven-
tion and innovation, why could not these inventions be copied and installed
in underdeveloped countries, thus raising their living standards’ Evident!),
more than invention is required for innovation What we have to do is to

explain the slow pace of innovation given the vast stock of accumulated
inventions We shall examine these problems m Chapter 57

SUMMARY OF THE BASIC THEORY OF COSTS

This concludes our study of the theory of costs A number of the subjects
covered, although important, arc digressions from the point of view of
developing a theory of the behaviour of the firm We shall now summarise
•iVit ‘cnisit putifcj xiuAein ncehs to 'Know 'orf ore pimteiffiig 'r> 'he-
"ftre

next few chapters on the theory of supply Anyone who is not clear about
the following points should review his understanding of them b) re-rcadmg
the relevant parts of the preceding five chapters

1 We h>pothesize that firms strive to maximise their profits, and, since

profits are the difference between the costs and revenues of the firm, it i*

I Thu meant as a positive ttateiDent of fact and not as a normative siaiement about
IS

v-hether this development was good or bad id an ethical sense

 THE VERY LONG RUN 293

necessary to consider what we mean by costs and how these var)-^ as output
varies.
2 The cost of any input used by the firm is measured in terms of the
benefit foregone by not using that input in its next best alternative use.
.This an opportunit\'-cost concept of the firm’s costs.
is

3 In the case of any factor service that the firm hires from outside, the
opportunity cost of the firm is adequately measured by the price paid, since
thatsum of money cannot be used for other purposes.
4 In the case of factors owned by the firm, the opportunity cost is
measured by the current market value that these services could command.
The current market value may bear little relation to the cost originally
paid to purchase these factors.
5 What its profits is the difference between its revenues
the firm calls
and its measured by its accountants. The economist calls profits
costs as
the difference between revenue and all the opportunity costs of the factors
used. Generally the accountant does not include a return to capital and
to risk-taking in his concept of costs. The economist does. When the econo-
mist speaks of costs he includes the return that must be made to induce
capital to remain in a particular industry (i.e., the opportunity cost of
capital). Thus profits as the economist uses the term are a return over and
above the full opportunity cost of all factors of production used by the
firm.
6 Since costs measure the value of the factors in other uses, the existence
of profits in a particular industry is a signal that that industry is particularly
profitable and this signal will serve to attract firms to switch into this
industry from other lines of production.
7 So far we have seen how to value inputs. To see how costs vary' as
production varies we need to see how inputs vary as production varies. This
relation is described by the production function. When we have costed the
inputs we can derive a cost function showing how costs vary as production
varies.
8 The long-run, average-cost curve shows the lowest attainable average
cost for each level of output on the assumption that the proportion in
which inputs are varied is low as is possible.
so adjusted as to get costs as
9 The short-run, average-cost curve will beU-shaped if there is some

best (i.e. least-cost) ratio in which the factors should be used and if a given
amount of at least one factor must always be used in the short run (i.e. the
inputs of one factor can neither be increased nor decreased). In the short
run, production can be varied by varying the inputs of some factors, keeping
inputs of others constant. Thus, in the short run, production is varied by
varying the proportion in which factors are used. If we start from the
optimum factor proportion then the more is production increased or
294 THE INTERMEDIATE THEORY OF SUPPLY

decreased above or below that level, the more will the actual proportions
change from the optimum one, and the higher will the costs per unit of
output become
10 In the very long run technological knowledge changes and the
question of how costs change as output changes depends on the question
of how changes in technological knowledge respond to changes in economic
incentives
CHAPTER 22

THE THEORY OF
PERFECT COMPETITION

In Chapter 17 we hypothesized that firms seek to maximise their profits,

such profits being the difference between total revenues derived from selling
its product and the total costs of making it. In Chapters 19 and 20 we de-

veloped a theory of the nature of costs and their variation with output. This
theory is meant to apply to all firms. We must now consider how revenue
varies with output. We cannot develop a single theory' that applies to all
firms and it is the different reactions of the firm’s revenues to its sales that
distinguishes the market forms of perfect competition, monopolistic compe-
tition, oligopoly and monopoly. Having developed a theory of how revenue
varies with output we then study how firms behave in order to maximise
their profits. This leads us to a theory of the behaviour of firms in each of
the four market forms listed above. In this chapter we consider perfect
competition.

BASIC ASSUMPTIONS

The theory' of perfect competition is built on two main assumptions, one

about the relation between the sales of a firm and its revenues, and one
about the nature of the industry in which the firm operates.
1 The firm is assumed to be able to alter its rate of production and sales
within any feasible range without this having a significant effect on the
price of the product it sells. The firm can double or treble its sales or stop
producing altogether Avithout affecting price. Thus the firm must passively
accept whatever price happens to be ruling on the market. We say that the
firm is a price-taker.
2 The industry is assumed to be one in which any new firm is free to
296 THE INTERMEDIATE THEORY OF SUPPLY

set Up production if it so wishes and in which any existing firm is free to

cease production and leave the industry if it so wishes This means that exist
ing firms cannot bar the entry of new firms and that there are no legal pro
hibitions on entry or exit We say that the industry displays freedom of
entry and exit

THE RELATION BETWEEN REVENUE AND SALES THE

FIRM S DEMAND CURVE
We may begin by considering as illustrative cases the demands for the pro
ducts of a car manufacturer and of a wheat farmer The car manufacturer
will be aware of the fact that his own
policies can influence the market for
his product He will know he makes a substantial increase in his
that if

price his sales will fall off He will also know that if he lowers his price
substantially he will be able to sell more of his product If he contemplates
making a large increase in hi$ production which is not in response to some
known or anticipated rise in demand (i e a rightward shift m the demand
curve for his product) then he is likely to have to reduce his price in order

to sell the extra output Thus he cannot control both pnee and quantity in
any way that he wishes the quantity that he is able to sell will depend on
the price that he charges Let us now make the point more formally the
manufacturer of care is faced with a downward sloping demand curve for
his product He is able to determine either the pnee of his product or the

quantity that he sells he may fix his price m which case the maximum
quantity that he can sell is determined or he may fix the quantity that he
wishes to sell m
which case the maximum price that he may charge is
determined
Now way of contrast an individual producer of wheat He
consider by
will be one of a very large number of farmers all growing wheat his own
contribution to the total production of wheat will be but a very small drop
in an extremely large bucket The pre war elasticity of demand for wheat
was estimated at approximately 25 * Ifwe take this figure as being a reason
able estimate of the present elasticity this means that if the farmer did
succeed m increasing the world supply of wheat by 1 per cent he would
bring down by 4 per cent The farmer is howeter not
the world price
between changes in u,orld output and the price of
interested in the relation
wheat but rather in the relation between changes in his oiin output and the
price of wheat Even \ery large percentage changes in hij own output will

1 W Ifred Malenbaum Tht Wnld Wheat Fememj 1885 1938 (Harvard Univen ty Prea
1953)
 THE THEORY OF PERFECT COMPETITION 297

represent only ver)' small percentage changes in world output. The elasticity
of the market demand for wheat is defined as

_ percentage change in total quantity of wheat

percentage change in market price of wheat

where stands for the elasticity of market demand for wheat.

Let us now introduce a new concept, the firm's elasticity of demanfi which
we define as

percentage change in firm’s production of wheat

yip
= ——
percentage change in market price of wheat
Although the denominators of these two elasticities are the same, the
numerators are not. Any change in one farmer’s wheat production will rep-
resent a much larger percentage change in his own production than it will
in world production. Thus, in the case of the wheat farmer, Tjp will be sub-
stantially larger than rj^f.

To see this, let us calculate the firm’s elasticity of demand for a very large
wheat farmer. We have noted that the elasticity of demand for wheat
has been estimated to be about -25. Total world production of wheat was
approximately 200 million tons in 1964, while the output of what would be
regarded as a very large Canadian wheat farm would have been in the
order of 50,000 tons.
Now suppose that such a wheat farm increased its production by 20,000
tons, a40 per cent increase in its own production but an increase of only
of 1 per cent in world production. Table 22.1 shows that this increase
would lead to a decrease in the world price of of 1 per cent (lOrf in j^lOO),
and give the firm an elasticity of demand of 1,000! This is a very high
elasticity indeed. The farmer would have output 1,000 per
to increase his

cent to bring about a 1 per cent decrease in the price of wheat. Since the
farmer is quite unable to var)’ his output this much, it is not surprising that
he regards the price of wheat as being unaffected by any changes in output
that he could conceivably make. Clearly, we will not be far from the truth
is unable to
if we say that, for all practical purposes, one wheat farmer

affect the world price.

We therefore assume, with only a very slight simplification of reality, that
the farmer is unable to influence the world price of wheat and is able
totally
to sell all that he could conceivably produce at the going world price.

The individual wheat farmer is thus assuhaed to be

faced with a perfectly elastic demand curve for his
product. He is a price-taker.
that this term has
1 In the first edition I called this own elasticity of demand. I find, however,
nomenclature.
already been used for a related but not identical concept, hence the switch in

10 *
29B THE INTERMEDIATE THEORY OF SUPPLY
Table 22 1

CALCULATION OF A FIRM’S ELASTICITY OF DEMAND

Gi\eu i?\, =0 25
World output = 200 million Ions
Firm s output increases from 50,000 to 70,000 tons

Step 1 Find the percentage change in world price

percentage change in world output

percentage change in world price

= percentage change m world output

IVrcentige change in world price

l/lOO of 1 per cent

25

= —4/100 of 1 percent

Slip 2 Compute the firm s elasticity of demand

percentage change in firm’s output
percentage change in world price

+ 40 per cent
+ 1,000
—4/100 of 1 per cent

The wheat farmer and the car producer is one of

difference between the
degree of power over the market The wheat farmer, because he is an in-
significant part of the whole market, has no power to influence the world
price of wheat On the other hand, the car manufacturer does have the
power to influence the price of cars, because his own production represents
a quite significant part of the total supply of cars
The first assumption listed on page 295 may now be re-worded as follows
every firm m a perfectly competitive industryis assumed to have ajirtn

elasticity of demand so high that it may neglect the influence that any
change in its output might have on market pnee Another way of saying
this IS that the firm acts as if the elasticity of demand for its product is

infinite
Students sometimes confuse the individual firm’s demand curve under
perfect competition with the market demand curve for the product The
market demand curve for the product downward sloping indicating that,
is
if there is an increase m suppfy, price will fell The demand curve facing a
 .

THE THEORY OF PERFECT COMPETITION 299

single producer is assured to be perfectly elastic because variations in his

production over the range that we need to consider for all practical purposes
will have such a small effect on price that the effect can safely be assumed
to be zero. Of course, if the single producer increases his production by a
vast amount, say ten-thousandfold, then this will cause a significant in-

100,000 100 000,000

. 200,000,000 300.000,000 400.000.000
Quantity produced (tons)

o
*w
CL

Pm

Fig 22.1 ‘The Demand Curve’

(i) Market demand curv'es.
{i
7 „ = -25 at 200 million tons)
(ii) Firm demand curve.
(77f
= CO)
(iii) Firm demand curve.
(’if
= 1,000)

crease in supply and he be unable to sell all he produces at the going

will
price. The perfectly elastic demand curve does not mean that the producer
could actually sell an infinite amount at the going price, but rather that the
variations in production that it will normally be practicable for him to
make \vill leave price virtually unaffected. This is illustrated in Figures 22.1

All three parts of the figure are drawn on an accurate scale. Part (i) shows
300 THE INTERMEDIATE THEORY OE SUPPLY

a market demand curve with an elastiaty of 25, parts (u) and (in) show

demand curves of infinite elasticity and with an elasticity of one thousand

Clearly, for most ordinary purposes, the two curves are indistinguishable

A DIGRESSION THE TRADITIONAL ASSUMPTIONS OF PERFECT COMPE

TiTiON We have said that to estabhsh ‘perfect competition’ it is necessary
that each firm be a price taker Rather than assume this directly we can
derive it from certain other assumptions about the nature of the product
produced and the market m which it is sold One such set is as follows

1 There is a very large number of sellers, no one of whom commands

a large share of the total market,
2 the products of different sellers arc identical, and buyers have no
preference among sellers

3 there arc so many buyers that sellers and buyers do not establish

personal relationships with one another, and

4 buyers are perfectly informed about the pnees of different sellers

These structural conditions arc often stated as being the assumptions of

perfect competition But the model of perfect competition depends upon
price taking, whether it is produced by this set of structural conditions or
by some other set Although we shall not go into them now, there arc other
‘
conditions that may also lead to pnce*taking behaviour

Total, Average and Marginal Revenue

We arc now able to use our theory of the demand for the product of a single
firm to discover how a firm s revenue is related to its output First, wc shall

define three revenue measures which are exactly analogous to the three cost
measures introduced m Chapter 20 total average and marginal revenue

1 Total revenue refers to the total amount of money that the firm
receives from the sale of its products, i e , to gross revenue This will vary
with a firm s sales so we may write

where TV? is total revenue and, as in the last chapter, Q is total production

1The significance of this point for testing the theory is very important failure to appreciate
It has occasioned much fruitless debate In the real t orld for example large numbers
of

buyers are never perfecOy informed Docs th % mean that perfect competition is never applic
able’ The answer is No provided that substantially less than perfect knowledge u sull con
ducive to sellers behaving as pnee talers Put diflerently the listed conditions may besuffcient
be price takers but they may not be necessary If they are not necessary an
for sellers to
empirical study that shows that they do not exist does not by itself refute the theory
 :

THE THEORY OF PERFECT COMPETITION 301

over some period of time.* Total revenue is obviously equal to the quantity
sold multiplied by the selling price of the commodity, i.e.,

TR=Q.p.
where p is the price per unit.
2 Average revenue is total revenue divided by the number of units

sold so as to give the average revenue per unit sold. Quite obviously, the
average revenue is the price of the commodity, so we may write^

AR = p.
It follows from this that the curv^e which relates average revenue to output
is demand cur\'e that relates price to output.
identical with the
3 Marginal revenue is the change in total revenue resulting from
an increase in the rate of sales per period of time (say per annum) by one
unit. The marginal revenue resulting from the sale of the nth unit of a com-
modity is thus the total re%'enue resulting from the sale of n units per annum
minus the total revenue that would have been earned if only n—\ (i.e., one
less) units had been sold per annum. The student should not think that n— 1

units are sold at some time and an extra unit

some later time. Marginal
at
revenue refers to alternative sales policies at the same period of time. Thus,
to find the marginal revenue of the 100th unit we compare the total revenue
resulting when 100 units are sold over some period of time \rith the total
revenue that would have resulted if 99 units had been sold over the same
period of time. In general we may write^

MR„ = TR„- TR„.,.

Total, average and marginal revenue of the firm in perfect

competition: In perfect competition the individual firm faces a perfectly
elastic demand curve for its product. Since the market price is unaffected
by variations in the firm’s output, it follows that the marginal revenue re-
sultingfrom an increase in the volume of sales by one unit is constant and

1 tVe are not considering problems of the Iiolding of stocks of commodities so, in this simple
theor)', we can equate production srith sales.

2 In case this should need demonstrating;

TR = Q.p

therefore AR = Q± = /'•

Q
3* This definition invokes finite changes. Students familiar with the calculus will recognise
marginal revenue as the derivative of total revenue with respect to quantit) sold;
ATR
302 THE INTERMEDIATE THEORY OF SUPPLY

IS equal to the average revenue If, for example, a farmer is faced with a
perfectly elastic demand curve for wheat at a market pnce of 13r per bushel
It follows that each additional bmhel sold w»U bring m that amount, c the i
,

marginal revenue is 13r, and the average revenue (equals total revenue/num
ber of bushels sold) ISrper bushel Tlic demand curve facingthe firm
IS also
IS thus identical with both the average and the marginal revenue curves
Alt three of these curves coincide in the same straight line showing that

p — AR = MR all remain constant as output vanes Total revenue docs, of

course, vary with output, since pnce is constant, it follows that total revenue

nses steadily as output nscs (The student should satisfy himself that, if he

draws a graph relating a perfectly competitive firm’s total revenue to its

output, he w ill obtain an upward sloping straight line which passes through
the origin ,

PROFIT MAXIMISATION IN A PERFECTLY-COMPETlTlVE

MARKET
We are now m a position to combine the theory of costs that we outlined in
the previous two chapters with the theor> of demand just developed, to
obtain a thcorv of the behaviour of the firm under competitive conditions
We assume throughout that in equilibnum, the firm maximises its profits
In this section we enquire into the short run equilibrium of the firm

Fig 22 2 The equilibrium of the firm in perfect competition

All relevant information shown m Figure 22 2 The cost curves of the

is
firm are repeated from Figure 20 5 and a perfectly elastic demand curve at
the going pnce of £Ou is superimposed Wc now wish to find the output
that maximises the firm s profits TTie d/C curve tells us the increase in total
costs that results from the production of one more unit The marginal
 .

THE THEORY OF PERFECT COMPETITION 303

revenue curve (equals price in perfect competition) tells us the increase in

total revenue resulting from increasing the rate of sales by one more unit.
Clearly as long as the former is less than the latter, an increase in the rate of
production and sales by one more unit will increase profits. If marginal
revenue of the nth unit exceeds the marginal cost of that unit, then the nth
unit adds more to revenue than to cost, and profit is increased by raising
production and sales to the level of n units per period. Thus, if profits are
to be maximised, production should be increased whenever marginal
revenue exceeds marginal cost (left-hand arrow in Figure 22.2). On the
other hand, if marginal revenue is less than marginal cost for the -nth unit
then that unit adds more to cost than it does to revenue, so that its produc-
tion lowers total profits. Clearly, if profits are to
be maximised, the level of
production should be reduced whenever marginal cost exceeds marginal
revenue (right-hand arrow in Figure 22.2). Since production should be in-
creased when marginal cost is less than marginal revenue and decreased
when marginal cost exceeds marginal revenue, it follows that profits will be
maximised when marginal cost equals marginal revenue (i.e., when the last
unit produced adds just as much to revenue as it does to cost while previous
unitsadd more and subsequent units add less to revenue than to costs)
In Figure 22.2 profits will be maximised at output Oa, where MC=MR=
ax= Ou. Price or average revenue per unit is ax {
— Ou) while average total
cost per unit is ar {
= 0t). Average price minus average
profit per unit is

cost per unit which equals {

rx = tu). Thus total
which equals average profit
profit per unit [rx] times the number of units produced {Oa), is indicated by
the area of the rectangle /«xr; total costs are indicated by the area of the
rectangle Oart (
= output, Oa, times the average total cost, ar).

It follows obviously from the above, that, if the marhet price changes, the
profit-maximising output will also change. If in Figure 22.2 price rises to
Ov, the output that maximises profits becomes Ob. The output at which
marginal costs equals marginal revenue is called the equilibrium output,
because this is the one level of output that the profit-maximising firm will
maintain. If any other level of output were temporarily achieved, either
marginal revenue would be less than marginal cost in which case profits
would be increased by lowering output, or marginal revenue would exceed
marginal cost in which case profits would be increased by raising output,

and in either case the firm would not be in equilibrium.

We have said that the firm will expand output until marginal cost equals
marginal revenue. When done, it will be producing at the most
this is

profitable level of output. It always possible that the most profitable level
is

of output will be less profitable than not producing at all. To cover this
possibility we must add the condition that the firm will not produce at all
unless average revenue exceeds average variable costs.
 THE INTERMEDIATE THEORY OE SUPPLY

Rules for the Profit-Maximising Behaviour of the Firm

Ue may now summarise the preccdinsj important discussion in terms of
'onie rules for profit maximisation
Rule 1 A firm should not produce at all if the total
revenues from its product do not equal or exceed its
total variable cost.

Tlic common sense behind this rule is that a firm always has the option
ot producing nothing If it produces nothing, it will have an operating loss

equal to Us fixed costs Unless actual production adds as much to revenue

as It idds to cost, it will increase the loss of the firm
Rule 2 Assuming that it pays the firm to produce at
will be profitable for the firm to expand output
all, it
whenever marginal revenue is greater than marginal
cost, and to keep expanding output until marginal
revenue equals marginal cost.
The common sense behind this rule is that any unit that adds more to
rr\ cnue than it does to cost met eases profits As long as marginal cost is less

than marginal revenue, increases in output uill increase profits On the

other hand, if the marginal cost of a unit exceeds us marginal revenue, that
unit IS decreasing the piofits of the firm and it should not be produced
nou restate these two rules as three necessary conditions for the
profit-maximising bclnviour of the firm (TTie conditions apply to a firm,
no matter in what type of market it operates
1 For a given output to be the profit-maximising
output, it is necessary that at that output

2 For a given output to be the proBt-maximising

output, it is necessary that for slightly smaller out-
puts MR>MC, and that for slightly larger outputs
MC>MR^
1 The graphical marpretation of these two conditions is that the profit maxiirasiog output

occurs at a point where \fC cuts MR from below The student should satisfy himself why m
(he adjacent sketch only output OA is a possible position of maximum profit

Fig 22 3 Nfargtnal cost = marginal

icvcnue as a necessary but not a
stifTicient condition for a profit
maximum

Output
 THE THEORY OF PERFECT COMPETITION 305

3 For a given output to be the profit-maximising out-

put, it is necessary that at that output total revenue is
equal to or greater than total variable costs. Another
way of putting thisis that it is necessary that average
revenue (price) be equal to or greater than average
variable cost.
When a firm has reached a profit-maximising position there is nothing
further it can do to increase its profits. It thus has no incentive to alter its
behaviour and we say that it is in a position of equilibrium. Economists are

not usually interested in equilibrium positions as such. Such positions are

required in order to do comparative static analysis which allows us to pre-
consequence of various changes in which we are interested. Through
dict the
comparative static analysis we hope to generate testable hypotheses about
the behaviour of firms in the real world. We must, however, discover the
equilibrium for the whole industry before we can use out theory to produce
hypotheses about how competitive industries will react to the changes in
which we are interested.

THE DERIVATION OF THE SUPPLY CURVE

The firm’s supply curve: So far in this chapter we have developed a
theory of revenue and then combined it with our theoiy' of costs to study the
conditions for profit maximisation. We now study how profit-maximising
output varies as price varies. We do this because we wish to develop supply
cuiT'es that show how output changes as price changes.

Figure 22.4(i) shows a numerical example of a firm’s marginal cost curve,

together with four alternative demand cuiT'es: if the market price were

Fig 22.4 The derivation of the supply curve from the marginal cost curv'e of the
price-taking firm.
306 THE INTERMEDIATE THEORY OF SUPPLY

£2, A if £3, D, if the market price were ^4, and

the market price were
Dj tf the market price were £5
The firm’s marginal cost curve gives the marginal cost corresponding to
each level of output We desire a supply curve that gives the quantity the
firm will supply at every pncc For paces below AVC the firm will supply
zero units (Rule 1) For paces above AVC the firm will equate price and
marginal cost (Rule 2) From this it follows at once that

In perfect competition the Arm’s marginal cost curve

above the A VC curve is its supply curve

This proposition is so obvious that it sometimes causes difficulty to the

student who is looking for something difficult and profound If you arc not
absolutely certain that you understand it, you should construct the firms
supply curve for yourself Given perfect competition, profit-maximisation,
and the actual cost curve of Figure 22 4(i), you can discover the output of
the firm corresponding to any given market price You can then plot the
firm s supply curve on a graph of your own by relating market price to
quantity produced by the firm Once you have done this, you will see that
the supply curve you have constructed is identical with the marginal cost
curve in Figure 22 4(i), above
In Figure 22 4(u), we have plotted such a supply curve for the firm The
points in this graph correspond point for point with those m Figure 22 4(i)

The short run supply curve for a competitive industry Now

that we have discovered how to find the supply curve for each firm in a
competitive industry, we have to determine how to find the supply curve
for thewhole industry To do this, we need to know the sum of the amounts
supplied by each firm for any given price Let us consider the simple
numerical example illustrated m Figure 22 5 Firm A’s supply curve is
shown in the first diagram and firm B s in the second The problem is to
construct an aggregate supply curve showing how the total supply of these
two firms will vary with price (on the assumption that they are in a perfectly
competitive industry} Below the pace of nothing is produced Betwwfl
and £2 only firm A produces, so the a^regate supply curve is identical
with firm A s curve At a pncc of firm A produces 400 units and firm B
produces 300 units, the total production at £5 is 700 units You should
check the construction of this a^regate supply curve for a number of alterna
tivepaces For any given pace (he a^regatc quantity is the sum of the two
quantities produced by the two firms at that pace
hundreds of firms the process is the same each firm’s MC
If there are
curve shows what the firm will produce at any given price, industry
p the
 THE THEORY OF PERFECT COMPETITION 307

supply curve relates the price p to the sum of the quantities produced by all

the firms. Thus we have the following result.

The supply curve for a competitive industry is the

horizontal sum of the marginal cost curves of all the
individual firms in the industry.^

In Part II we used short-run industry supply curves as part of our theory

of price. We have now derived these curves for a competitive industry, and

Price

Price

Fig 22.5 The derivation of the supply curve for a

competitive industry.

ii Quantity in hundreds

we have they are related to the behaviour of individual, profit-

seen how
maximising firms. An industry is said to be in short-run equilibrium when
price is such that supply equals demand. This was discussed in
Chapter 9
above. Figure 22.6 illustrates such an equilibrium.
We know that at price pE the quantity supplied by each firm setting
definition: we con-
Pe=MC will add up to the quantity OE. This is true by
structed the industry supply curve in such a way as to make it true. The
industiy^ issaid to be in equilibrium when supply equals demand and thus
when, for each firm, marginal cost equals price.
understanding
Students who have committed this proposition to memory
without fully it
1

sometimes get confused between the horizontal and vertical summation of curves You should
now reread footnote 1 on page 88
308 THE INTERXIEDIATE TIIEORV OF SUPPLY

Pt
Ftg 22 S The equilibrium of a
competitive mdustn

Output

Profits and Losses tn the Short Run

Wc have shown tint the firm ts run cquiltbrium at a rate of output
in short

for which Its marginal cost equals its marginal revenue Although we know
from this that the firm is maximising its profits, ue do not know how large
these profits are It is one thing to know thtt a firm is doing ns w cll as it can
in the circumstances, it is quite another thing to know how well it is doing

Fig 22 7 Short run equilibrium of a

competitive firm
p, li the pnee at which p = 'fC
Pfc
IS the price at which ATCa^
minimum
IS the pnee at which A^Cisz
minimum

Output
 THE THEORY OF PERFECT COMPETITION 309

Figure 22.7 shows three possible positions for a firm in short-run equi-
librium. In all cases, the firm is ma.\imising its profits at price by pro-
ducing an output of Oq, but in case (i) the firm is making profits in excess
of all costs: in case (ii) it is just covering all costs; and in case (iii) it is

making losses.’ In case

might be better to say that the firm is mini-
(iii) it

mising its losses rather than maximising its profits, but both statements
mean the same thing: the firm is doing as well as it can do, given its costs
and prices.^
All three of these positions represent possible short-run equilibrium posi-
tions for the profit-maximising firm in perfect competition. They are all

equally possible in the short run. But only one of them will persist in the
long run. Before reading on you should make a serious effort to discover for
yourself which one ivill persist in the long run and why.

Long-Run Equilibrium of the Competitive Firm and Industry

In the long run the firm would continue to adjust to the market price by
producing where p — MC. But should price drop below ATC and remain
there (below pg in Figure 22.7), the firm would want to withdraw its re-
sources from this industry. You will recall that, as we have defined them,
total costs measure the returns on the best alternative use of the resources.
When average revenue is below average cost in this industry, the firm is not
using its resources to maximum advantage. When it gets the chance, it will

shift them. It gets this chance as its fixed factors, such as machines,
gradually are used up. When they are used up, the firm will not replace
them; it will gradually liquidate its investment and withdraw (or exit) from

the industry. In the long run, the firm ndll remain in production only if
price at least equal to average total cost. Thus, firms in situation (iii) will
is

withdraw from the industr)" in the long run. As firms withdraw, the supply
curve of the industry' will shift to the left and prices will rise. Firms will con-
tinue to prices will continue to rise until the firms remaining
withdraw and
in the industry are able to cover all their costs - until, that is, they are in
situation (ii).

Now, assume that, in the short-run equilibrium position, firms are in

situation (i). The individual firms are making profits in excess of the returns

available on the alternative uses of the resources devoted to the industry.

This fact will attract new investment into the industry in the form either
Do not forget that costs include a return on capital which firms themselves call profits.
1

When we say losses therefore we do not necessarily mean the company s statement will show
losses, what we do mean that all opportunity costs are not being
covered.
is
ivell, just to indicate
2 In case (iii), we have drawn in the average \'ariable cost cun’e as
that, in the short run, it pays the firm to stay in production. If price fell
below /iz, it ivould pay
the firm to shut down production.
310 THE INTERMEDIATE THEORY OF SUPPLY

of new firms entenng the industry and/or of existing firms expanding their
plants The cases are not different For simplicity we shall illustrate for the
*
case of new firms entering the industry
Suppose, m response to high profits for 100 ejtisting firms, ten new firms
enter an industry Market demand docs not change, and the market supply
curve that formerly added up the outputs of 100 firms now must add up the
outputs of 10 firms At any price, more will be supplied because there are
1

more suppliers Bui this shift in supply will mean that pE m Figure 22 6 is
no longer the equilibrium price As Figure 22 8 shows, pnee will have to
fall and both new and old firms will have to adjust their output to this

new pnee
In Figure 22 8, we assume that the profits pE attract new entrants
whose output shifts the supply curve to ^ This causes price to fall to pf At
this price before entry an amount equal to only Oa would have been pro

duced The extra output aF is supplied by the new productive capacity

After the new entry has occurr^ and market price has fallen to pf, there
are three possibilities
1 At Pf the individual firms may still be earning profits in excess of

total costs (In terms of Figure 22 7 Pf may be above pg ) There continues

to be incentive for resources to enter the industry, and a further increase in
supply and a fall m price are to be expected
2 At pf the individual firms may be suffering losses (m terms of Figure
22 7 pf is less than Pg) In this case no new firms will wish to enter the
industry, and existing firms will not wish to replace fixed equipment as it

wears out Ultimately there will be exit from the industry, and the market
supply curve will shift to the left
3 Atpf there may be aero profits (in terms of Figure 22 7 pf=pg~ATC]
*
In is no incentive to enter or to exit
this case there
One feature of what we mean by long run equilibnum is that the industry
is neither expanding nor contracung Ibr a given state of demand Only the
third of our three possibilities has this feature From this it follows at once

that, if there is freedom of entry and exit, existing firms in equilibrium must
be at a position wherep=:A7C’

Economies of Scale
If the industry is to be in equilibrium, it is necessary that each firm should
just be covering total costs when it is doing as well as it can with its existing

1 But which of the two possibiUues that ocean a not merely a matter of chance See note 2
page 312
2 Recall once again that we have included m
total cost the wages of management,
returns

to capital nsk premiums and all other payments to resource* at the level of their best aliema
live use (see Chapter 19) Zero profluover opportunity cost implies no hardship to the firm
 ,

THE THEORY OF PERFECT COMPETITION 311

plant. But this is not enough. If the firms are subject to economies of scale,
they will be able to reduce their average costs of production by building
larger plants. Since they assume that they can sell all they want at the going
price, they will see an expansion of the scale of their operations as necessarily
increasing their profits. This because price { = average revenue) does not
is

fall, whereas average costs do. Thus, firms will expand their scale of opera-
tions as long as they are in a perfectly competitive market and are subject
to falling long-run costs. From this it follows that, in long-run equilibrium.

Fig 22.8 A shift in the supply curve

caused by new entrants.

each firm must be at a point at which no further unexploited economies of

scale remain within its grasp.
This is depicted geometrically in Figure 22.9 which shows three different
sets of short-run cost curves. If the cost curves are those marked MCi and
SRACi (where long-run costs are decreasing), apparent that the indus-
it is

tr)' will not be in long-run equilibrium, even if price is p„, because a firm can

make by building a bigger plant thus driving costs down below

profits
output. If the cost curves are those labelled MC^ and SRAC2 (where long-
run costs are increasing), apparent that the industry will not be in
it is

long-run equilibrium, even is pf„ because a firm can make profits

if price
by building a smaller plant. Only if price is p'^ output is OE, and firms are

operating plants with the cost curves MC^ and SRAC^ will it be true that
no firm can increase its profits by changing either its output or its scale of
plant. Thus output OE with price p^ represents the long-run equilibrium
situation for a competitive industry.
Given two assumptions, first, that individual sellers are price-takers and,
second, that there is freedom of entry and exit, the conditions of competitive
equilibrium may be restated as follows:

1 Every firm produces where price equals marginal cost.

2 Every firm produces where price equals short-run average total cost.

3 Every firm produces where price equals long-run average total cost.
312 THE INTERMEDIATE THEORY OF SUPPLY

These three conditions are all met only at the minimum point of the long
run average cost curve, as illustrated at output OF m Figure 22 9 The '

first condition results from the firms maximising their profits for a
given
price The second and third result from the incentives for new resources to

enter into (or existing resources to exit from) the industry if they are not
satisfied with the profits that they arc earning *

1 If the long run cost curve of a firm m a competmve industry does not have a minimum
point there is no competitive equtlibnum Whether long run costs eventually increase u thus
of paramount importance to theonsis who depend upon the model of a perfectly compeuuve
industry
2* From this we can see that whether the response to a nse m demand is an increase m the

size of existing frms or an expansion in the number of firms is not arbitrary In long run
equilibrium alt firms are at the minimum point of their long run /17C curve Unln: fl ciangt in

demand shifts this m n mum point the fall tong ran adjustment must be borne bf a change in the numiir

offrms Consider a demand

rise in In the short run all existing firms expand output unul
MC= MR The new f rms enter and existing firms contract along iheir MC curves until by
repetition of the argument in the lext they reach the low point on iheir long run 4TC curve
again Existing firms will expand capaaiy m the long run only if the lowest po nt of their
A TC curve shifts to the right This requires what economise call external economics of scale
when the industry expands i( uses more ofihe input ofaiBwie/wlwrfiiirfarlry subject to economies
of scale so that the price of this input falls If die optimal proportion of input combinations
vary as output varies then it is possible that the change in the pnee of this one input will shift

the low point of the firm s curves to the nght and thus lead existing firms to produce more
ATC
in long run equilibrium Otherwise all long run adjustments take the form of variations in the
nvirobei of firms the output of surviving firms remaimi^ constant in long run equil bniint
 :

APPENDIX TO CHAPTER 22

MATHEMATICAL PROOF OF
THE RULES FOR THE PROFIT-
MAXIMISING BEHAVIOUR
OF A FIRM^

This appendix proves, for the student and third say that, if tt„ is a maximum,
who is uneasy about common-sense a change of one unit in either direction
arguments, the propositions we asserted cannot increase profits.
on pages 304-5. The proofs illustrate Rule 1 A profit-maximising firm
clearly how a little bit of mathematics should not produce at all if the
can settle some things once and for all.
total revenues from its product do
Let ?:„ = TR„-TC„, not equal or exceed its total vari-

where Tr„ is the profit when n units are

able cost.
sold.
PROOF OF RULE 1

Let zJ77„ = From above

(1)
where A7t„ is the change in profit result- ^ 770 ,

ing from the sale of the «th unit. which can be re-written
If the firm is maximising its profits
TR„-TVC„-TFC„ >
by producing n units, it is necessary TRo-TVCo-TFCo. (4)
that the profits at output Q„ are at least
But we note by definition that
as large as the profits at any other out-
put. Consider three specific alternative TRo = 0, (a)

outputs : 0, n—\, n+\. If the firm is TVCo = 0, (b)

maximising its profits at output n, then TFC„ = TFCo = K, (c)

^ TTq (1) where A is a constant.

By substituting (a), (b) and (c) into
^ 77„_i (2)
(4), we get
^ 77„+i (3)
The condition says that profits
first
TR„-TVC„ > 0,

from producing must be greater than from which we obtain

profits from not producing. The second TR„ ^ TVC„.
1 This .Appendix is optional. This is the result desired.
314 THE INTERMEDIATE THEORY OF SUPPLY
On a per unit basis, it becomes ALTERNATIVE PROOF OF
TR„ TVC„
RULE 2
Using elementary calculus, this result
may be shown m a manner that sharp,
where Q„ is thenumber of units Since ens the final conclusion
TJi„ = Q„p„, where is the price when

n units are sold, this may be re-written „

= TR„~TC„
p„^ A VC„ each of which is a function of output Q
Rule 2 Assuming that it pays the To maximise it it is necessary that
firm to produce at all, it will be
profitable for the firm to expand (d)

output whenever marginal revenue

IS greater than marginal cost, d^
and to keep expanding output until
<0 (e)

MR = MC. dn dTR dTT

d<?
'
dQ dQ
PROOF OF RULE 2
From (d), a necessary condition of
By definition,
maximum n is AIR—AICssO, or AIR-
A7r„ ^ AiC For <?„ to maximise profits requires
= {TR„-TC„)-{TR„ i-TC^.i) AiR,= AfC„ Now,
^ {TR„-TR,.,)-(TC,-TC,.^)
^7r„ = MR„-A/C„ d^ir _
Now, if ^^R^ — AfC„ > 0, It IS possible
d^ “ dQ " d<?

to increase profit by increasing output,

From (e), a necessary condition of
thus this situation does not represent
maximum v is
profit maximisation (More formally, it

implies jr„+,>7r„, which violates (3) dAIC

dAfR „
above ) Similarly if A^R„ — AIC„<0, it
IS possible to increase profit by decreas-
^~~dQ
ing output (This inequality implies which says that the slope of MC must
7r„_ I
> 7T„, which violates (2) above ) be greater than the slope of Affl Taken
Hence, for maximum profits, AfR„= with the previous result it implies that,
A{C„, m the neighbourhood where for Q„ to maximise n, AIR„—AiCf at a

AfR„ i-AfC„ i>0 point where AlC cuts AfR from below
CHAPTER 23

THE THEORY OF MONOPOLY

In Chapter 22, we developed a theory' of the behaviour of a firm and indus-

try under conditions of perfect competition. In this chapter, we deal with
a second, verj' different type of market structure, monopoly.
It is convenient for the present discussion to think of monopoly as the
situation in which an entire industry is supplied by a single seller. We shall
call this seller the monopolist. Further on in the chapter, we shall define
monopoly in a less restrictive way.
In the case of perfect competition each single producer has no power
whatsoever over the market; he can greatly increase or diminish his pro-
duction without affecting the market price significantly. A monopolist, on
the other hand, has power to influence the market price; by reducing his
output he can force up the price and by increasing his output he can force
price down.

THE RELATION BETWEEN AVERAGE REVENUE AND SALES:

THE MONOPOLIST'S DEMAND CURVE

The reader should no^v review the section on total,average and marginal
revenue on pages 300-302. Since the monopolist is the only producer of the
product, the firm’s elasticity of demand (see page 297) is the same as the
market elasticity of demand. A 10 per cent variation in the firm’s output is
a 10 per cent variation in the industry^’s output. The demand curve facing
the monopolist is therefore identical with the market demand curve for the
product. Since price is identical with average revenue, the demand cur\'e is

also the curve relating average revenue to quantity' produced.

In the case of perfect competition the marginal and average revenue
curves are identical. Since the price is unaffected by variations in the firm s
output, it follows that the addition to revenue resulting from increasing the
316 THE INTERMEDIATE THEORY OF SUPPLY

levelof sales by one more unit is the market price of that unit (sec page 301)
In the case of a monopolist, however, the marginal and average revenues
are not identical The monopolist faces a downward-sloping demand curve
and the sale of additional units of his product forces down the price at which
all units must be sold

The relation between the average and marginal revenues of a

MONOPOLIST We start by considering a simple example Assume that a
monopolist is selling 100 units of some commodity per year at a price of
£\ (=240d) Total revenue IS 100 x240d=24,000if Now assume that the
rate of sales is stepped up to 101 units per year and, as a result, the pnee is
driven down by 2d to 238rf per unit The price and average revenue when

Fig 25 I The change in a

monopolist s revenue resulting from

a small decrease in his selling

pnee

101 units are sold » 193 lOd but what is the marginal revenue’ Clearly the
total revenue is 101 x 238i/=:24,038</ The marginal revenue of the lOl't
unitIS defined as the increase m
total revenue when sales are increased from
100 units to 101 units per period, and, in this example, it is 38rf or 3 j 2d, m
spite of the fact that the lOIst unit sold for I9j lOrf The reason, of course is

that if only 100 units had been sold they vrould have commanded a higher
price per unit The must be cut by 2d on all 100 units m order to m
price
loss
crease the rate of sales from 100 to 101 units per year Thus there is a
of 2<f per unit on all 100 units, resulting from the increase of sales from 100
to 101 This loss, which amounts to 100x2rf=200rf, must be set against the
238rf gained from the sale of one more unit Thus, the net change in total
revenue is 238(/— 200<f=38i/ TTie maiginal revenue associated with the
101st unit sold will alwaysbe less than the pnee at which all 101 units are
sold, aslong as 100 units could have been sold at a higher price Only if lOO
units could be sold at 240i/and 101 could be sold at the same price would
the marginal revenue be equal to the price The reader should now calcu
 THE THEORY OF MONOPOLY 317

late for himself the marginal revenue of the 101st unit if the price is driven
down from 240«/ to 239^(7 when sales are stepped up from 100 to 101 units
per annum.
may be illustrated graphically. In Figure 23.1
Exactly the same argument
sales are assumed
from the rate of OJ units per year to OA' units.
to increase
The extra revenue gain from the sales of the extra unit is equal to the area
of the rectangle labelled ‘gain in revenue’. This, however, does not represent
pure gain, for OA of the units could have been sold at price Op, but they
must now be sold at price Op', because the extra unit is to be sold as well.
This makes a loss of p'p per unit on all OA units, which is indicated by the
area of the rectangle labelled ‘loss in revenue’. Marginal revenue of the
extra unitis equal to the difference between these two areas.*

Thus, we conclude that, for any given unit, the marginal revenue result-
ing from the sale of that unit will be less than the price obtained for the
unit. The reason for this is that the monopolist has power over the market
if he tries to increase his volume of sales he will spoil the market for the
quantity he is already selling. In Figure 23.2 both the demand curve and
the marginal revenue curve for a monopohst are plotted. The curve D shows
the price corresponding to any given level of output, while the curve MR
shows the marginal revenue corresponding to the unit of output indicated.
Thus, for example, the price obtained when OC units are sold is Or, while
the marginal revenue resulting from a rise in sales from OC— 1 units to OC
units is only Cs. The price obtained when OB units are sold is Bi, but the
increase in revenue when sales go from OB — 1 unit to OB units is nothing
at all! (This is, of course, perfectly possible as the following numerical
example will show; 4 units are sold for ;^10 each, while a cut in price to £8
is necessary to raise sales to 5 units.) When
OB' units are sold the price is Bi/,
but the change in total revenue resulting from the sale of one more unit is
negative and equal to Ev, so that total revenue is decreased by the sales of
one more unit! (That an increase in sales could reduce revenue is, of course,
not at all surprising if a given percentage fall in price is met by a smaller
:

percentage rise in sales, total revenue earned will decline.)

1 It is easy to prove algebraically that, if the demand curve slopes downwards, marginal
revenue is always less than price.

= TR„^,-TR„
= {n+i)p„+i—npn
— ^Pn+I 1

= n{p„+i—p„)+Pn+i
slopes do\\nwards, p„+\ (the price ruling when a+l units are
sold)
Since the demand curve
will be than p„ (the price
less ruling when n units are sold). Thus, the of the (n+ MR
l)th unit

is less than //„+ by n times the tall in price from/>, to p„+i.

,
318 THE INTERMEDIATE THEORY OF SUPPLY

Fig 23 2 Average and marginal

revenue curves for a monopolist

PROFIT MAXIMISATION IN A MONOPOLISED MARKET

In Figure 23 3 we repeat the relevant cost curves from Figure 22 2 Accord-
ing to our theory of costs, the technological facts of life arc the same for the
monopolist as for a competitive firm, so that the short-run cost curves have
the in both cases The difference lies in the demand condmons
same shape
The compeuuve firm {of Figure 22 2) is faced with a perfectly
perfectly
elastic demand for its product while the monopolist of Figure 23 3 is faced
with a downward-sloping demand curve By exactly the same argument as
was used inChapter 22, pages 302-3, the output that maximises the
monopolist’s profitsis the output for which marginal cost equals marginal

revenue If marginal cost is less than marginal revenue, profits can be in-
creased by producing more because maigmal units are adding less to cost
than to revenue If marginal cost «iceeds marginal revenue, profits can be
increased by producing less, because marginal units are adding more to cost
than to revenue In Figure 23 3 the output that maximises profits is OA At
 THE THEORY OF MONOPOLY 319

Fig 23.3 The equilibrium of a monopolist.

this level of output, marginal cost and marginal revenue are both Ad. The
price is Aa (since the demand curve shows the price consumers are prepared
to pay for various quantities offered) and the average cost per unit is Ab.
The average profit per unit is ba. Thus, total revenue of the firm is indicated
by the area of the rectangle OAap (^prices times quantity) and total profit
by ebap ( = profit per unit times number of units sold).
If the firm produces at less than OA, say OA', in Figure 23.3
some output
then marginal cost than marginal revenue. This means that another
is less

unit produced and sold adds more to revenue than to cost, and so would
320 THE INTERMEDIATE THEORY OF SUPPLY

increase profits (by^^ m this case) If profits arc to be maximised the output
should be increased \vhcnc\cr marginal cost is less than marginal reventie
If tlie firm produces more than OA, siy OA', then marginal cost exceeds
marginal revenue Thelast unit produced adds more to cost than it does to
revenue, clearly the production and sale of the last unit causes a reduction
m total profits (by hk m this case) If profits arc to be maximised, output
'
should be reduced whenever marginal cost exceeds marginal revenue

ABSENCE OF A SUPPLY CURVE UNDER MONOPOLY

The next stage m our study is to consider the relation between the price of
the product and the quantity supplied under monopoly In the case of per-
fect we were able to discover a unique relation between pnee
competition
and quantity supplied and this gave rise tosuppl> curves for each firm, and,
by aggregation, to a supply curve for the industry In perfect competition
we know the industry short-run supply curve as soon as wc know the mar-
ginal cost curves of tlic individual firms This is because the profit-maximis-
ing firms equate marginal cost to price so that given marginal costs, wc
know how much wiH be produced at eacli price
exactly
In monopoly such a unique relation bcivveen market pnee and quantity

supplied does not exist I.ike all profit maximising firms, the monopolist
equates marginal cost to marginal revenue, but, unlike firms in perfect
competition, marginal revenue docs not equal pnee, hence the monopolist
docs not equate marginal cost to pnee Under these conditions it is possible
for difTercnt demand conditions to give nse to the same output but to
differing prices In order to know the amount produced at any given price,
we need to know the demand curve as well as the marginal cost curve
The proposition that, when a firm faces a downward-sloping demand
curve, there is no unique relation between price and output is illustrated in
Figures 23 4 and 23 5 In Figure 23 4 wc see two different demand curves
both resulting in the same output, Oq^, but in two different prices When
demand is D, marginal revenue equals marginal cost at output Oqi and
price is Op When demand is
I
marginal cost again equals marginal
revenue at output Oq^ but pnee « at OPi Thus, it is possible to have the

be
1 It u a common mistake to think that a monopoltst who is not maximising proriu must
making losses This is of course not correct At output O 1 where UC= t/f? total profits are
total
as large as possible \i output O where UC> t//f, additional units sold are reducing
profits but total profits are still positjxe (they are equal to a'i' times In fact total profits
O i")
remain positive until output reaches OB units At average
that level average total cost equals
revenue so that total costs equal total resenue Should output be increased beyond OB total

profits would become negative

 :

aj ' ' D,
MR, MRj 4
Quantity Quantity

Fig 23.4 Two different prices Fig 23.5 Two different outputs
associated with the same output. associated with the same price.

same output sold at various different prices even though the marginal cost
curt'e does not shift.’

In Figure 23.5 we illustrate the same general point by showing a case in

which the same price is associated with two different quantities; when de-
mand is 2)3 marginal revenue equals marginal cost at output Oq^ and the
resulting price is Op^. When demand is D4 marginal revenue equals ,

marginal cost at output Oq^ but price is again Op^.^

1 Strictly speaking we have asserted the result but have not proved it. tVe can get closer to
a formal proof as follotvs:
(a) There is a unique relation between a demand curve and a marginal revenue curve; for

any given demand curve there is only one marginal revenue curve consistent \vith it.
(b) Take the MR, curve and pivot it through the point at which it intersects the MC curt'e;

this ensures that output remains constant at Oqj.

(c) As thishappens the demand curve consistent with the new marginal revenue curve will
shift and (unless the D curve shifts through the point with coordinates Op, and OQ2) price
tvill change.
For a formal proof we would have through what point the demand curve w'ill
to establish
pivot. It is quite easilyshown demand curve the quantity ordinate is
that for a straight-line
twice that of its corresponding marginal revenue curve for any given price. Thus, pivoting the
marginal revenue curve through the point at which MC—MR, pivots the D cutve as shown
in Figure 23 . 4 .
2 Again we have merely asserted this result rather than proved it. If we make use of the

result quoted in the previous footnote, we may proceed as follows

(a) Draw any demand curve D^', draw its corresponding marginal revenue curve, and
determine the price O/13 and the quantity Oq^.
desired
(b) Now pick a new quantity Oq^. greater than the original quantity Oq^. Since it is

follows that the new de-

to have the new quantity Oq^ associated with the original price O/13 it

mand curve must go through the point which has O/13 and Oq^ as its coordinates.
(c) Since we require Oq^. to be the profit maximising position, the
new curve must MR
II
322 THE INTERMEDIATE THEORY OF SUPPLY

FIRM AND INDUSTRY, SHORT AND LONG RUN

A monopolist, as we have defined him, is the only producer in an industry
There is, thus,no need to have a separate theory of the firm and the in
dustiy, as was necessary with perfect competition The monopolist u the
industry
In a monopolised industry, as in a perfectly competitne industry, profits

provide an incentive for new firms to enter the industry If the monopoly is
to persist in the long run, other firms must be discouraged from cntenng the
industry There must be, in technical terms, parr/ers to entry These
may take several forms patent laws may create and perpetuate monopolies
by conferring on the patent holder the sole right to produce a particular
commodity The government may grant a firm a charter or a franchise that
prohibits competition by law Monopolies may also arise because of econo
mies of scaleThe established firm that can produce at a loiver cost than any
new, and necessariVy smaU, tompewtor may retawv a monopoly tWroogh a
cost advantage A monopoly may also be perpetuated by force, or by threat
potential competitors can be miimidaied by threats ranging from sabotage
to a price war in which the established monopoly has sufficient financial
resources to ensure victory
is no entry mio a monopolistic industry, the profits of a
Because there
monopolist may persist over time In perfect competition, the long run
differed from the short run because of the process of entry In monopoly,
where entry is prevented, the long run equilibrium is no dilTcrent from the
short run equilibrium

THE NATURE AND EXTENT OF MONOPOLY POWER

The monopolist that we considered above was able to select the price-
quantity combination that maximised his profits The price he selected
turned out to be the market pnee of the industry at equilibrium, and the
quantity he supplied was the quantity demanded at equilibnum We as

sumed that his demand curve was the industry demand curve and that no
other seller, present or potential, threatened his position as the sole supplier.
In what way docs ihis theory of a monopolist differ from the situauon of
pro-
any producer who faces a downward-sloping demand curve’ Such a

mtersect the MC curve at output Oq, i e it must pass through the point at which the A/Chne
intersects a vertical line raised from 9*
(d) Now demand cufve must have a q coordinate twice
using the previous result that the
that of Its MR curve for each price
it follows that the demand curve must go through
the point
which has coordinates 2(094) and O^sthebe^htof^/^aal O9* Steps (a) and (d) eachpve
us a point on the demand curve Ifwe confine ourselves to linear curves then these are sufficient
to determine the whole curve
 THE THEORY OF MONOPOLY 323

ducer would also choose the price-quantity combination (where MR=MC)

that maximised his profits. But this price might turn out not to be the long-run
equilibrium price, because his choice of a particular price-quantity combina-
tion might itself lead to changes in the behaviour of others that would shift
his demand curve. We saw earlier that a demand curve represents the re-
lationship between price and quantity when all other things remain un-
changed. But the very act of choosing a price may lead to changes that are
assumed not to occur when demand curves are drawn. Such changes will
lead to a demand curve.
shift in the
, The Coca-Cola Company is the sole producer of
Consider an example.
Coca-Cola and faces a downward-sloping demand curve for its product.
Suppose that this demand curve shows that, if the Coca-Cola Company
cuts the price of its product by 20 per cent, and if all other soft-drink sup-
plierskeep their prices at their present level, the Coca-Cola Company will
experience a 50 per cent increase in sales, which would result in a significant
increase in profits. Since Coca-Cola is unquestionably free to cut its price
by 20 per cent why does it not do so? The answer is that if it did, it strongly
suspects that ceteris paribus will not hold - that its very action of reducing
prices will lead other sellers to reduce their prices. If they do, Coca-Cola’s
demand curve will shift to the left, and its sales may end up increasing not
by 50 per cent but perhaps only by 10 per cent. In such an event, profits
would decrease rather than increase.
Shifts in the demand curve of a seller as a result of the seller’s pricing de-
have two main sources. The first source (illustrated in the Coca-Cola
cisions
example) is found in the reactions of firms that sell products that compete
to some extent with the seller’s product. The second source is found in the

entry of new sellers who succeed in capturing part of the sales that the seller
included in ‘his’ demand curve. Shifts of either of these kinds threaten the
monopoly power of the seller, and will reduce the profits the seller would
have received if they had not occurred.
Is itnot possible to imagine a firm -without any competition whatsoever ?
A firm may have a complete monopoly of a particular product, but every
product has some substitutes for the ser\’ices it provides.
Some products have fairly close substitutes, and even a single seller
producing such a product will have rivals for his customers’ expenditure.
Not only will his demand be reladvely elastic, but new entrants into closely
allied fields may shift his demand curve. Even if there is no very close sub-
stitute for his product at any moment, high profits may induce rivals to de-

velop a close substitute to cut into his market.

Monopoly power exists when a firm is at least to some extent insulated
from loss of customers to other sellers. Since no firm is perfectly insulated
from all other products or for all time, perfect monopoly power does not
324 THE INTER^fEDIATE THEORY OF SUPPLY

exist But the extent of shifts in a seller s demand curve due to other pro
ducers’ actions is a quantitative variable m some cases, such shifts may be
very minor, in others, they may be very large Thus the degree of monopoly
power may vary and it is useful to think of it as a variable
In general, the extent of monopoly power will be greater, the smaller the
shifts indemand caused by the reactions of sellers of other products and the
smaller the shifts in demand caused by the entry of new sellers How laigc
these shifts will be depends upon a great variety of other variables which
will be considered in Chapter 26
It IS well to note at this point that a single seller is neither sufficient for,

nor necessary to, monopoly power The Coca-Cola example demonstrated

that It is not sufficient It is not necessary, because two sellers (who are

not faced by the threat of entry) could agree between themselves to set a

common pnee and to share the market Their behaviour would be no

different than if they were a single seller

How do we Measure Monopoly Power?

Our theory predicts that behaviour in monopolistic markets will differ from
behaviour in perfectly compeutive markets If we are to test this theory, we
must be able to give some idea of the extent of monopoly power in various
markets Also it is often felt by government agencies (acting on behalf of

the people) that uncontrolled monopoly power is undesirable These govern

mental agencies must know where monopoly power exists if they are to
control or eliminate it For both of these reasons, and for others as well, it

becomes important to measure the extent of monopoly power in various

markets This is not an easy thing to do
Several measures of monopoly power have been suggested ‘ Ideally, one
would like to compare the prices, outputs, and profits of firms m any in
dustry with what prices outputs, and profits would be if all firms were under
unified (monopoly) control and were fully insulated from entry But this

hypothetical comparison docs not lend itself to measurement

In practice, two alternative measures are widely used The first of these is

.4^wis'r,swV-jAv5iv.nat'S'Ab5svTAVv?.0iarA\nrjiCtcvtaljnarkel
sales controlled by the largest group of sellers Common types of concentra
tion ratios cite the share of total industry sales of the largest four or eight
firms ^ Whether concentration ratios measure effective monopoly power is

1 Two of the most famous early discussums are A P Lemer The Concept of Monopoly
and the Measurement of Monopoly Power Bnum vf Etmomic Stadia I 157-75 and
Joe S Bam The Profit Rate as a Measure cf Monopoly Power , Quarterly Journal of Economics
February 1941 pages 271 93 Neither rf these are suitable for beginning students
2 In Chapter 38 we present some concentration ratios for UK industries
 THE THEORY OF MONOPOLY 325

a matter of some debate among economists. Clearly, market share is one

measure of the potential power to control supply and set price. The inclusion
in concentration ratios of the market shares of several firms rests upon the
possibility that large firms will adopt a common price-output policy that is
no different from one they would adopt if they were in fact under unified
management.^ Concentration ratios measure actual exercise of monopoly
power only if this ‘possibility’ occurs. High concentration ratios may be
necessary for the exercise of monopoly power, but they are not sufficient.^
It is, nevertheless, interesting to know where potential monopoly power does
exist.

The second approach to measuring the extent of monopoly is to examine

profits. If profits are and remain ‘high’,^ so goes the logic of this measure, it
is indirect evidence that neither rivalry among sellers nor entry of new firms
prevents existing firms from pricing as if they are monopolists.
While neither of these measures is ideal, each is of some value and each is

widely used. In fact, concentration ratios and high profit rates are them-
selves correlated.*^ Because of this, alternative classifications of industries,
according to their degree of monopoly power measured in these two ways,
do not from each other very much. Because monopoly power is
differ
measurable, and in spite of some difficult problems of measurement, the
theory of monopoly is often very useful to both economists and policy-
makers, and it is widely used by them.

A Perfect Monopoly?^
We have seen in the previous section that the concept of monopoly has
certain difficulties. No firm has complete monopoly power. Some firms
which have fairly close competitors will face quite elastic demand cur\'es;
other firms with fewer close competitors will face less elastic demand curves.
In order to clarify the concept of a monopoly, some economists have tried
to define a perfect monopoly which would be at the opposite end of the
spectrum from perfect competition. Two ideas have been put forward, both
of which are unsatisfactory. Some people have said that a perfect monopolist

1 Such behaviour might occur in various ways. The most common suggestion is that they

are in tacit collusion.

2 We shall discuss this point more fully in Chapter 26.

3 By ‘high’ profits the economist means returns sufficiently in excess of all opportunity costs

that potential new entrants desire to enter the industry'. For empirical applications this will
usually be taken to mean rates of return that are substantially higher than the average of all
industries having roughly the same degree of risk.
4 The classic by Joe S. Bain, ‘Relation of the Profit Rate
study is to Industry Concentration’,
Quarterly Journal of Economics, August 1951, pages 297-304.
5 This is an optional section.
326 THE INTERMEDIATE THEORY OF SUPPLY

would be a firm selling a good for which the demand was perfectly mtlastic
(this seemed the opposite to a perfect competitor who faces a perfectly

elastic demand) The reader should be able to see the contradiction in the

idea of a profit-maximising monopolist facing a perfectly inelastic demand

curve
Such a case would never come about No profit-maximising monopolist
would ever sell at a price at which the demand for his product was perfectly
inelastic Clearly he would go on raising price while not losing sales and thus
go on increasing profits In fact he would continue to raise his price, not
only until the demand ceased to be perfectly inelastic, but until the demand
ceased to be inelastic at all (for marginal revenue is negative as long as
'
demand is less than one)
elasticity of

The second idea put forward was that a perfect monopolist would be a
firm that had a monopoly of all goods and hence a firm which took in the
consumer’s entire income (ignonng saving) The monopolist would thus be
faced wtth a demand curve of unit elasticity any increase in price would be
met by a proportional fall in quantity bought, so that total expenditure
(=total consumers’ income) would remain constant
The problems here are more subtle than in the case of the first idea pro-
posed Let us fint consider what a profit maximising monopolist would do
if faced with a demand curve of unit elasticity The marginal res enue corre-

sponding to such a demand curve is, of course, zero A I per cent increase in

IS always accompanied by a
sales 1 per cent fall m pnee, while a 1 per cent
reduction in sales causes a 1 per cent rise in pnees, so that total revenue re

mams unchanged Clearly the monopolist will cut down output, thus reduc-
ing costs but leaving revenue unaffected, and hence increasing profits If
the demand curve were truly of unit clastiaty throughout its whole range the
monopolist’s output would approach zero, while his price would approach
he sold virtually nothing at all and sold it at an infinitely
infinity until finally
high pnee, absorbing the whole of income as his profits Now all of income '

becomes the monopolist s profits and, since nothing is produced, no money

will be paid out of production Thus, next period the only pur-
to factors
chaser left for the monopolist to exploit would be the monopolist himself’
What is the reason for these absurdities^ The whole theory of the firm is
a partial equilibrium theory based on cetens panbus assumptions ^ Other
things, prices, incomes, etc arc assumed to be constant, i e not affected
, ,

by any change in the industry under consideration The assumption that

income is constant depends upon the industry being small If the industry n
very large then a change m the level ofproduction will affect factor earnings,
and thus incomes, significantly and so cause a shift in the demand curve Clearly,
1 For an elaboration of this point see pages 316-17
2 See Chapter 37
 THE THEORY OF MONOPOLY 327

we could not use the theory of the last few chapters if the demand curve for
a firm’s product shifted every time the firm’s output changed.
The attempt to consider a perfect monopolist as one who takes all income
founders on the rock that it is no longer a partial equilibrium situation. The
attempt to apply the tools of partial equilibrium analysis to such a situation
can only produce absurd results. The theory of the firm that we have out-
lined here is concerned with productive units which are small in relation to
the total economy. We conclude that the two ideas of a perfect monopoly
this context. We should not worry
considered here are self-contradictory in
about the idea of a perfect monopoly; we should rather think of firms which
are more or less monopolistic in the sense of being able to earn larger or
smaller profits in the short run and being more or less able to protect these
profits in the long run. The measurement of such degrees of monopoly power
has already been discussed in the previous section.
CHAPTER 24

PRICE DISCRIMINATION

Ravv milk IS often sold at one price if it is to go into fluid milk, but at a
lower price if it is to be used to make ice cream or cheese Doctors in private
practice, lawyers and business consultants often charge rates for their services
that vary according to the incomes of their clients Cinemas charge lower
admission pnees for children Railways m many countnes charge different

of products In other countries, such as

rates per ton mile for different kinds
the UK, have passed laws specifically prohibiting
the central authorities
such discriminatory practices Manufacturers often sell their product abroad
at a lower price than at home Electricity authorities ofien sell electricity

more cheaply for industrial use than for home use

Such price differences could never pcnist under perfect competition Yet
many of the examples quoted have existed for decades Persistent price
differences clearly require the cxerasc of some monopoly power, since the
seller is exerting some influence over the price at which his product is sold

Why should a seller want to sell some units of his output at a price well
below the price that he gets for other units ^ Why, m other words, does he
practice price discrimination^
Price discrimination occurs when a producer sells a specific commodity to
different buyers at two or more different prices, for reasons not associated
with differences in cost The theory of price discrimination is sufficiently
*

complicated that a thorough discussion is best left for a more advanced

course ^ But since price discrimination of some sort occurs in a very large

1 Thus quantity discounts difTcrenccs between wholesale and retail prices and prices that
vary with the ume of day or the season of the year ate not generally considered pnee d s
crimination The latter are not usually d senminatory because the same physical product sold
at a different time u a different product if the time of purchase is important to the purchaser
A long distance phone call placed after 8pm
costs less than the same call placed at 3 p
m
But It IS not the same service as any btmnessman knows

2 The interested student will find the topic discussed in most intermediate price theory teals
 PRICE DISCRIMINATION 329

fraction of all markets, often with the approval of public authorities, it is

important to understand the phenomenon.

WHY PRICE DISCRIMINATION PAYS

Why does persistent price discrimination occur? It occurs because different
buyers may be willing to pay different amounts for the same commodity,
and it may be profitable for the seller to take advantage of this willingness.
Some of the issues involved in discriminatory pricing can be seen looking
at the demand curve in Figure 23.1 on page 316 and thinking of it as
describing a market containing a large number of individual buyers, each
of whom wishes to buy one unit, and each of w’hom has indicated the price
he is prepared to pay for it. Suppose a single price, p, is charged. The
quantity OA will be sold because the buyers of each of the OA units are
willing to pay at least p per unit. Although one buyer was willing to pay
only p all of the other buyers were willing to pay more. They have benefited
because the price was limited to p} If the seller could negotiate wdth each
buyer individually, he might be able to charge some buyers more than p.
Infact, by charging each person the maximum that person is prepared to

pay, the seller could greatly increase his profits. To raise the rate of sales
from OA to OA' the seller would have to lower the market price from p to
p'. Marginal revenue is less than the new price charged because of the effect

that the reduction in price has on the revenues from the first OA units, a
reduction that was not necessary to sell them (since they were already being
sold at the higher price). If the seller had been able to sell the amount AA'
at p' without reducing the price on the first OA units (if, in other words, he
had been able to discriminate), he would have profited thereby.

WHEN IS PRICE DISCRIMINATION POSSIBLE?

The conditions under which a seller can succeed in charging discriminatory
prices are, first, that he can control the supply of the product in the sense of
controlling what is offered to a particular buyer, and, second, that he can
prevent the resale of the commodity from one buyer to another. However
much the local butcher would like to charge the banker’s wife twice as much
for a he charges the street sweeper he cannot succeed in doing
lamb chop as
so. always go into the supermarket for her meat where
Madame Banker can
her husband’s occupation is not known. Even if the butcher and the super-

market agreed to charge her twice as much, the banker’s wife could hire the
street sweeper to do her shopping for her. The surgeon, on the other hand,
may succeed in discriminating (if all reputable surgeons will do the same)

1 The area under the demand curve above the line p ( = the area pia) is sometimes called
consumers’ surplus. It represents the amount consumers would have been willing to pay, unit by
unit, for the quantity OA, above the amount they actually paid at the fixed price />.

11*
330 THE INTERMEDIATE THEORY OF SUPPLY
because it will not do the banker much good to hire the street sweeper to
have his operations for him
The first of the two conditions - control over supply - is the feature that
makes price discnmmation an aspect of the theory of monopoly Monopoly
power in some form is necessary to (but not sufficient for) price discrimina-
tion Competition among sellers for customers leads to a single price at (or
near) the level dictated by the costs of production
The second of the two conditions - ability to prevent resale - tends to be
associated with the character of the product or the ability to classify buyers
into readily identifiable groups Services are less easily resold than com-
modities, goods that require installation by the manufacturer (like heavy
equipment) are less easily resold than arc movable commodities (like house-

hold appliances) ’ Transportation costs, tanff barriers or import quotas

serve to separate classes of buyers geographically and may make discrimina-
tion possible
To summarise, price discrimination will be both profitable and possible

where the supplier(s) can control the amount and

distribution of supply,
where the buyers can be separated into classes among which resale is either
impossible or very costly,* and where there are significant differences m the
*
willingness to pay among the distinct classes of buyers

THE POSITIVE EFFECTS OF PRICE DISCRIMINATION

The positive consequences of price discrimination are summarised m the

following two propositions, which we state and try to make intuitively

plausible, but do not prove

1 For any given level of output the best system of

discriminatory prices will provide higher* total
scH
1 An interesting example of nonresalabiliiy occurs in the case of plate glass Small pieces
much more cheaply per square Toot than bigger pieces, but the person who needs a 6 X 10
platewindow cannot use four pieces, each of which is 3 xS
2 The discussion of this paragraph relates rbrcctly to discrimination among classet of ioftrs
Discrimination among units of output follows sirmlar rules Thus the tenth unit purchased by a
fifth
given buyer in a given month can be sold at a different (higher or lower) price than the
unit only if the seller can keep track of who buys what This can be done by the seller of elec
tncity through his meter readings, or by the magazine publuher, who can distinguish between
renewals and new subscnpcions The owner ofa car wash establishment and the manufacturer
of aspirin find it more difficult, although by such devices as coupons or ‘one penny' sales they
too can determine which unit is being purchased
3 ‘Willingness to pay’ is reflected in the demand curves The fact that demand curves slope
downward shows thatsome units could always be seJd at a higher pnee if sellers are permitted
to devitate from a single pnee
4 A more careful statement would say ‘higher than or equal to’ m order to recognise the
ibility that the best single price may be the best system of prices in a given situation
 PRICE DISCRIMINATION 331

revenue to the firm (and thus also higher average

revenue) than the best single price.
This should be intuitively obvious at once. If it is not, you should review
the first pages of this chapter.^

2 Output under monopolistic discrimination will

generally be larger than under single-price monopoly.

To see this quickly, look again at the discussion on pages 315-17.

Marginal revenue will tend to be higher, given the possibility of price dis-
crimination, because the lower price the producer must charge in order to
sell an additional unit will not apply to all The com-
previous units saleable.
mon sense of this is as follows ; the monopolist who must charge a single price
produces less than the perfectly competitive industry, because he
is aware

that by producing and more he drives down the price against him-
selling
self. Price discrimination allows him to avoid this disincentive. To the extent
that he can sell his output in separate blocks, he can sell another block with-

out spoiling the market for the block already being sold. In the case of
perfect price discrimination, where every unit of output is sold at a different
price, his output would be the same as the output of a perfectly competitive
industry.^
The predicted combination of higher average revenue and higher output
does not in have any normative significance. We cannot say that price
itself

discrimination per se better or worse than any other scheme of pricing. It

is

will often lead to a different distribution of output as well as to a different

level of output it will typically lead to a different distribution of income.
;

The particular patterns will depend upon more facts than we have intro-
duced. Having specified the differences, men can debate their desirability.
Economic analysis can predict consequences, but it cannot finally evaluate
them.

THE NORMATIVE ASPECTS OF PRICE DISCRIMINATION

The foregoing discussion may not satisfy the student who is aware that price
discrimination has a bad reputation among economists and lawyers as well

1 We are saying here that the removal of a constraint (the limitation to a single price) upon
a seller will never worsen his situation, and will generally permit him to improve it.

2 If each unit can be sold at a separate price, the seller does nothing to spoil the market for
previous units by selling an additional unit. The marginal revenue of selling an additional unit
isthe price of that unit. Thus, the demand curve becomes the marginal revenue curve, and the
monopolist reaches equilibrium at a point at which the price (in this case, marginal revenue)
equals marginal cost. This is also the point of competitive equilibrium. Under perfect price
discrimination, as under perfect competition, thus, in both cases, profit-maximisation

leads to the output where p = MC.

332 THE INTERMEDIATE THEORY OF SUPPLY

as among laymen The very word discrimination has odious connotations

In the UK, railways have been prevented by law from charging dis
cnminatory prices since the time of their original construction In the US
the Robinson Patman Act makes many kinds of price discrimination illegal
Much of the impetus for railroad regulation in the US came from the out
raged cries of farmers and their organisations that they were being
discriminated against and forced into bankruptcy by the railways who were
not legally prohibited from charging discriminatory prices as were their

British counterparts
Whether or not an individual judges price discrimination to be an evil is
likely to depend upon the details of the case, as well as upon his own

personal value judgments The following examples should serve to illustrate

the varying aspects of price discrimination

Example I A very large oil refiner agrees to ship his product to a market
on a given railway only if the railway gives his company a secret rebate on
the transportation cost and docs not give a similar concession to rival
refiners The railway agrees, and is thus charging discriminatory pnees
This rebate gives the oil company a cost advaniagc that it uses to dnve its

rivals out of business or to force them into a merger on dictated terms

(John D Rockefeller used similar tactics in forming the original Standard
Oil Trust in the US m the late nineteenth century )

Example 2 British Railways are not allowed to discriminate between

passengers in different regions To prevent discrimination, a fixed fare per

passenger mile has been laid down and must be charged on all lines what
ever the density of their passenger traflic and whatever the elasticity of

demand In the interests of economy,

for the services of the particular line
branch lines which cannot cover costs arc often closed down This means
thatsome lines are closed even though the users would prefer rail transport
toany of the available alternatives and the strength of their preference is
such that they would voluntarily pay a price sufficient for the line to yield
a profit The lines are nonetheless closed because it is thought inequitable
to charge the passengers of their line more than the passengers of other
lines

Example 3 When the Aluminium Company of America (alcOa) had a

virtual monopoly on the production of aluminium ingots, it sold both the
raw ingots and fabricated products (such as aluminium cable) made from
the ingots At one time, alcoa sold cable at a pnee 20 per cent below the
pnee it charged for ingots, although of course the cable price was above its
'ost of producing cable It did so because users of cable could substitute
 PRICE DISCRIMINATION 333

copper cable, but many users of ingot had no substitute for aluminium. In
return for its bargain price for cable, alcoa made the purchasers of cable
agree to use it only for transmission purposes. (Without such an agreement,
any demander of aluminium might have bought cable and melted it down.)

Example 4: Doctors in private practice very often charge discriminatoiy^

prices for their seirdces. When they are accused of sharp or immoral
practices in this respect they point out that, if theyhad to charge a uniform
fee for all patients, would have to be so high - if the doctors were to
it

obtain a reasonable income - as to price their services quite out of the reach
of the lower income groups. The discriminatory price system, they argue, is
what allows them to make their services available to all income groups
while still securing an income sufficiently high to ensure a continued supply
of doctors.

Example 5: A product that a number of people want has the cost and
demand structure pictured in Figure 24.1. There is no single price at which
a producing firm can cover total costs. However, if the firm is allowed to
charge discriminatory prices, it will be willing to produce output OA, and
it will make a profit.' (Public utility companies are often said to operate

under these conditions.)

Vi

o
O

Fig 24.1 No single price will cover

total costs but a set of discriminator)'
prices will do so.

Example 6; The government decides to offer primar)' school education to

all children. The cost per child is estimated at £300 per child per year.
Instead of charging tuition to each child’s parents, the government chooses
1 Given perfect discrimination, total revenue would be the shaded area Odc.t, whereas total

costs are OabA. As long as the triangle ade is larger than ebc, there are profits.
334 THE INTERMEDIATE THEORY OF SUPPLY

to the school free and to raise the monc> b> a school tax that »
make
proportional to the value of the houses of the people v\ho live in the com-
munity, whether or not they have children

Each of these examples, as well as those at the beginning of this chapter,

involves price discrimination Few readers would regard them all as equal!)
good or bad The point to be stressed is twofold first, tlie consequencci of
price discrimination can differ m
man) ways from case to case, and, second,
no matter what any individual s values are, he is almost bound to evaluate
the individual cases differently

PRICE DISCRIMINATION SYSTEMATIC

AND UNSYSTEMATIC
Fvery thing that we have discussed in thischipter has been concerned «iib
systematic and persistent price discriminTtion Another sort of price di«

cnmmation is frequently found any seller who occasionally gives a

favourite customer a few pennies off, or shades his price to secure a new
account, is engaged in price discnmination If these practices are used
infrequently they arc called uns\stmati( dismminotm Such discrimination
is not really part of the price structure and we have ignored it in this chapter
This docs not mean that it is unimportant - indeed, unsystematic price

discnmination plavs a v cry real role in the dynamic process by which pnecs
do change in response to changed conditions of supply and demand
Systematic pnee discrimination most often consists of classifying buyers
according to their age, location, industry, income, and so forth, or according
to the use they intend to make of the product, and in charging diffcrcin
prices for the different classes of buyers even though costs do not vary
according to these classifications Such discnmination may also take other
forms, such as charging an individual more for the first unit he buys than
for subsequent units, or ttce ima, or by setting an all or nothing price on a
specified quantity of the product that is different for different quantities of
the product
The causes and consequences of systematic and unsystematic price dn
crimination are very different Control legislation will, howevcr.gencraTly^f
unable to distinguish between them and will hit at both If legislation is
moiiv ued solelv by a desire to attack systematic discrimination, it may have
unforeseen and possibly undestrrd effects on unsystematic discrimination
I Foresample if uniyscemalic pnee dMcnimnalion ii important Tor the working of
pennon, prohibiiing »uch legittanon may aid the mainienanee of monopoly power
CHAPTER 25

COMPETITION AMONG THE

MANY; MONOPOLISTIC
COMPETITION

‘There’s a drugstore on every comer.’ So goes a key line in the play Brother
Rat. This is only a slight exaggeration in many big cities in the United States
and in Britain, where there are more drugstores and chemists shops than are
required to meet the needs of the population. Most of these shops could
handle more business than they get. On the shelves of each are a half-dozen
or more brands of aspirin, each of which is identical to all the others, except
for the label, the package and the price. Most and most highways have
cities

many more petrol stations than are needed to provide all the cars on the

highway with fast and effective ser\ace.‘ Walk into almost any supermarket
and count the number of different brands of toilet paper. Notice that,
although price, packaging and quality differ, each brand is selected by at
least some customers.
These phenomena, as well as the very large role played in the economies
of most Western countries by advertising and ‘salesmanship’, have led many
economists to seek a theoretical model for markets that are characterised by
a large number of firms selling similar but differentiated products, with
much devoted to nonprice competition.^ The theory of mono-
effort
polistic COMPETITION is just such a model. In the first section of this
chapter, we consider the theory; in the second section, we consider two
cases in which its predictions differ considerably from those of the theory of
perfect competition.

1 Evidence presented to a recent Royal Commission in the Canadian Province of British

Columbia suggested that there was 40 per cent excess capacity in the retail petrol industry in
the sense that the same standard of service could have been maintained with 40 per cent fewer
petrol stations.
2 Nonprice competition means competition in product quality, labelling, advertising and so
forth.
336 THE INTERMEDIATE THEORY OF SUPPLY

THE THEORY OF MONOPOLISTIC COMPETITION

Consider an industry m which there is a large number of producers wth
free entry into, and exit from, the industry, but m which each producer sells
a product that is somewhat different from that sold by his competitors The
technical term for this is that the product is differentiated There might,
for example, be a large number of competing firms selling soup The pro
ducts of the firms would be rather similar but by no means identical Each
soup would differ m physical composition from competing products, it
would also have a different packaging, and, as the advertisers say, it would
have a different brand image* from its competitors Industries of this kind
are referred to as being monopolistically competitive or imperfectly
COMPETITIVE The terms monopolistic and imperfect competition describe
a situation similar to perfect competition, with the single important differ

encc that each producer sells a product that is somewhat differenUatcd from
that sold by his competitors
The fact that the product is differentiated between firms means that each

firm does not face a perfectly elastic demand curve for its products We now
construct a demand curve for the firm, showing how much it can sell at

various prices on the assumption that competing firms do not vary their

prices If the firm raises its price, it will lose business to its competitors, but
It will not lose all of its customers just because its prices go slightly above
those of Its competitors The fact that the product is differentiated from
competing products means that some people will prefer it to other products
even though it is somewhat more expensive As prices arc raised above those
of similar products, the firm can expect that fewer and fewer customers will
persist in buying this good On the other hand, if price is lowered below that
charged by competitors the firm can expect to attract customers but every
one will not be attracted by a small pnee differential Thus the firm will be
faced with a downward sloping demand curve for its product Generally
the less differentiated is the product from its competitors, the more elastic

will this curvebe (In the limit if there is no differentiation, the demand
curve will be perfectly elastic because the smallest increase in price above
those of competitors will lose all the firm’s customers while the smallest
decrease in price below those of competiton will attract all the competitors
customers We
arc then back at perfect competition )
Thus we must picture, as m Figure 25 1, a firm faced with a downward
sloping, but rather elastic demand curve, for its product The firm wiH,
course, have the usual U-shaped short run cost curve The short run eqni
librium of the firm is exactly the same as that of a monopolist The firm is
not a passive price taker, it may juggle price and quantity until profits arc
maximised This is at output Oqi and price Opy in the figure
 COMPETITION AMONG THE MANY 337

We may now ask about the long-run equilibrium of the industry. The
firm that we have shown in Figure 25. 1 isearning profits above aU oppor-
tunity costs {=p^uvw) and, if this firm is typical of the others in the industry,
there will be an incentive for new firms to enter the industry. As more firms
enter, the total demand product must be shared out amongst this
for the
larger number of firms so that each can expect to have a smaller share of the
market. Thus at any given price, each firm can expect to sell less than it
could before the influx of new firms. Thus the demand curve for the firm’s
product ivill shift to the left. Clearly this will continue until profits are no

MC

Fig 25.1 The short-run equilibrium

of a firm under monopolistic
competition.

longer being earned, for as long as profits do exist there an attraction for is

new firms to enter, and the industry' will continue to expand.

Before the student reads further he should make a genuine effort to see if
he can discover for himself the final equilibrium position. Start from Figure
25.1 and observe that, as new firms enter the industry, the demand curve, d,

facing our one firm will shift to the left. Also observe that this will continue
until there are no further profits to be earned. WTiat will be the position of
final equilibrium? Assume to begin with, that the demand cur\'e shifts to
the position indicated in Figure 25.2 in w'hich the average revenue curve
intersects the average total cost curve at the point of low'est average cost.

Will do ? Surely if the firm produces output at Oq^ then it will be earn-
this
ing no profits.But if it restricts output below' 0^2 increase average
revenue more than average costs and, hence, it will move into a range of
output at which profits can be earned. In Figure 25.2 costs are just covered
at outputs 0^2 and Og^, while profits are earned at any output betw'een
these levels. We have not yet found an equilibrium in which the profit-
maximising firm will be earning only normal profits.
If w’e are in the situation shown in Figure 25.2 the firm will be producing
338 THE INTERMEDIATE THEORY OF SUPPLY

at some output between O92 and Oqs and will be earning profits so that
expansion of the industry will continue The final position will be that
indicated in Figure 25 3 The average revenue curve touches the aierage
total cost curve at only one point, the point x corresponding to quantity
0^4 and price Op^, the average re\enuc curve is tangential to the a\eragc
cost curve at point x When output is at Oq^ all costs are just being covered
since average revenue equals average total costs At any other level of out
put, losses would exist since average revenue is less than average total cost
Thus we see that a long run, zero profit equilibrium is possible under
conditions of monopolistic competition in spite of the fact that the individual
firm IS faced with a downward sloping demand curve Each firm is forced

into a position in which it is has excess or unused capacity The firm

m Figure 25 3 could expand its output from 0^4 to Oq^ and reduce average
costs, but It does not make use of this productive capacity because doing so
would reduce average revenue by even more than it would reduce average
costs If the demand curve cuts the average total cost curve, as m Figure
25 2, It is always possible to make profits by producing in the range over
 COMPETITION AMONG THE MANY 339

which the demand curve lies above the cost curve. An equilibrium in which
only normal profits are possible requires that the demand curve should just
touch (i.e., be tangential to) the average total cost curve, and this in turn
implies that in equilibrium the firm will have some unused capacity (equal
to q^q 2 in Figure 25.3).

PREDICTIONS OF THE THEORY

In this section we shall consider the two major predictions of the theory of
monopolistic competition which differ from the theory of perfect competi-
tion. Other predictions of the theory which are similar from those of perfect
competition are considered in the following chapter.
The first prediction is as follows:

The equilibrium output of the firm occurs at an output

less than the one at which average total cost is a mini-
mum. (This is known as the excess-capacity theorem.)
Free entry pushes firms to the point at which the demand curve is tangent
to the average total cost curve. But, if the demand curve slopes downward, *

such a tangency must, mathematically, be in the declining portion of the

curve.^
Potentially, this is one of the most important insights in the whole theory
of the firm. It says that a long-run, normal-profit equilibrium can occur even
though each firm is like a monopolist in the sense of having a downward-
sloping demand curve. It will occur because so many firms will enter the
industry that each individual firm will be unable to utilise all of the
capacity at its command. The theory predicts that industries in monopolistic
competition will exhibit a continual tendency toward excess capacity. It also
would have been under perfect
predicts that prices will be higher than they
competition, and that price will be greater than marginal cost. These are
real differences between the theories of perfect competition and monopolistic
competition. If one is interested in making predictions about tendencies
need to be tested in order to establish
for capacity utilisation, the theories
which one of them provides more accurate predictions about the real world.
It is an interesting comment about the debate on this part of modern

economics that although literally dozens of well-known articles both attack-

ing and defending this prediction on a priori grounds exist, there is not a

1 Buyers are supposed to think in such terms as: ‘I prefer Crosse and Blackwell soup ;
‘I

trust Mr Green, even if more expensive’; and ‘Isn’t that the brand the Beatles use?’.
he is a bit
2 The proof of this is quite easy: two curves that are tangent have the same slope at the
point of tangency. But the demand curve slopes dowmward by assumption; therefore, average
total cost must also slope downward.
340 THE INTERMEDIATE THEORY OF SUPPLY
Single well-known article reporting an empirical attempt to test these pre-
dictions (A reason will be suggested soon )
The second prediction is as follows

It may pay the monopolistically competitive firm to

engage in nonpnce competition of a kind that it would
not pay it to use in perfect competition.

To see that this is a prediction of the theory, recall that a firm in perfect
competition can sell as much
wants to at the going price and that it
as it

regards itself as a price-taker Therefore, it docs not pay it to spend money

to increase the amount it can sell But, in monopolistic competition,
expenditures on product differentiation, product quality or advertising can
change the slope or shift to ihc right the demand curve of the seller

Expenditures on advertising and other forms of nonpnce competition may

increase short-run profits We are thus led to the prediction that industries
‘

characterised by the conditions of monopolistic competition will be found

to engage in nonpnce competition, whereas those in perfect competition

will not

THE GREAT DEBATE

A great debate exists between the devotees and critics of the theory of
monopolistic competition The devotees call the theory ‘a revolution of

major importance’ the cntics say ‘In the general case [the theory does not
,

help us to] make a single statement about the economic events m the world
we sought to analyse *
The student may find it hard to understand why such a debate exists

After he will say, the two theories make different predictions Surely,
all,

one or the other must predict more accurately it is a legitimate point Part
of the trouble is that it is difficult to get the two sides to agree on a fair test
Consider a popular example of monopolistic competition the market for
soaps and detergents the well-known brands on sale in the United
Among
States are Ivory, Dash, Joy, Comet, Cascade, Camay, Lava, Duz, Tide,
Cheer, Dreft, Oxydol, Spic ’n’ Span, and Zest Surely this is impressive
1 It is a conaection between the elasticity of
can be shown formally that there a firms

demand curve and maKimising amounts of advertising product differentiation and

the profit
the
quality improvement Thuswill pay the firm to engage in these activities if it can affect
ic

elasticity of the demand curve in certam ways These propositions are proved in R Dorfman

and P O Steiner Optimal Advertising and Optimal Quality Amencan Econorme Rewu!,

December 1954
The first quotation is from Robert L Bishop hfonopohsuc Competition After Thirty
2
Years The Impact on General Theory* Anuncan Econemte Rtvicw May 1964 p 33 The
second IS from George Stigler /ipf Zer/wes
an £f«wmtc iVoWrw (Longmans 1948), p 18
 COMPETITION AMONG THE MANY 341

differentiation - and the fact that most of the names are familiar to most
Americans is impressive evidence of the advertising on their behalf On first
glance, this might appear to be a perfect
example of monopolistic competi-
But, every one of the products named above is manufactured by a single
tion.
company, Procter and Gamble, which alone accounts for more than half
of the American sales of soaps, cleansers and detergents.^ Will Procter and
Gamble really believe that if it lowers its prices its largest rival, Lever
Brothers, will not lower theirs? Does the soap industry exhibit revenues that
do not exceed costs and does it have free entry ? The answer is no to each
question. The fact that the theory of monopolistic competition is consistent
with nonprice competition does not mean that the presence of nonprice
competition is due to monopolistically competitive behaviour.^

Monopolistic competition after 30 years; Monopolistic competition,

as a theory, is more than 30 years old.^ Its survival over so long a period

cannot be explained by the number of tested and confirmed predictions, for

there are few of them.
Whatever may be the outcome of the debate on the predictive value of
monopolistic competition, there can be little doubt that it has contributed

two important things to the development of the subject; at the time that it
was first developed, perfect competition was under severe attack because of
the lack of realism of its assumptions. The theory of monopolistic com-
petition recognised the facts of product differentiation, the ability of firms
to influence prices and advertising. By incorporating all of these new
assumptions into a new theory, economists were encouraged to consider the
question of their effect on the operation of the price system. Also, by show-
ing that such dramatic changes in the assumptions might not affect the
predictions of the theory, economists have become a little more sceptical of
attacks based solely on assertions that the assumptions of their theories are
unrealistic.'^
The second major contribution of the theory is that many economists have
been profoundly influenced by it. It has served to call attention to the

The three-firm concentration ratio exceeds 80 per cent. This illustration comes from
1

J.W. Markham’s article in the American Economic Review, May 1964, p. 54. His point is that the
multi-product firm is not the multi-product industry of monopolistic competition, and that we
require a different theory to explain its behaviour.
2 We have met this point of logic before. In general, the point is that if d is a sufficient

reason for the existence of B, the observation that B exists does not imply that A must exist,

since there may be sufficient reasons for B’s existence that do not include A.
3 basic books on the subject were first
The two published in 1933, one on each
side of the

Atlantic:E.H. Chamberlin, The Theory of Monopolize Competition (Harvard University Press,

1933); and Joan Robinson, The Economics of Imperfect Competition (Macmillan, 1933).
4 See note 1, page 14.
342 THE INTERMEDIATE THEORY OF SUPPLY

potential importance of the extent to which the sellers recognise that they

are interdependent, the extent to which entry ts free, and the nature of the
product These things were freed, by the theory of monopolistic competition,
from their polar models of monopoly and perfect competition This has led
to the development of other important models that we shall discuss in the
next chapter
The major impact of the theory of monopolistic competition has been in
changing the way economists think about problems of market behaviour
This impact has been particularly noticeable in the field of oligopoly
 CHAPTER 26

COMPETITION AMONG THE FEW;

THE THEORY OF OLIGOPOLY

In this chapter we shall consider the behaviour of industries in which there

are only a very few firms. In most Western countries a high proportion
of
manufacturing industries fall within this category. In the British car in-
dustry, for example, 90 per cent of the production is in the hands of five
large firms. The selling of matches in Britain is dominated by only two firms,
one British and one Swedish. Similar patterns are to be found in many other
industries such as the motor-tyre industry, the chemical industry, the
synthetic fibre industryand the electric wire and cable industry. The most
casual observation shows that industries in which there is a small number
of firms are quite common. Such industries are called oligopolies and the
special case inwhich there are only two firms is given the name duopoly.
In order to establish some points of comparison consider the situation
with respect to rivalry in the three market forms we have considered so far.
Monopolists are firms with no close rivals.* Firms in either perfect or mono-
competition recognise that other sellers of the same or very similar
polistic

products exist, but they do not engage in personal rivalry because there are
so many firms it is useless to try to forestall the actions of some of them -
that
the competitors become an impersonal mass which merely constitutes the
‘behaviour of other producers’. When we move to oligopoly, however, the
whole price-output problem of the firm takes on a new dimension that of :

the possible reactions of the firm’s few competitors. The firm’s policy now
depends on how it thinks its competitors will react to its moves, and the
outcome of the firm’s policy depends on how they do in fact react. There is
now no simple set of rules for the equilibrium either of the firm or of the
I All firms are in a sense rivals in competing for consumers’ limited expenditures. When a
monopolist changes his price, however, there are no other sellers to react. Thus, the
monopolist’s ceteris paribus demand curve is his actual demand curve.
344 THE INTERMEDIATE THEORY OF SUPPLY

small group of firms that constitute the industry Neither is there a set of
simple predictions about how the firms will react, either individually or
collectively, to various changes in taxes, costs and demand Everything
depends on the policy that the firm pursues, on the policies that its com
petitors pursue, on how each reacts to the other’s changes and on how each
thmks the other will react
It IS often said that, under these circumstances, price and output is

indeterminate Such a statement is misleading (the price and output do of

course, get determined somehow), and what is meant by the statement is
that, under oligopoly, price and output arc not determined by the same
factors as in monopoly or in large*group cases In small group cases an
additional set of factors - competitors* real and imagined reactions to each
other’s behaviour — contributes to the determination of price and output

THEORIES OF OLIGOPOLY
In view of the complexity of the oligopoly problem it is not surpnsmg that
there is no single, welUdevcloped theory of oligopolistic behaviour In fact
there are two quite different attacks being made on the oligopoly problem
and we must give some mention to each

The Development of Theoretical Models

One attack on the problem is a senes of theoretical models by
to develop
assuming that individual firms react in particular ways, and then see what
follows from these assumptions Economists have done a good bit of this
from 1838 to the present ‘ The trouble with this approach is that the
number of actions, reactions and interactions is so many that there can be
no hope of an exhaustive analysis of all possible cases and it would take

centuries to test the predictions that arise from the ver> large number of
^
conceivable cases
The student who goes on to a further study of economics should probably
read one of the standard textbooks on oligopoly theory This will give him
an impression of the complexity of the problem But if, when he finishes his
study, he asks himself what he can predict about oligopolistic behaviour in
the real world he will probably answer ‘httle’ or ‘nothing at all’ Until there

! Cournot m 1838 made the fii*l known attack on the duopoly (two-seilcrs) problem
A A
He had each seller always assume that his nval would not change his price The rival always
did of course (This sort of quaintness pervades the early attempts to deal with oligopoly
2 For a review of the literature and some of the special cases the student who goes on m
economics should consult Fcllner GuRpeftboa damtg Market
At Few Oligopoly and Similar
Slruclures New York Knopf 1949 Chapters II and III
 :

COMPETITION AMONG THE FEW 345

is empirical knowledge of how firms actually do react in small-group situa-

tions sufficient to narrow drastically the range of possible cases, there can
be little hope of obtaining propositions about oligopolistic behaviour which
are of real use in positive economics.
The modern development of the Theory of Games, which is a study of
rational strategies in small-group situations, gives promise of providing an
analytical structure suitable for the handling of these problems.^ But an
analytical technique only as useful as the real-world information that it is
is

used to analyse. Even the most powerful new techniques \rill be empty
without empirical knowledge of how firms behave in typical small-group
situations.

Piecemeal Generalising from Hypotheses about Observed Behaviour

The second major attack on the problem of understanding oligopolistic
behaviour is an attempt to build up a theory piecemeal by developing hypo-
theses to explain actual bits of obseir^ed behaviour, in the hope that these
piecemeal explanations ivill eventually shape into a general theors' which,
because it began with actual observations, will be successful in explaining
and predicting the oligopolistic behaviour that actually occurs in the world.
The theory of oligopoly is in transition; economists have rejected a number
of very simple models and are searching for the elements that will produce
a more complex theory. Although we cannot enter into a detailed discussion
of these elements in this book, we can mention a few of the specific pre-
dictions that have been suggested in order to convey some of the flavour of
these modern developments.

The HYPOTHESIS OF QUALIFIED JOINT PROFIT-MAXIMISATION: The hypo-

thesis of qualified joint profit-maximisation may be stated as follows

Firms that recognise that they are in rivalry with one

another will be motivated by two opposing forces, one
moving them toward a set of policies that maximise
the combined profits of the existing group of sellers.
(These are called joint profits.) The other set of tend-
encies move them away from the joint profit-
maximising position. Both sets of tendencies are
associated with observable characteristics of firms,
markets and products, and thus we can make pre-
dictions about market behaviour on the basis of these
characteristics.
For an entertaining general introduction see
1 J. D. IVilliams, The Compleat Strategyst,

McGraw-Hill. New York, 1954.

346 THE INTERMEDIATE THEORY OF SUPPLY

This hypothesis explicitly allows for the fact that the actions of the n\als
affect the size of the ‘pie’ as well as its division among them This is, of
course, not surprising, since the group must face a downward sloping
demand curve and, unless the clastiaty of this market curve just happens
to be unity, total revenue earned by the group will vary as their collectne
pricing policies vary If the firms behave as a single firm, they can act as a
monopolist and adopt the policy that will maximise their joint profits If
the firms depart from monopoly behaviour, they will reduce their joint
profits It may pay one firm to depart from the joint profit-maxinusing
But
position by doing so, it can increase its share of the profits If a firm
if,

adopts such a strategy in order to raise its share of the profits, it must
balance what it expects to gain by securing a larger share against what it
expects to lose because there will be a smaller total to go around among alt
firms The hypothesis says that there are forces operating on the individual
firm that lead it to alter its behaviour toward the joint profit-maxinusmg
position and other forces that lead it to alter its behaviour away from the

joint profit-maximismg position To make the hypothesis testable, wc must

specify which push the firm
forces m
which direction ^
We may now list a few of the hypotheses that have been advanced as
explaining and predicting the strengths of the tendencies toward and away
from joint profit-maximising

1 THE INDUSTRY WILL TEND TO BE CLOSER TO THE JOINT PROFIT'

MAXIMISING POSITION, THE GREATER THE DEGREE OF MUTUAL RECOCNI
TiON OF INTERDEPENDENCE Mutual recognition of interdependence will
tend to be greater in each of the following circumstances
(i) the smaller the number of sellers,
(ii) the more nearly equal the sellers are in market shares and in methods

of production,
(ill) the more nearly identical the products of the sellers

2 The industry will tend to be closer to the joint profit-

maximising POSITION, THE EASIER IT IS FOR FIRMS TO REACH TACIT
AGREEMENT^ The ability to reach and abide by tacit agreements will be
greater in eadn Of fne lo’i'iowing arcumstances
or
(i) in a market in which the price that maximises joint profits is stable

u
1 Notice that if the hypothesis were correct it would explain why the theory of monopoly
not sufficient to predict the market behaviour of oligopolies Monopoly theory predicts full
joint profitmaximising and jt cannot predict the extent of departures from it
policies
2 Tacit agreement is one term for the process by which firms may come to common
without explicit discussion Other terms for the same process are tacit collusion qu**'

agreement, conscious parallel action and impbat

coordination
 ; ;

COMPETITION AMONG THE FEW 347

rising rather than falling (since

it is easier to get orderly agreements on

maintaining or raising prices than on reducing them)

(ii) the less the degree of uncertainty attached to the firm’s
estimates of
future demands, costs and other relevant factors;
(iii) in an industry with a dominant firm rather than in an industry
\rithout one;
(iv) among firms wdth similar expectations of the future than among
firms with widely differing expectations.

3 The industry will tend to be closer to the joint profit-

maximising POSITION, the greater THE BARRIERS TO ENTRY; This is
because strong barriers to entry give the existing firms more scope to exploit
their monopohstic position. Such barriers %vill tend to be greater in each of
the following circumstances:
(i) the greater the economies of scale;
(ii) the greater the degree of product differentiation indulged in by
existing firms
(iii) the greater the brand-name advertising.
This problem of barriers to entry an extremely interesting and complex
is

one. The first a technological one in the sense that

barrier listed above is

it is created by the state of technolog)'. Economies of scale mean that new

firms, which inevitably begin with only a small share of the market, will
have high costs. They will thus find it hard to compete with large established
firms which will have low costs because they are large enough to exploit
existing economies of scale. This very important point is illustrated in

Figure 26. 1 where the A TC curve shows the long-run, average costs of a
single firm subject to economies of scale.*
The established firm has a demand curve and it is able to earn profits
above all its opportunity costs. If a new firm enters the industrv' with a
demand curve D„i there which the firm can cover its costs.
is no output at
If it takes time for the firm to establish itself in the market and to have the
demand for its product build up to a higher level, the firm must accept
losses. These losses will continue until its demand has expanded to D„2 at

which time it can cover all its costs by producing at the output correspond-
ing to the point of tangency betrveen D„2 and A TC. Figure 26. 1 (ii) differs
from Figure 26.1(i) in that the cost curve shows that quite a small level of

output is sufficient to exploit all the significant economies of scale.

The
curve Z)„j for the new entrant is shown in exactly the same position as in

Figure 26.1 (i) but this time the firm can make profits even though it has a
ver)’ small share of the market. Clearly, entry is much easier in an industry

1 If you are not sure about this important concept of economies of scale, you should review
pages 269-70.
 .

Competition among the few 349

with costs like those in Figure 26.1 (ii) than those in 26.1(i). This is the
meaning of point (i) above.
If all firms in an industry' have costs like (ii), is there anything that they
can do to increase the barriers to entr)' so that they can move in the
direction of joint-profit maximisation without fear of an immediate rush of
new firms into the industry.^ There are two possibilities. First, if the product
is one in which consumers switch brands frequently, then increasing the
number of brands sold by existing firms will reduce the expected sales of a
new entrant (thus keeping Z)„i as far over to the left as is possible). Say, for
example, that an industry contains three large firms each selling one brand
of cigarettes, and say that 30 per cent of the users of any brand give up their
brand each year to switch in a random fashion to other existing brands.
This means that 30 per cent of all customers will be changing brands each
year. If a new firm enters the industry, it can expect to pick up 25 per cent
of these switches (it has one brand out of a total of four available brands)
This gives it 1\ per cent of the total market just as a result of picking up the
custom of random brand switchers. If, however, the existing three firms had
five brands each, then there would be 15 brands already available and a

new small firm selling one new brand could only expect to pick up one
sixteenth of the brand switchers, giving it only 1-9 per cent of the total
market on this account.
This discussion may be summarised as follows. Where brand switching is
important, there will be a large floating population of brand switchers who
constitute potential customers for a new firm. The greater the number of
brands sold by existing firms the more diffused will be the efifect of this
brand switching and the less will be the percentage of the brand switchers
gained by the single brand of a new firm. The proliferation of brands by
existing firms is thus seen to constitute a defensive reaction making it harder
for a new firm to pick up customers who are switching brands. This explains
point (ii) above.
Finally we come second defensive policy which is commonly used
to a
when technological barriers to entry are weak. This policy attempts to shift
the cost curve to the rightby creating advertising costs. If there is a vast
amount of brand-image advertising, then a new firm will have to spend a
great deal on advertising its product in order to bring it to the public s
attention. If the firm’s sales are small, advertising costs per unit sold will be
very large. Only when sales are large, so that the advertising costs can be
spread over a large number of units, will costs per unit be brought down to
a reasonable level. Thus, heavy advertising expenditures in an industry
without economies of scale in production have the effect of changing total
cost curves from the general slope illustrated in Figure 26.1 (ii) to that
illustrated in Figure 26.1(i).
350 THE INTERMEDIATE THEORY OF SUPPLY

This proposition is illustrated in Figure 26 2 The curve labelled ATC

(Production) copied from Figure 26 l{u) and shows a case m svhich
is

economies of scale arc all exhausted at a quite modest level of output, with
constant long run average costs beyond that level Such an industry is, as
we have already seen, easy to enter We
now add a fixed level of advertising
costs necessary to establish a new brand against the heavy advertising of
existing brands When we divide this fixed cost by the number of units pro
duced we obtain the curve (Advertising) which shows how advertising
cost per unit sold declines as output is raised, thus spreading the fixed cost
over more and more units ‘ If we add these two curves together we obtain
the curve A7X^ (Production plus Advertising) which shows how all costs

Ftg 26 2 Average total cost curves

allowing for both production and
costs

vary as output vanes The curves are drawn to scale and it is clear from
inspection that the new cost curve doesn’t flatten out until a much higher
output is reached, compared with the output at which the curve without
advertising flattens outA small new entrant will be at a substantial cost
disadvantage over a large established firm This explains point (iii) above
The discussion may now be summarised as follows advertising costs
necessary to establish a brand image increase the level of output at which
the long run total cost curve Battens out, and thus makes it more difficult

for a small new firm to operate profitably Advertising thus creates a bamcr
to entry
The above hypotheses about barriers to entry which are created by exist

1 This curve IS a rectangular hyperbola with thcforimila where A' u the

advertising cost and Q the number of units produced and sold

 COMPETITION AMONG THE FEW 351

ing firms help to explain two apparently paradoxical aspects of everyday

industrial life : the fact that one firm sells many different brands of the same

product, and that each firm spends much money on advertising, competing
not only against the products produced b\ rix.il firms 1ml also against other
products which arc produced by the same firm. The soap and cigarette
industries provide classic examples of this behaviour. In both of these in-
dustries there is only a small number of firms but there is a very large
number of only slightly differentiated products. The explanation is that
technological barriers to entry are weak in these industries - a small plant
can produce at an ATC which is just as low as that of a large plant. Product
differentiation and brand-image advertising create substantial barriers
where technological ones are weak, and they thus allow existing firms to
move in the direction of joint-profit maximisation without fear of a flood of
new entrants attracted by the high profits.

Of course, there is a great deal more to the theory of oligopoly than can be
summarised here, but enough should have been show that the council
said to
of despair ‘you will never explain something so complex as small group
behaviour’ should be rejected. The assertion that something is impossible
is a powerful challenge to the creative mind. Enough has already been
established in this field to show not only that we can already explain and
predict some parts of oligopoly behaviour, but also that we should have
substantial confidence that 20 years from now we shall be able to explain
and predict much more. There can be no doubt, however, that even in our
introductory treatment of this subject we have come ver^' close to one of
the frontiers of modern economics.
CHAPTER 27

SOME PREDICTIONS OF THE

THEORIES OF COMPETITION
AND MONOPOLY
In this chapter we sliall use comparative static analysis* to discover certain
implications of the theories of competition and monopoly We shall devote
our attention to the two extremes of perfect competition and monopoly
and we shall say very little about monopolistic competition and oligopoly
Wc omit monopolistic competition because, unless we have much more
knowledge than we are assuming at present, the theory gives no definite
predictions concerning the kinds of changes in which we are interested
This is also true of the theory of oligopoly and if we had a more specific
theory of particular oligopolistic reactions it would necessarily be very com
plex and therefore the denvations of predictions even from a single oligo-
polistic model IS not an easy task Nonetheless, the study of firms in other
market situations may help us to gam some idea of how firms might behave
m some oligopolistic situations
The shall derive in this chapter can be viewed m
propositions that we
two different waysthey arc implications of the theories already
First,
developed From this point of view vve are engaged in a purely logical pro
ccss of finding out what propositions are implied by our assumptions This
IS a simple matter of right or wrong either a certain proposition is, or is
not, implied by our theory This is a matter on which vve can come to a
perfectly definite conclusion and the student must accept that the prob-
abfkty of emjrv m
Ajgnr sii!! renrannng- m
fftis wnV-wufievf .y jwm.I'

If we wish, however, to see how well our theories fit the real world, if we
wish that IS, to test our theories, then the propositions developed in this
chapter must be taken as hypotheses about the world Whether or not these
hypotheses are consistent with the facts or are refuted by them is a matter for
to
testing and, in the absence of empirical evidence, it is not necessary
accept these propositions as statements about what happens in the world
1 See page 153 if you have forgotten the precise meaning of this term
 THEORIES OF COMPETITION AND MONOPOLY 353

To summarise, we may say that the question whether or not a given

proposition is implied by, or follows from, some theory, is a purely logical
question that can be settled definitely without reference to facts; but the
question whether or not a given proposition {which follows from a theory)
fits the facts or is refuted by them is an empirical question that can only be
settled by an appeal to real-world observations. In this chapter we are
concerned solely with the former type of question; in subsequent chapters
we shall consider questions of the latter type.
We shall begin by considering the response of various industries to shifts

in the demand for their product.

A RISE IN THE DEMAND FOR THE PRODUCT OF A

COMPETITIVE INDUSTRY
Figure 27.1 shows the cost and demand conditions for a single firm (i) and
for a whole industry (ii), under perfect competition. When the demand

(i) Equilibrium of a firm in perfect competition.

(ii) Equilibrium of a perfectly competitive industry.
Fig 27.1

curve D^, both the individual firm and the industry are in equilibrium
is

at price Op^. (If you have any doubts about this, you should review
incentive for any firm to change its output, nor
is
Chapter There is
22.) no
there incentive for entry or exit of firms.
Now assume that the market demand curve in Figure 27.1 shifts from
Di to D 2 This causes a shortage of the product to develop. This shortage
.

12
 ?

354 THE INTERMEDIATE THEORY OF SUPPLY

causes the price to rise, and in response to this, existing firms increase their
output In the short run, the market price rises to OPi, at which pace the
total industry supply is equal to the total demand for the product At price

Opi the individual firm that we arc considering will produce Oqi units and
the total production of all firms will be OQj Our firm will be making profits
on each unit equal to the difference between the price per unit and the
average cost per unit — 2 t=Ai) Thus the firm’s total profit is equal
to the area of the rectangle htjp2
These then are the predictions concerning the short run effects of a rise

in demand
1 Price will rise
2 There will be an increase in the quantity supplied
by each firm and hence by the industry
3 Each firm will now be earning profits over oppor-
tunity costs

(i) Constant long run costs

(ii) Falling long run costs
(ill) Rising long run costs
F,g272
The long run effects follow from prediction 3 Since firms arc now
making profits, this industry becomes an attractive one in which to invest
New firms will enter the industry, and existing firms will tend to expand
Thus, there will be an increase in supply that will tend to bring the price
below the previously established, short run equilibrium level pi Supply
will increase and price will fall until the profits arc eliminated and there
ISno longer any incentive for new firms to enter the industry or for exist
mg firms to expand their capaaty Whether the long run equilibnum
price IS above, equal to, or below the original price will depend upon
the shape of the long run supply curve Figure 27 2 illustrates the three
possibilities
In each case, is the original demand curve and pi is the equilibf'nm
its
price at which each existing firm in the competitive industry is covering
 ;

THEORIES OF COMPETITION AND MONOPOLY 355

cost (see Figure 27.1 on page 353 above). SRSi represents the short-run
supply curve. In response to the shift in the demand curve to
2 short-run D ,

price rises to p2 In response

. to the profits earned at p2 new firms enter the
,

industry and the supply curve shifts to the right. The new equilibrium price
is
p2 LRS is the long-run supply curve.
.

We must now ask why it is that the equilibrium price might change when
the scale of the industry expands. There are two possibilities. Suppose that
the industry incurs increasing costs as it expands. Cost cuiY'es rise as the
industry grows* and, were price to fall to its original level, there would be
losses. The expansion of the industry under the incentive of profits will stop
before price returns to its original level. The long-run supply curv'e of such
industry will be upward-sloping.
Now suppose that long-run costs are falling. When the industry has
expanded sufficiently to drive price down to its original level, costs will have
fallen and so there wdll still be a gap between average revenue and average
cost. The existence of profits will still cause entry into the industr)' which
will continue until price has been driven down to the level of average total
cost. The long-run supply curve of such an industry will be downward-
sloping.^
Thus the long-run consequences of a rise in demand are as follows^

1 The scale of industry will expand.

2 Profits will eventually return to zero.
3 Price will be above, below or equal to its original
level according as the industry is one of increasing,
decreasing or constant costs. ,

A FALL IN THE DEMAND FOR THE PRODUCT OF A

COMPETITIVE INDUSTRY
Now consider the effects of a fall in demand. Figure 27.3 shows the firm and
the industry in long-run equilibrium before the fall in demand. When
demand falls a glut of the commodity is created ; supply exceeds demand at

the original price Opi ;

price then falls to Op^- At price Op^, the individual
firm was just covering all opportunity costs. At the new price, Op^, the firm
will produce Oq^, the output for which marginal cost equals marginal
revenue.
The firm will, however, now be making losses. Average total costs are

1 The firms in the industry display long-run costs similar to those of firm in Figure 20.2 (ii)
on page 269.
2* Advanced students svill realise that such economies must be external to the industry.
3 These are predictions. We do not know if they are empirically correct until we have tested
them. The student must beware of this sort of terminology. When he is told that something is
true he must be clear whether we mean a logically correct deduction from a theory or true
empirically.
356 THE INTERMEDIATE THEORY OF SUPPLY
and a\erage re\enue is q^k = 0^ 3
) so that losses per unit are kj The
(

question immediate!) anses would

not be worth the firm’s while to close
it

down its operations altogether instead of running at a loss’ Tlie answer to

this question prosides an extremely important proposition m the theor) of
the firm

If we divide the firm*s costs into fixed costs, which

must be paid even if nothing is produced at all, and
variable costs, which must be paid only if the firm
actually produces commodities, then it pays the firm
to continue production as long as total receipts exceed
total variable costs.

(
11 ) The equilibrium of a perfectly competitive industry

F.e273
Since fixed costs must be paid whether or not the firm produces, they do
not affect the decision whether or not to produce Variable costs can, how-
e\ er, be avoided by the expedient of stopping production, and it thus pays

to continue production only if receipts cover variable costs As long as

receipts are even slightly greater than variable costs, then they can make at

least some contnbutton towards fixed costs and total losses will be less than
they would be if the Firm ceased production
Tins important point may be clarified by a numerical example Consider
a firm with fixed costs of ;(^10,000 per annum Assume that, at the most
profitable level of output, variable costs arc ^^5,000 Assume total receipts
exceed £5,000, say they arc £6,000, then u pays the firm to continue m
production, for total losses of £9,000 arc less than they would be if pro*
duction ceased, in winch case they would be £10,000 If receipts were to fall

below variable costs, say to £4,000, then it would pay the firm to cease
 ,

THEORIES OF COMPETITION AND MONOPOLY 357

production, for total losses would then only’ be ^^1 0,000 as opposed to losses
of j(^l 1,000 if production were to continue.
We conclude, therefore, that if price falls below average total cost, but
exceeds average variable cost, then the firm wll make losses, but that it will
stay in production at least in the short run. Only if price falls below the
average variable cost, -ivill the firm cease production.
Thus the short-run effects of a fall in demand in a competitive industry’
are as follows:

1 a fall in price;
2 a
fall in quantity produced and sold;
3 the existence of losses;
4 firms will go out of production immediately if they
are unable to cover their variable cost of production.

The long-run effects follow from point (3). Since firms in the industry'
are not covering all costs, the industry is not an attractive place in ^vhich to
invest. No new investment funds will enter the industry; as old plant and
equipment wears out, it u’ill not pay to replace it (because the expected
profits on the funds used for this equipment will be negative). Thus the
scale of the industry will contract. The short-run supply curv'e, showing how
production varies with price, capital equipment being held constant, shifts
to the left as the capital in the industry' wears out and is not replaced. But,
as the supply diminishes, the price of the product begins to rise. This price
brought about by a contraction of output in the industry, will continue
rise,

remaining in the industry can just cover their total costs.

until the firms
We may now repeat the argument of pages 354—5 to show what finally
happens to price under various long-run cost conditions. If long-run costs
are constant, then price must rise to the level which obtained before the
demand shift occurred. This is easily’ seen: constant costs means that the
contraction in the scale of the industry leaves the costs unaffected. There-
fore, as long as price remains below its original level {Op^ in Figure 27.1),
losses will be the and the industry u-ill continue to contract. Should
rule,

price rise above Opi profits ivill occur and expansion will take place.
If the industry’ is a falling-cost industry, then the contraction of scale will
raise costs. Thus price must rise above its original level before the con-
traction ceases. (When price returns to its original level Opi, costs will have
risen, so losses will still occur. If the industry’ is a rising-cost industry' then

the fall in scale will low'er costs. Thus price will finally settle at a le\’el belorv

its original one. If contraction continued until price had risen to Opi then,

costs having fallen, profits would be earned. All of this is illustrated in

Figure 27.2 in which SRS2 is the initial short-run supply curve, SRS^ is the
358 THE INTERMEDIATE THEORY OF SUPPLY
new short run curve after scale has been contracted and LRS is the long
run curve
We have established the following long run implications of a fall in

demand
1 the quantity produced and sold will fall and the
whole scale of the industry will contract,
2 losses will eventually disappear and all costs will
be covered,
3 price will be above, below, or the same as the price
ruling before the demand shift occurred according as
long>run costs are falbng, rising or constant

A CHANGE IN THE DEMAND FOR A MONOPOLISTS

PRODUCT
We now come to the question of the effect on a monopolist s pnee and out

put of a shift m the demand for his product We shall consider the case of

Fig 271 h nsc m demand

associated with a ma
fall

monopolist s price

an increase in demand and the anafysu can easify be applied to tke c^se of
a decrease In general we cannot predict that a rise m
demand will cause
an increase in either price or in output * It is possible providing the elasticity
of the demand curve changes sufliciently for a rise in demand to cause a fall
either in price or in production Figure 27 4 illustrates the case which am
rise in demand causes a fall in price The demand curve shifts from Di to
1 Th i follows immed ately from the discuas on in Chapter 23 on page 320 on the absence
ofa supply curve under monopoly
 THEORIES OF COMPETITION AND MONOPOLY 359

D2 ,
the quantity produced and sold increases from Oq^ to 0 q2 but price falls
from 0/)j to Op2 The reason
. for this is that, at all prices at or above the
original price of Opi, the new demand curve is veiy' much more elastic
than the old curv'e so that it pays the monopolist to increase his output to a
point at which the final price below the original one. We conclude that a
is

rise in demand combined with a sufficient increase in elasticity can result in

a fall in the monopolisms price.
Figure 27.5 illustrates a case in which a rise in demand causes both the
price and output of a monopolist to increase. It is easily proved %rith
elementary mathematics that any rightsvard shift of a straight-line demand
cuiv'e causes both price and output to rise {assuming an upward-sloping
MC cuivt).
.A.t this level of generality we are thus left \rith the conclusion that a rise
in demand for a monopolist can cause both his price and output to rise.

Fig 27.5 A rise in demand

associated \vith a rise in both the
monopolist’s price and output.

but that it is possible for either his price or his output to fall. For many
purposes in applying the theon- of the firm it is important to have some more
definite prediction about the effect demand on a monopolist’s
of a change in
price-output policy and we may
mention hosv one might proceed to
briefly
investigate this problem further. On the one hand, one might proceed \rith
a fuller investigation of the theoretical model. Mathematical tools would be
necessary and one ivould seek to determine precisely ivhat changes in
demand elasticity- would have to accompany a rightward shift in the demand
cun^e if either price or output were to fall. Having carefully tvorked this

out, one might then appeal to empirical eridence for particular commodities
to see if it appeared at all likely that such combinations of shifts in cur\-es
and changes in elasticity might occur. On the other hand, one might go
directly to empirical evidence and see what price and output changes have
in fact accompanied changes in demand. If it were found, for e.xample, that
360 THE INTERMEDIATE THEORY OF SUPPLY

increases in demand were always accompanied by increases in both price

and output, this could then be taken to be the normal real world case with
the other cases as possible, but unbkely, exceptions

ELASTICITY OF DEMAND AT EQUILIBRIUM

In a competitive industry, equilibrium occurs where the industry supply
curve (which is the sum of the marginal cost curves of the firms that com-
pose the industry) cuts the demand curve The elasticity of the market
demand curve at the equilibrium price may be anything from zero to
infinity ‘
In the case of monopoly, however, equilibrium occurs at the level

of output at which the marginal cost curve cuts the marginal revenue curve
Since marginal costs are always greater than zero, it follows that marginal
revenue must also be greater than zero at the equilibrium point But if

marginal revenue is greater than zero, it implies that the elasticity of

demand is greater than one * TTius, a profit-maximismg monopolist will

always move to some point on his demand curve at which elasticity exceeds
one, if he finds himself momentarily at a point at which demand is inelastic,
It will obviously pay him to reduce output, thus simultaneously increasing

his total receipts and reducing his total costs, hence increasing his total

profits Thus, while it is quite possible for a competitive industry to be in

equilibrium in a position in which a reduction m output would result in

an increase in consumers’ expenditure on the product, such a situation is
not possible for a profit-maximising monopolist

Producers Cooperatives
We sec that the theory of competitive price predicts that the market will
come into equilibrium at the pioint at which the quantity demanded equals
is no restriction whatever on
the quantity supplied, and that there the

elasticity of demand at the point of equilibrium

We may now ask what the theory predicts about the possible scope for
increasing profits m a competitive market that is m
equilibnum Note first
that, in competition, individual farmers will act as price-takers, their
behaviour be the same whatever the elasticity of market demand at the
will
point of equilibrium This is because there is no point in any one of them
cutting back on his output, since any one of them faces a perfectly clastic
demand for his own product If he did reduce his output, he would leave
1 This IS also the case m manopoluuc competitum
by a
2 If the elasticity of demand is less than one a I per cent increase in sales will he met

fall in the price of more than 1 per cent hence lotal receipts will fall and marginal revenue
(the change in total revenue resulung from selling a little rnore) will be negative
 !

THEORIES OF COMPETITION AND MONOPOLY 361

price unaffected and so would merely reduce his income as his output fell.
The first prediction of the theory is as follows.

Producers in a perfectly competitive industry can

always increase profits above those in competitive
equilibrium if they all agree simultaneously to restrict
output by a quota system.

The common sense behind this prediction in the case of inelastic demands
is that if all farmers form a producers’ co-op and each agrees to cut pro-
duction by establishing quotas for every member’s output, then the total
supply will be affected, price will rise by a larger percentage than the fall in
output, and all the farmers will be better off. Since their revenue is up and
their costs are down, because they are producing less, their profits mu&t be
up.' The theory' has, however, a second very important prediction.

Once output has been restricted by the Co-op, each

single producer will have an incentive to increase his
output and the producers’ co-op wiU, therefore, tend
tp exhibit unstable behaviour.

The common sense behind this prediction is that the co-op raises prices by
cutting production but if any one farmer could raise his output, his own
actions would not affect the price, and his income would rise, since he could
go back to selling his pre-co-op output at the post-co-op prices. Thus, unless
the co-op very carefully policed and has the power to enforce its quota
is

restrictions on everyone’s output, there will be a tendency for members to

begin to violate quotas once prices have been raised. Furthermore, the
co-op must have power over all producers, not merely over its members;
otherwise a producer can avoid the quota restriction merely by leaving the
co-op.
Figure 27.6(i) represents the market conditions of supply and demand
for some agricultural product that is produced under conditions of perfect
competition.
Figure 27.6(ii) represents the conditions of demand and cost for an
individual farmer. Before the co-op is formed, let us suppose that the market

is in competitive equilibrium at price pji and output OA and the individual

farmer is producing output Oa, and just covering costs. The co-op is formed

1 We have concentrated on the case of inelastic demand because here the gains from
cooperative (monopolistic) action are obvious and the farmer’s grievances most vocal. The
sections on pages 368-70 shows, however, that such cooperative action can always improve the
farmer’s position whatever the elasticity of demand

12 *
 y>.' Ttir <srrKU»t>iATr Titronv nr irfriY

111 nnJfr I'j ollim pnvjufrM l»r\r»r 4 n*nfK»j»ititin* m’liKiKrrn tJn^nuil't

ll) )> f u» ft l»* mliKf >ii» ij toJum fuiln wjijsijt

to (H! ulirtr nurviiul mrmjr i» njiul in mttettia! f»«» vi ihtl |}f tr;'j|

pinJin »>f all ilif pn-ilm ru » I ilif *tr miYifnxrt! Pir ptur al i* u
ir\rl (4 I*
f'p
n»f imlniiliial faitnft in | iruir 'J? (1 m lui 4
« f 0^ Imt rvrii |}h>iic)s lif l»a* fnliKrtt I n {><• hit injpfmril hs
|«Hiliofi hr n liov» rjtl m * in tin* BtnntjiH «!»mn In tlx- aira rn ll<-

If* will'll tlif linrt u>f In ihr Irfi Tlin «|fn*rif nialfi fnir fint f-tr

tltrix'ii

ISf r) f*rnirr vnll Ik* iM-tirrofrir • U forni«J •nd

«uccr^<l« {n r«(«mR price t« f« tti«n if no civop it
fortnrtl •nd price remaint *1 ^4.

I lihniini t‘( » cftm'wliinr irultnin wi«!i jirwl *iI*!>«mI a ru'fefainf

ciP{» rr*inr«Kjn Klirfftf
II » f 4 lifin IB • »i’iiHj*cmne iniluiirY » tiB'In unrr^irKifl
miiijif iilinn li hiiIi a | i«lj| rnttimniBif tii>p'»n«n<i'''0 K^mr
t wlifti < nl) ilm fmn twJain tu qxH ><4
/.f .VA

Hill, onre fincc li rancfl 10 /*j, iheimiiMilii »l Tirmrc wmild lil^c ii’

lin ouiptii l)rr.imr prirr it Rtealrr iHm nnnjiiul f««t left 10 liiimctf. hf
wituld lilkC lo memte iiit otiipui lo Of» .ind earn the prolltt iliowii
I’lRiirc 27 »« 11) In the .irct in which llie lm« h»e m ihe nsliL Tlin profit o
 :

THEORIES OF COMPETITION AND MONOPOLY 363

necessarily greater than the profit he earns by producing Ob} This leads to
our second prediction

When the co-op has maximised the total profits of its

members by restricting their output by quotas, every
co-op member will be in a position in which he can
increase his profits by violating his output quota, pro-
viding the other members do not violate theirs.

These two predictions highlight the dilemma of the attempts of pro-

ducers’ co-ops to raise theirmember’s incomes; every member is better off
if the co-op is formed and is effective, but he is even better off if everyone

else plays ball, but he does not, yet if everyone cheats (or stays out of the

co-op), all will be worse off. Thus we are led to the following summary
prediction.

Producers’ co-ops formed to raise incomes above their

competitive levels will be able to raise producers’
incomes, provided they are able strictly to enforce
quotas on the outputs of all producers. Such co-ops
will, however, exhibit unstable tendencies, for it will
always be in the interest of any single member to
raise his output. If many producers do so, the co-op
will collapse and all producers will lose.

The history of schemes to raise farm incomes by limiting crops bears

ample testimony to the empirical applicability of this prediction. Crop
restriction often breaks down and prices then fall as individuals exceed their
quotas. The and occasional violence that is sometimes
great bitterness
exhibited by members of crop-restriction schemes against nonmembers (or
members who cheat) is readily understandable.

CHANGES IN COSTS
We may now very briefly consider the case of cost changes both in perfect
competition and in monopoly. We shall find that in both cases a fall in costs
leads to a rise in output and a fall in price. Consider the case of an innova-
tion which lowers the average variable cost of production by X per cent.

Such an innovation reduces marginal costs since it now costs less than it did
previously to produce any unit of output. Thus each firm’s marginal cost
curve shifts downwards and the industry supply curve, which is the sum
of the marginal cost curves, also shifts downwards: as a result of the

1 The reason is that at Ob, price is greater than marginal cost.

364 THE INTERMEDIATE THEORY OF SUPPLY

innovation, any given output will be produced at a lower price than before
while any given price will call forth a higher output Thus the cost saving
innovation lowers price and raises output Figure 27 7{ii) shows this

prediction
In a monopoly, the marginal cost curve also shifts downward, thus, it is
obvious that the marginal cost and the marginal revenue curves will now
intersect at a higher level of output than they did previously Since the
demand curve is unchanged, it follows that the price must fall \Vc conclude
that It IS a prediction both of the theory of competition, and of the theory
of monopoly, that a fall in costs will cause an increase m output and a
reduction in price Tlius, the benefit from any fall in costs will be to some

Fig 27 7 The effect on price of a fall in the marginal cost curve

(i) for a monopolist

(ii) for a competitive industry

extent passed on to consumers in terms of lower prices in the cases ofboth

competition and monopoly
But while prices will fall in both cases, they will not usually fall to the
same degree, as Figure 27 7 illustrates The demand curve is the same m
cost curve for the monopolist (i) ««
both (i) and (ii), and the marginal
identical with the supply curve of the competitive industry (n) The
monopolist equating MR and A/Cj is led to the price The competitive
industry equating demand and supply is led to the price pi The lines MCi
and an equal r^uction in costs in the two cases, and
^2 reflect the effect of
lead to the new pnees and p2 Notice that the amount of the price change
IS greater in the compcutive case than in the monopolistic one The reason
IS that the monopolist is guided by the marginal revenue curve, which is
steeper than the demand curve the same vertical fall in marginal costs
leads to a lesser increase in quantity and thus a lesser fall in price than m
perfect competition
 :

THEORIES OF COMPETITION AND MONOPOLY 365

The theory thus gives rise to the following prediction:

Other things being equal, prices and quantities will

change less in monopoly than in competition in
response to a change in marginal costs.

THE EFFECT OF TAXES ON PRICE AND OUTPUT*

There are many kinds of taxes. ^Ve shall here consider only three of them
a tax thatis a fixed amount per unit produced a tax that is a fixed amount; ;

and a tax that is a fixed percentage of profits. We shall call the first kind a
per unit tax, the second kind a lump-sum tax and the third kind a profits
tax.^

Per unit tax: A per unit tax increases the cost of producing each unit by
the amount of the tax. The marginal cost curve of eveiy firm shifts vertically
upward by the amount of the tax. In perfect competition, this means that
the industry' supply cur\'e shifts upward by the amount of the tax.
wth a little re-interpretation, illustrates this case. Suppose
Figure 27.7,
that MC2
and S2 reflect the cost and supply conditions (under monopoly
and perfect compedtion respecdvely) before the tax. A tax of £t per unit is
imposed, and the cuiY'es shift to MC^ and 6’,. It is immediately evident that
in both cases such a tax will lead to an increase in price and a decrease in
output. As we have drawn the diagram, the increase in price is less than
the amount of the tax, even in the case of perfect competition. Is this
necessarily the case ?

The answer is No. In perfect competition, whether the price effect of a

per unit tax is less than, equal to, or greater than the amount of the tax
depends upon the shape of the supply cur\'e. If the supply curv'e is rising,
price ^vill increase by less than the amount of the tax; if the supply curve is

horizontal, the price rise will be equal to the amount of the tax: if the
supply curve is declining, the increase in price rvill be greater than the
amount of the tax. Figure 27.8 illustrates these three cases. In (i), (ii)and
(iii), the imposition of a tax of £t per unit shifts the supply curv'e S up
vertically to S', and leads to a change in price of Ap?
In competition, in the short run, since costs are rising, prices will increase

1 subsidy is a negative tax. We shall discuss the tax case, and leave extension to subsidies

as an exercise for the student.

2 As an exercise the reader should try' to extend the analy'sis to a further type of tax — a tax
that a fixed percentage of the price. Such a tax is called an ad valorem tax.
is

3 This case was analysed in greater detail in Chapter 11. If you have any difficulty with the
present argument, re-read pages 134-9.
366 THE INTERMEDIATE THEORY OP SUPPLY

by less than the tax, m the long run, this need not be the case, as sve have
‘
just seen

Lump-sum tax Consider now the cflTect of a lump sum tax Such taxes
increase the fixed costs of the firm but do not increase marginal costs ^ The

(f) Ruing costs

(ii) Constant costs

(ill) Falling costs

Ftg 27 B The effect of taxes on price

and output m a competitive industry

and output of a change in fixed costs is zero, both

short-run effect on price
in perfectcompetition and in monopoly Since both marginal costs and
marginal revenues remain unchanged, the profit -maximising level of output

1 In the monopoly case it is also possible for ihe pnee change to be greater than the amount
of the tax but the conditionis not a simple one and we shall not present it ith a horizontal

marginal cost curve the monopolist s price change wiU be less than the amountofa per unit tax

2 Denoting the lump-sum tax by T and using an astenslt to indicate post tax values

AfC,- TC,-TC. ,

TT? - 7C.-t-7*

7C? 1
= 7C. ,+ r
A/C* = 7C?-7Cr ,

= TC,+ r-(7V. ,+ T)
= TC.-TC. ,

= Af(^
 THEORIES OF COMPETITION AND MONOPOLY 367

cannot be affected. Hence we deduce the implication that a lump-sum tax

leaves priee and output unehanged in the short run.*
In the long run, the tax has no effect on a monopolist’s price and output.
Assuming that the monopolist was previously making profits, then the tax
merely reduces the level of these profits. But, since the monopolist was
making as much money as he possibly could before the tax, there is nothing
any of the tax burden onto his eustomers. (Of course,
that he can do to shift
if the lump-sum tax was so large that, even at the profit-maximising level

of output, profits were reduced to less than zero, the monopolist would
cease production in the long run.)
In the case of perfect competition, we would expect the tax to affect price
and output in the long run. If the industry was in equilibrium with zero
was instituted, then the tax would lead to losses.
profits before the tax
Although nothing would happen in the short run to price and output.

Fig 27.9 The effect of a profits

tax on a monopolist’s output.

equipment would not be replaced as it wore out. Thus, in the long run, the
industry would contract, and price would rise until the whole tax had been
passed on to consumers and the firms remaining in the industry were again
covering costs.

Profits tax; Suppose one price-quantity combination (say/;*, q*), gives

the firm higher profits (without considering taxes) than any other. If the
central authorities impose a 20 per cent profits tax, the firm will have only
80 per cent as much after-tax profit as it had before this will be true for ;

each possible level of output. Therefore q* will still be the profit-

maximising output. This is illustrated in Figure 27.9. We conclude that a
tax on profits will have no effect on price and output in the short run.
In the long run the same result would occur if the tax only applied to
profits in excess of all opportunity costs. In fact, however, a tax on profits as

1 Unless the tax is so high that it causes the producers simply to abandon the business at

once. Because of this, a lump-sum tax is a little bit different from a fixed cost in that it can be
avoided by quitting the industry.
366 THE INTERMEDIATE THEORY OF SUPPLY

by less than the tax ;

in the long run, this need not be the case, as wc have
‘
just seen

Lump-sum tax Consider now the efTect of a lump sum tax Such taxes

increase the fixed costs of the firm but do not increase marginal costs ^ The

(i) Rising costs

(ii) Constant costs

(ill) Falling costs

fig 27 6 The effect of taxes on pnee

and output in a competitive industry

and output of a change in fixed costs is zero, both

short-run effect on price
in perfect competition and m
monopoly Since both marginal costs and
marginal revenues remain unchanged, the profit-maximising level of output
1 In the monopoly case it is also possible for the pnee change to be greater than the amount
of the tax but the condition is not a simple one and we shall not present il \\ iih a honzonul
unit tax
marginal cost curve the monopolist s price change will beless than the amount of a per

2 Denoting the latnp-sum lax by 7" and using an asientl to indicale post lax
values

A/C, = 7C;-7C,.,
7C* = TCj+r

A/C* = 7C*-7CJ_,

= 7C.-7C._,
= A/C,
 THEORIES OF COMPETITION AND MONOPOLY 367

cannot be affected. Hence we deduce the implication that a lump-sum tax

leaves priceand output unchanged in the short run.^
In the long run, the tax has no effect on a monopolist’s price and output.
Assuming that the monopolist was previously making profits, then the tax
merely reduces the level of these profits. But, since the monopolist was
making as much money as he possibly could before the tax, there is nothing
that he can do to shift any of the tax burden onto his customers. (Of course,
if the lump-sum tax rvas so large that, even at the profit-maximising level
of output, profits were reduced to less than zero, the monopolist would
cease production in the long run.)
In the case of perfect competidon, we would expect the tax to affect price
and output in the long run. If the industry' was in equilibrium with zero
profits before the tax was instituted, then the tax would lead to losses.
Although nothing would happen in the short run to price and output,

Fig 27.9 The effect of a profits

tax on a monopolist’s output.

equipment would not be replaced as it wore out. Thus, in the long run, the
industry would contract, and price would rise until the w'hole tax had been
passed on to consumers and the firms remaining in the industr)' were again
covering costs.

Profits tax: Suppose one price-quandty combination (say p*, q*), gives

the firm higher profits (without considering taxes) than any other. If the
central authorities impose a 20 per cent profits tax, the firm will have only
80 per cent as much after-tax profit as it had before; this will be true for
each possible level of output. Therefore q* will still be the profit-
maximising output. This is illustrated in Figure 27.9. We conclude that a
tax on profits wll have no effect on price and output in the short run.
In the long run the same result w^ould occur if the tax only applied to
profits as
profits in excess of all opportunity costs. In fact, however, a tax on

abandon the bus

I Unless the tax is so high that it causes the producers simply to
e
once. Because of this, a lump-sum tax is a little bit different from a fixed cost in
avoided by quitting the industry.
368 THE INTERMEDIATE THEORY OF SUPPLY

defined by the tax authorities is a tax on the return to some very essential
factor services (seepage 252) Thus a so-called profits tax may have the
effect of reducing the return to factors of production This means that a
profits tax may have important effects on the allocation of resources in the
long run
In a perfectly competitive industry profits as they are defined by the
economist are zero in the long run, but profits as they are defined by the
tax authorities are positive This means that perfectly competitive industries
will pay profits tax in the long run If profits as defined by the tax authorities

are the same proportion of total costs in all industries then this will not

matter, for the relative attracUon of various industries will not be affected

If, however, profits as defined by the tax authorities make up a different

proportion of total costs in various industries, then the relative earning in

different industries of factors that arc paid out of what the tax authorities
call profits, will be changed and the allocation of resources will be affected
by the tax An example of (his effect is elaborated in detail in Chapter 38

A COMPETITIVE INDUSTRY THAT BECOMES

MONOPOLISTIC

As well as yielding certain testable predictions based on the mmirhal amount

of knowledge normally assumed, the theory of the firm and industry pro-
vides a framework into which more detailed knowledge can be fed if we
have It, further testable predictions can then be extracted from the theory
If, for example, we know the elasticity of demand and how the demand

curve IS shifting through time, and if we have detailed information about

costs, the theor> can be made to yield a large number of predictions In this

section we shall consider one example of the use of the theory in a specific

context This example, although hypothetical, is not unlike many situations

found in the real world, particularly in the retail trades

Assume that in a particular city the barbers all belong to a single trade
association Therefreedom of entry into the industry in the sense that
is

anyone who has obtained a stated set of qualifications can set up as a barber
All barbers, however, must join the association and must abicfe 6y its rufes
The association sets the pnee of haircuts and strictly enforces this single
price Thus there is no price competition between barbers
Periodically the association raises the fixed price of haircuts in an attempt
to raise the incomes of members The association is strong enough to
its

prevent any illicit and to resist all attempts to secede on the

price cutting
part of individual barbers The question is will the barbers succeed m
’aising their incomes by raising the price of haircuts^ If >ou were a con
 THEORIES OF COMPETITION AND MONOPOLY 369

suiting economist called in to advise the barber’s association, what would

you predict be the consequences of this price rise ?
to
Clearly we need to distinguish between the short- and the long-run effects
of this increase in the price of haircuts. In the short run the number of
barbers is fixed. Thus in the short run the answer is simple enough; it all
depends on the elasticity of the demand for haircuts. If the demand elasticity
is less than one, total expenditure on haircuts will rise and so, therefore,

will the income of barbers if demand elasticity exceeds one, the barbers’
;

incomes will fall. The problem in the short run amounts to that of getting
some empirical knowledge about the elasticity of demand for haircuts. If
you were actually advising the barbers’ organisation you might be lucky
enough to be able to refer to a full scale econometric study of demand. In
the case of haircuts this is unlikely and you would probably have to try to
gain some idea of demand elasticity by studying the effects of changes in
haircut prices either at other times or in other places. We cannot go into
this matter except tomention that you should not fall into the trap of
reasoning as follows; ‘Haircuts are a necessity for no one goes without one.
Therefore the demand for haircuts will be almost perfectly inelastic.’ The
reader should easily spot the fallacy in this argument. There are many

reasons why the demand for haircuts is not perfectly inelastic and suffice it
to mention only one the time between haircuts is by no means a constant.
;

An increase in the average period between haircuttings from three to four

weeks would cause a 33^ per cent decline in demand; if this change in
habits were occasioned by a 20 per cent rise in price, then the demand
elasticity over this price range would be 1-66!
Let us say, however, that, on the basis of the best available evidence, you
estimate the elasticity of demand over the relevant price range to be substantially
less than one, say 0-3. You then predict that the barbers will be successful in
raising incomes in the short run a 20 per cent rise in price will be met by a
;

6 per cent fall in business so that total income of the typical barber would
riseby about 13 per cent. In predicting the consequences you would also
need to estimate the length of the short run for this industry (a couple of
years?).
Now what about the long run? If barbers were just covering all costs
before the price change, they will now be earning profits. Barbering will
become an attractive trade relative to others requiring equal skill and
training, and there will be an inflow of barbers into the industry. As
the number of barbers rise, the same amount of business must be shared
out amongst more and more barbers, so that a typical barber will find a
steady decrease in the amount of business that he does. His profits will thus
decrease. Profits maybe squeezed from another direction. Faced with
also
increasing excess capacity - a typical barber could handle much more
370 THE INTERMEDIATE THEORY OF SUPPLY

business than does in fact come his way - barbers may compete against
each otlier for the limited number of customers Since they are unable to
compete through pnee cuts they can only compete m service Sho|Bmaybe
spruced up, expensive magazines purchased, etc In these ways competition
will raise costs Thus profits of the individual barber will be attacked from
two directions, falling revenues and rising costs Tins movement will con
tinuc until barbers arc once again just covering all costs Once this comes
about there will be no further attraction to new entrants The industry
would settle down m a new position of long run equilibrium m which
individual barbers arc just covenng all of their opportunity costs There will
be more barbers than m the onginil situation, but each barber will be
working a lower fraction of the day and will be idle for i larger fraction
{: e there will be excess capacity), the total number of haircuts given by
,

each barber would be diminished and possibly the level of costs and service
will have increased Tims your report would say ‘you may succeed m the
short run (if demand is sufficiently inelastic) but the policy is bound to be
self defeating in the long run
The general moral of the story is that if you cannot control entry you
cannot succeed keeping earnings and profits above normal m the long
m
run If price compctuion is ruled out, then profits will be driv en down by the
creation of excess capacity Producen associations which are successful m
keeping earnings up arc those which arc successful in restricting entry
 THEORIES OF COMPETITION AND MONOPOLY 371

that the solution depends only on the ‘objective’ factors of costs and market
demand, for the attitudes of each competitor to the stratagems of his few
opponents becomes important, and, for the same costs and market demand,
the equilibrium of the industry will var>' considerably as the
psychology of
the competitors varies. In this sphere the traditional theory has had the
least success, and it may be true that an entirely different framework will
have to be worked out in order to deal successfully with the problems of
oligopoly.
As w'ell as providing some general insight into the possible workings of
the price system, the theor)' of the firm provides a series of predictions, or
testable hypotheses, about how firms and industries will react to various
changes in the data. Some of these predictions have been developed in the
present chapter. Generally, the more information do we start \vith, the
richer is the set of predictions that can be derived from the theor)'. If we
knew assumed) no facts, we could deduce no consequences; for, in spite
(or
of occasional appearances to the contrary, economic theory' is unable to
produce something out of nothing. Out of ignorance comes only ignorance.
If we knew (or assumed) only one thing, for example, that firms maximise
profits, we could deduce only this one thing. We need to know at least two

things before we can

begin to deduce any further consequences of these facts.
Generally, the more things do we know (or assume we know), the more
likely is it that we shall be able to deduce interesting and possibly un-
suspected consequences from these facts.

The traditional theory of the firm assumes a quite modest degree of

knowledge about the real world, and the testable implications which can be
shown to follow from the theory are on a correspondingly modest scale. The
theory assumes;

1 firms try to maximise profits;

2 demand curves slope downwards;
3 short-irun cost curves slope upwards (eventually);
4 the number of firms in the industry is known;
5 the degree of similarity between the products of
competing firms is known.

The implications deduced are of the sort which we have been considering

in this chapter.
CHAPTER 28

MONOPOLY VERSUS COMPETI-

TION: PREDICTIONS ABOUT
PERFORMANCE

Monopoly has been regarded with suspicion for a very long time It is

often held that modern economic theory has proved that monopoly is a
system whereby the powerful producer exploits the consumer while com*
petition always works to the consumer’s advantage Indeed, the founder of
English Classical economics, Adam Smith, in his Wealthof Nations,
a ringing attack on monopolies and monopolists and, since that time, most
economists have been advocates of free competition and critics of monopoly
In this chapter we shall first consider this Classical case against monopoly
and then go on to see what else can be said about monopoly and com*
petition on the basis of positive economics

THE CASE AGAINST MONOPOLY

The Classical case againstmonopoly is to a very great extent based on a
single prediction a perfectly competitive industry should be monopolised
if

and the cost curves of all productive units be unaffected by this change the
pnee will rise and the quantity produced will fall Thus, given identical

cost and demand conditions, monopoly leads to a lower output and a

higher price than does perfect competition We shall first prove that this

proposition does in fact follow from the theories of competition and

monopoly
Equilibrium under perfect competition occurs where supply equals
demand Since the supply curve is the sum of the marginal cost curves, it
follows that in equilibrium, marginal cost equals price ‘ This is illustrated

I See Chapter 22, pages 302-3 Since each firm is faced with a perfectly elastic dentin'!
curve at the prevailing market price, it will maximise profits at
by producing up to the point
which marginal cost equals price
 MONOPOLY VERSUS COMPETITION 373

in Figure 28.1 in which the competitive supply curve is MC, the competitive
output is Oq^ and the competitive price Op^.
Now assume that this industry is monopolised as a result of a single firm
buying out all the individual producers. Further assume that each plant’s
cost curve is unaffected by this change. In other words we assume that there
are neither economies nor diseconomies resulting from the coordinated
planning of production by a single decision unit. This means that the
marginal costs will be the same to the monopolist as to the competitive
industry : the competitive industry’s supply curve will be the marginal cost
curve to the monopolist. But the monopolist who seeks to maximise profits
will equate marginal costs not but to marginal revenue. We must
to price,
now draw a marginal revenue curv'e in Figure 28.1 which, as we have

Fig 28.1 The effect on price and

output of the monopolisation of
a perfectly competitive industry.

already seen, will lie below the demand curve at every level of output. Now
the output of the industry falls from Oq^ to 0q„ while the price rises from

^Pc to Op^.

The argument may be summaiiscd in the following steps:

1 In perfect competition, marginal cost equals price.
2 In monopoly, marginal cost equals marginal revenue.
3 For any given output, marginal revenue is less than price as long as
the demand curve is downward sloping.
4 Therefore at the perfectly competitive output, marginal revenue is

less than marginal cost.

5 Since the marginal revenue curve slopes downward, and the marginal
cost curve slopes upward, these two can only be equated by reducing output
below the perfectly competitive level.

6 Since the demand curve slopes downward, the lower level of output
will necessarily be associated with a higher price.
The prediction that we have derived does have a strong common-sense
374 THE INTERMEDIATE THEORY OF SUPPLY

appeal In perfect competition no one firm is big enough to affect the price
of the product by varying its production Every firm therefore fixes its
production at the most profitable level on the assumption that the market
price IS given When the industry is monopolised, it becomes possible to
driveup price by restricting output, and this is what we have shown will
happen What is possibly not obvious from the purely common sense
argument is that, whatever the cost and demand conditions, it will always
pay the monopolist to restrict output below, and to raise price above, the
*
perfectly competitive level

THE CASE FOR PERFECT COMPETITION^

If prices are higher and output lower under monopoly as compared with
competition, are we justified in saying that one price-output situation is
in more nearly optimal, than the other’ The second
any seme better, or
prong of the Classical case against monopoly consisted m showing that
perfect competition led, m equilibrium, to a use of resources that was, m a

clearly defined sense, the best’ one

We cannot go into this enquiry m any detail here and we shall have to

satisfy ourselves with a rather intuitive statement of why the allocation of

resources resulting from perfect competition was regarded as superior to the

one resulting from monopoly
Under perfect competition mai^inal cost equals price Marginal cost
shows the cost of producing the last unit of the commodity actually pro
duccd and the price indicates what consumers are prepared to pay for the
last unit of the commodity purchased ’ It follows that consumers are pre

pared to pay for the last unit they actually purchase an amount exactly

equal to the cost of producing that last unit In monopoly, price exceeds
marginal cost From this it follows that consumers pay for the last unit they
actually purchase an amount greater than what it actually costs to produce
it Furthermore consumers would be prepared to buy further units for an
profits
1 This IS why on page 361 above we said that a producers co-op can always raise
above their competitive equilibrium level by restricting output
taking us in the direction ofweUare economics a subject usually thought
2 This discussion is

to be a normative one founded on value judgments Personally 1 accept the case argued
by

GA Archibald Welfare Economics Ethics and Esscnlialism ,

fcMcmico Vol 26 1959 that

welfare economics is either mathematics (the conditions for an optimum) or positive economics

(predictions about what changes wiU expand the consumer s range of choice) My reason for
(or can t be
omitting a detailed discussion of welfare economics at this stage is not that it isn t

rather
made to be) positive economics but that in my
opinion u is difficult to handle this
subtle subject adequately in a first year course To try to do so is to court either naive and
slavish acceptance or uninformed rejection of all welfare principles - the reaction varying with
the philosophical and political position of the student
ce
3 Consider the individual
demand curve illustrated in Figure 28 2 When the market pr
’Oj the individual buys 100 higher
units Ifthepnceis21sheonlybuys99 units Ifthepnceis
 :

MONOPOLY VERSUS COMPETITION 375

amount greater than the cost of producing these units ; consumers are, ho\v-
ever, not allowed to purchase these extra units because the monopolist is
restricting output in order to maximise his profits. In Figure
28.1, for
example, the marginal cost at the monopolist’s equilibrium is q„u and price
is q^v so that consumers are prepared to pay uv more for another unit than it
would actually cost to produce that unit. They would also be wilhng to buy
at a price in excess of the marginal costs of production a total of
q„q^ units
more than they are permitted to buy.
There is a strong intuitive appeal to the idea that consumers will in some
sense be ‘better off when marginal cost equals price, than when it is less
than price. It can in fact be shown, and it is a well-known proposition in
welfare economics, that, given a very large number of very strict conditions,
the equating of marginal costs to price in all lines of production will ^aeld
an optimal situation in the sense that it will be impossible to make some consumers
better off without simultaneously making others worse off', while when marginal
costs do not equal prices in some industries, this wll result in a sub-optimal
situation in the sense that it will be possible to make some consumers better off without

making any worse off by changing some prices and outputs in the economy.
The demonstration of the above proposition belongs to a course in
welfare economics. There is no doubt that this proposition has had a
strong effect on the attitudes that many economists and students of eco-
nomics have to the question of the value of the price system. If the unhindered
price system produces an optimal situation, then interference by the central
authorities can only move us away from this optimum and it may be
condemned for this reason. Because the proposition has exerted so much
influence and has led so many to the conclusion stated above, it is probably
worth while pointing out just how far away it is from any applicability for

the world of our experience.

The argument that perfect competition leads to an optimum allocation

of resources requires among other things, the following

1 There should be no divergence between private and social costs and

he buys fewer market price does settle at 20s so that he buys 100 units, he gets all
units. If the
but the last unit at a price less than he would be prepared to pay for
them. (He is prepared,
for example, to pay 2 Is to get 99 units.) Thus, assuming
his demand curve be downward sloping, the price
to
measures what the consumer is prepared to pay for the

last unit that he purchases; he would be prepared to pay

more than the current price in order to obtain all but

the last unit (which is why he would continue to buy
them and to obtain additional units,
at higher prices) ;

he is only prepared to pay an amount less than the market

price (which is why he does not at present buy these
additional units).
374 THE INTERMEDIATE THEORY OF SUPPLY

appeal In perfect competition no one firm is big enough to affect the price
of the product by varying its production Every firm therefore fixes its
production at the most profitable level on the assumption that the market
price IS given When the industry is monopolised, it becomes possible to
driveup price by restricting output, and this is what we have shown will
happen What is possibly not obvious from the purely common sense
argument is that, whatever the cost and demand conditions, it will always
pay the monopolist to restrict output below, and to raise pnee above, the
‘
perfectly competitive level

THE CASE FOR PERFECT COMPETITION'

If prices are higherand output lower under monopoly as compared with
competition, are we
justified in saying that one price-output situation is
in any unse better more nearly optima), than the other’ The second
or
prong of the Classical case against monopoly consisted in showing that
perfect competition led, in equilibrium to a use of resources that was in a
clearly defined sense, the ‘best’ one
We cannot go into this enquiry in any detail here and we shall have to

satisfy ourselveswith a rather intuitive statement of why the allocation of

resources resulting from perfect competition was regarded as superior to the
one resulting from monopoly
Under perfect competition, marginal cost equals pnee Marginal cost
shows the cost of producing the last unit of the commodity actually pro
duced, and the price indicates what consumers are prepared to pay for the
last unit of the commodity purchased ^ It follows that consumen are pre

pared to pay for the last unit they actually purchase an amount exactly
equal to the cost of producing that Iasi unit In monopoly, price exceeds
marginal cost this it follows that consumers pay for the last unit they
From
actually purchase an amount greater than what it actually costs to produce
It Furthermore, consumers would be prepared to buy further units for an
raise proliu
1 This IS why on page 361 above we said that a producers co-op can always
above their compeCiUve equilibrium level by restnetmg output
2 This discussion is taking ui m the direction of welfare economics a subject usually thought
to be a normative one founded on value judgments Personally I accept the case argued by
GA Archibald Welfare Economics Elbica and Essentialism , Economiea Vol 26 1959 that

economics
welfare economics iseither mathematica (the conditions for an optimum) or positive
what changes will expand the consumer s range of choice) My reason
for
(predictions about
isn (or can t e
omitting a detailed discussion of welfare economics at this stage is not that it t

rather
made to be) positive economics but that in my opinion it is difficult to handle this
an
subtle subject adequately in a first year course To try to do so is to court either naive
slavish acceptance or uninformed rejection of all wettare principles the - reacnon varying wit
the philosophical and political position of the student
price
3 Consider the individual
demand curve illustrated in Figure 28 2 When the market
’Os the individual buys 100 higher
units If thcpnccisZUhc only buys 99 units If the pnceis
 MONOPOLY VERSUS COMPETITION 375

amount greater than the cost of producing these units consumers are, how- ;

ever, not allowed to purchase these extra units because the monopolist is
restrictingoutput in order to maximise his profits. In Figure 28.1, for
example, the marginal cost at the monopolist’s equilibrium is q„u and price
is q„v so that consumers are prepared to pay uv more for another unit than it
would actually cost to produce that unit. They would also be willing to buy
at a price in excess of the marginal costs of production a total of q„qc units
more than they are permitted to buy.
There is a strong intuitive appeal to the idea that consumers will in some
sense be ‘better off when marginal cost equals price, than when it is less

than price. It can in fact be shown, and it is a well-known proposition in

welfare economics, that, given a very large number of very strict conditions,
the equating of marginal costs to price in all lines of production will yield
an optimal situation in the sense that it will be impossible to make some consumers
better off without simultaneously making others worse off-, while when marginal
costs do not equal prices in some industries, this will result in a sub-optimal
situation in the sense that it will be possible to make some consumers better off without

making any worse off by changing some prices and outputs in the economy.
The demonstration of the above proposition belongs to a course in
welfare economics. There is no doubt that this proposition has had a
strong effect on the attitudes that many economists and students of eco-
nomics have to the question of the value of the price system. If the unhindered
price system produces an optimal situation, then interference by the central
authorities can only us away from this optimum and it may be
move
condemned for this reason. Because the proposition has exerted so much
influence and has led so many to the conclusion stated above, it is probably
worth while pointing out just how far away it is from any applicability for
the world of our experience.
The argument that perfect competition leads to an optimum allocation
of resources requires among other things, the following;
1 There should be no divergence between private and social costs and

he buys fewer market price does settle at 20j so that he buys 100 units, he gets all
units. If the
but the last unit at a price less than he would be prepared to pay for them.
(He is prepared,
for example, to pay 2 Is to get 99 units.) Thus, assuming
his demand curve tobe downward sloping, the price
measures what the consumer is prepared to pay for the

last unit that he purchases; he would be prepared to pay

more than the current price in order to obtain all but
the last unit (which is ^vhy he would continue to buy
them at higher prices) ;
and to obtain additional units,
he is only prepared to pay an amount less than the market
price (which is why he does not at present buy these
additional units).
 )

376 THE INTERMEDIATE THEORY QF SUPPLY

between private and social revenues anywhere m

the econom> (see pages
253-4) Producing a good up to the point at which the revenue gamed
from the last unit just equals the opportunity cost of making u, makes
sense from society’s point of view only if the firms’ private costs refieci
the opportunity costs to society of using the resources elsewhere and if the
firms private revenues reflect the gains to society of having an extra unit
of this good produced We have already seen on page 253 that the social
md private costs often diverge and we shall See in Chapter 38 that it is

also true that social and private revenues often diverge This cntical
condition is therefore by no means always fulfilled
2 There should be perfect competition m all sectors of the economy
ensuring that the correct result obtains throughout This we know to be
wildly at variance with the facts (see Chapter 1^6)
3 All the perfectly competitive industries in the economy should have
neither economies nor diseconomies of scale In other words they must have
perfectly flat long run supply curves
There should be perfect information provided to all firms at zero
4
costsThis IS obviously untrue of our world iq which information is im
perfectand is only gathered at considerable cost
5 There should be perfect foresight about the future No comment is
needed on this condition in the real world
6 That there should be no learning by doing’ so that costs do not vary
with the number of times a process is repeated *

7 The process of innovation and growth should be strictly autonomous

(See Chapter 21 for a discussion of possible hy{>oiheses about innovation
If market forms other than perfect competition are more conducive to
growth future income will be higher if they are allowed to exist
8 The conditions necessary to produce this optimum should exist
simultaneously everywhere in the economy If this is not so, we have no
idea what the effect will be of fulfilling them somewhere the economy m
Specifically, if in a world of mixed market stt-uctures, we break up one
monopoly and make it into a competitive industry, we have no general

presumption even in a theoretical model about the likelihood of this moving

us closer to or further away from an optimum position ^
These are but a few of the conditions that must be fulfilled if perfect
1 See Kw Arrow The Implication of Learning by Doin^ Revitu/ of Economic Studies June
1962
2 This very important proposition destroys the basis of pieccm-al welfare economics V\e
know how to identify the best of all possible tvorlds (fioii\ the limited point of view of the
optimum we are discussing), but we have little clear idea how to order two states of the very
imperfect world in which we live If this were not so econonuju could not disagree as much as
they do about specific policy measures Sec
RG
Lipsey And k
J Lancaster
The General
y of Second Best , Renew of Eeennme Sluiui 1958
 . MONOPOLY VERSUS COMPETITION 377

competition is to lead to an optimum allocation of resources. It is

clear even
to the casual observer that each of these conditions is wildly at variance
with actual facts. Why
then is it that this theory has had such a profound
effecton so many economists and has led some to a worship of the un-
constrained price system ? The honest answer is T have no idea how it could
possibly happen’.
It is as if we w'ere trying to
decide w'hether abolishing capital punishment
•would raise or lower the number of murders in Britain in 1967 and a group
of ‘social scientists’ kept pointing out that in Heaven were only
^^’here there
Saints, who behaved in a certain heavenly way, capital punishment (=sent
to Hell) had a drastic effect in increasing the number of people attending
early-morning mass.
Of course, we do not have a world of perfect competition, and the above
theory does little to help us decide whether a bit more or less oligopoly
would help matters. Nonetheless we can still pass judgments on certain
implications of competition and monopoly which as ordinarj' citizens \ve do
not like. What we cannot do is to assume that economic theory predicts
that any increase in the degree of competition in the economy ahvays - or
even is likely to - increases the efficiency of resource allocation in the
economy.

The political appeal of perfect competition: Perhaps some of the

potent appeal of the propositions that perfect competition leads to an
still

optimum allocation of resources, is to be explained by the great appeal of

the perfectly competitive model to the liberal who was frightened b)' the
exercise of power either by private organisations or by the state. For one
who believes in the individual and hates all pmver groups, the perfectly
competitive model w'as almost too good to be true.
In the perfectly competitive world, no single firm and no single consumer
has any power over the market whatsoever. Individual consumers and pro-
ducers are both members of a group of many similar consumers or pro-
ducers, and no single one can affect the market. Individually they are
add to this the assumption that firms are
passive quantity adjustors. If w'e
profit maximisers, all firms become passive cyphers responding to market
signals and always doing what is most desirable from the society s point of
view. great impersonal force of the market produces an appropriate
The
response to all important changes. If, for example, tastes change, prices will
change and the allocation of resources will change in the appropriate
direction, but throughout the whole process no one will have any power
over anyone Millions of firms are reacting to the same price changes.
else.

If one refuses to react there Mil be countless other profit-maximising firms

all eager to make the appropriate changes. If one firm refuses to take
378 THE INTERMEDIATE THEORY OF SUPPLY

coloured employees or makes any other decision based on prejudice, there

will be millions of other firms that will recognise that profit maximisation

ISnot consistent with discrimination on the basis of race, colour or creed or

anything other than how hard a man is prepared to work
It was a noble model no one had power over anyone and yet the system
responded in a systematic way to changes in needs Many will feel that it is a
pity that it corresponds in so few aspects to reabty, as we know it Also we
should not be surprised to find that it still has a very strong appeal and that

some people still cling tenaciously to a belief that it describes the world in

which we live it would remove so many problems if only it did

IS MONOPOLY REALLY EVIL?

Looking only at the theories of perfect competition and monopoly, many
economists have been tempted to conclude that economic analysis has
proved that monopol> is evil, that it exploits the consumer, and that it
should be condemned and stamped out whenever possible ' It is extremely
important for the reader to realise, however, that nothing in positive
economics allows us to speak of monopoly as an evil Positive economics
seeks to establish valid predictions about the real world If our economics is
done well, we shall be able to say with some confidence that, if an industry
is monopolised, certain changeswill occur in pnee and output But there is

absolutely nothing m
economics that allows us to draw the con
positive
elusion that such changes are cither good or bad Positive economics at the
very best tells us the consequences of our actions whether or not we like
these consequences is a subjective matter We can be in complete agreement
about the consequences and yet disagree irreconcilably about whether they
arc desirable or undesirable, good or evil
It IS a prediction of the simple theory of monopoly and competition that,

providing demands and costs arc unaffected, pnee will be higher and output
lower when an industry is monopolised than when it is competitive, and
that the owners of firms will cam higher profits under monopoly than under
competition Which of these situations we prefer, and whether our margin
of preference is sufficient to justify incurring any substantial costs involved
inmoving from one to the other, arc matters that involve value judgments
and that take us, therefore, beyond the scope of positive economics

THE PROBLEMS OF CURRENT POLICY

The classical condemnation of noncompetitive forms of behaviour was based
on two basic propositions First, that the alternative to monopoly was perfect
Inlrtductoty Analysis 5th Edition
1 See for example P A Samuekon Etwuanus Alt
Graw Hill New York 1965 on page 555 there « a section on the effects of monopoly
"opoly entitled The Evils of Oligopoly
 MONOPOLY VERSUS COMPETITION 379

competition and second that the results that competition would produce
(e.g., a high level of output) were judged to be desirable. Today we realise

that very often the effective choice is not between monopoly and perfect
competition but between more or less oligopoly so that we are not sure what
the effects on price and output will be of a specific intervention we may
make. Thus, even if we accept the perfectly competitive result as being more
desirable than a completely monopolistic one, this does not in itself tell us
much about the real decisions that face us. In the remainder of this chapter
we shall about competitive and monopolistic situations in general terms
talk
with a view to considering the effects of encouraging a little more or less
competition than we now have.

Policy Implications of Accepting the Classical Position

The belief that competition produced ideal results and monopoly nonideal
ones led at once to the notion of prohibiting by law the practice of monopoly.
Antimonopoly laws, perhaps the first manifestation of the classical down-uath-
monopolies policy, outlaw attempts to monopolise an industry and give
the courts the power to dissolve an existing monopoly into a larger number
of independent companies, if they deem it necessary. (We discuss anti-
monopoly policy further in Chapter 38.)
A second policy - public-utility regulation - grew out of the classical econo-
mists’ belief that, insome fields (for example, in transportation and public
utilities), competition was impossible. These fields were regarded as areas

of ‘natural monopoly’ - as exceptions to the classical assumption that market

organisation does not affect levels of cost. The cost advantage in having one
railway between two points rather than fifty railways (or one water com-
pany in a city, or one telephone company in a country) is evident. But, said
the classical economists, a natural monopoly, one in which a single seller is

able to operate more efficiently than if he were subject to competition,

cannot be allowed to maximise profits by charging any price it ivill. Public-
utility regulation gives, to appropriate public authorities, control over the

price and quantity of sendee provided by a natural monopoly prices are ;

geared to provide only a ‘fair’ return on capital. The principle behind the
policy is to regulate a natural monopoly in such a way as to secure the

price, output and profit results that competition would produce, were it able
to operate efficiently.

COSTS AND EFFICIENCY UNDER MONOPOLY

AND COMPETITION
The basic proposition monopoly leads to higher price and loiver output
that
than does competition depends on the assumption that costs are unaffected
380 THE INTERMEDIATE THEORY OF SUPPLY

when an is monopolised If any savings are effected by combining

industry
numerous competing groups into a single integrated operation, then the
costs of producing any given level of output will be lower than they were
previously If this cost reduction does occur, then it is possible for output to
be raised and price to be lowered as a result of the monopolisation of a
perfectly competitive industry Such a situation is illustrated in Figure 28 3
The competitive equilibrium price is Op^ and quantity is Oq^ If the industry
were monopolised and costs were unaffected, production would fall to

0q„ and price would rise to Op„ If, however, the integration of
, the in-
dustry into a single unit causes some increase in the efficiency of organisa-
tion and thereby reduces costs, then the marginal cost curve will shift

downward If it shifts to then production will rise to 0q„ and pnce

Fig 2$ 3 The effectt on price and

output of a monopoly that is more
than a large number of
efficient
competing firms

will fall to Op^ This shows that the monopolisation of an industry, com
bincd with a sufficiently large consequent increase efficiency, can result m
in a fall in price and a nsc in quantity produced, as compared to the
‘
compctitne industry
Of course, it is an industry may
also possible that the monopolisation of
reduce the efficiency of production and so shift the marginal cost curve up
ward In this case, monopolisation will, a fortton, raise price and lower out-
put as compared to the competitive industry The reader should draw his
own diagram showing the effects on pnee and output of a monopolisation

that caused costs to rise above those ruling under competition

It IS sometimes argued that monopolisation will lower costs because
wasteful duplication will be clinunatcd, and because economies of scale will
1 The reader should be able to show for himself that, if the elasticity of demand were les*

an one at the competitive price the monopolist would reduce output and raise pnce no
atter how large the reduction in h» costs
 MONOPOLY VERSUS COMPETITION 381

result from, for example, establishing one coordinated management body

for the industry. On the other hand, it is often argued that competition
forces the individual firm to be efficient because the firm will not survive
unless it keeps its costs as low as its competitors’, whereas inefficiency is more
common under monopoly because, although inefficiency may reduce profits,
it will not result inbankrupting the monopolistic firm.
Wecannot predict the effects on price and output of monopolising a
competitive industry unless we know the effect of this change on the
industry’s costs. If, for example, monopolisation usually results in large cost
savings, then it may be that monopolisation usually results in a fall in price
and an increase in output. If, on the other hand, monopolisation usually
either leaves costs more or less unchanged or increases them, then it will
cause a and a fall in output. This is as far as our theory can take
rise in price

us; it can predict what empirical magnitudes are important, but, until we
have some evidence of the effects of monopolisation on an industry’s cost
structure, we cannot predict the effect that it will have on price and output.

The Incentive to Innovate

Does our theory predict anything further about the relative incentive, under
monopoly and under competition, to keep costs down by being efficient or
to reduce costs by introducing new innovations ? As far as profits are con-
cerned, both the monopolist and the perfect competitor have an incentive
to reduce costs. A monopolist can always increase his profits if he can reduce
his costs. We saw in Chapter 27 (page 363-5) that a cost reduction will
cause the monopolist to produce more, to sell at a lower price, and thus to
increase his profits. Furthermore, since he is able to prevent the entry of
new firms into his industry, these additional profits will persist into the long
run. Thus, from the standpoint of maximising his profits, the monopolist has
both a short- and a long-run incentive to reduce his costs.
The firm in perfect competition or in monopolistic competition has the
same incentive in the short run, but not in the long run. In the short run, a
reduction in costs will allow the firm that was just covering costs to earn
profits. In the long run, other firms will be attracted into the industry by
these profits. Existing firms will copy the cost-saving innovation, new firms
will enter the industry using the new techniques, and the profits of the

innovator will eventually disappear. The effectiveness of profits as an in-

centive to reduce costs for a firm in competition depend on the

will

magnitude of the extra profits and the length of time over which they persist.
If, for example, it only takes a few months for existing firms and new entrants
firm profits will
to copy and install the new invention, then the original s

be above normal for only a very short time, and the extra profits actually
382 THE INTERMEDIATE THEORY OF SUPPLY

earned may not be sufficient to compensate for the risks and the costs of
developing the new innovation In such the direct incentive to
cases,
innovate would be absent from a competitive industry On the other hand
if It takes several years for other firms to copy and install the cost-saving
innovation, then the profits earned over these years by the innovating firm
might be more than sufficient to compensate for all costs and risks, and
yield a handsome profit as well In this case, the incentive to innovate is
present in a competitive industry
Evidently, the key issue that we are faced with is the extent to which
market organisation affects the rate of innovation Before reading on >ou
should refresh your memory of the discussion of invention and innovation
by rereading pages 28&-9I
The greatest opponent of the classical position on monopoly was the

distinguished Austrian (and later Amcncan) economist, Joseph A

Schumpeter ‘
His argument m bnef was that only the incentive of profiu led
men to take the great risks of innovation and that monopoly power was
much more important than competition in providing the climate under
which innovation occurred The short-run profits of the monopolist pro-
vided the incentive for other men to find thor special advantage He called
the process of one monopoly being replaced by another the process of criattvi
destruction Speaking of the large firm with monopoly power, he said

What we have got to accept is that a has come to be the most powerful engine of
that progress m particular of the long run expansion of total output not only
and
in spue of, but to a considerable extent through, this strategy which looks so restne
live when viewed in the individual case and from the individual point of time In
this respect, perfect competition is not only impossible but inferior, and has no title
to being set up as a model of ideal efficiency It is hence a mistake to base the theory
of government regulation of industry on the principle that big business should be
made to work as the respective industry would work in perfect competition

Economists of the classical view were not unaware of this consideration

patent laws Patent laws represent an attempt
and they usually supported to

lengthen the short-run period during which the innovating firm can earn
supernormal profits as a reward for its innovation Once the patent expires,
other firms can copy the innovation and, when they do so, production will
expand until revenues fall to just cover all costs There is little doubt that,
were there no patent laws, many innovations could be copied with greater
speed than at present, and that the onginal innovators would not earn as

I Hu most famous book is Tht Tktetj rf Ecmiame Dntlopmmt (English ed

Harvard Urn
,

veniiy Press, 1934) The beginning student is referred to the lucid but less technical CafiM/ui"!
Socialism and Democracy {3rd ed , Harper, 1950) Both works are available in paperback
’’
Schumpeter, Capitalism, Socialism and Demacratj, page 106
 MONOPOLY VERSUS COMPETITION 383

much extra revenue to compensate

them for the costs and risks of develop-
ment. On
the other hand, patents can be imitated, so their real advantage
to the small firms should not be exaggerated. Indeed, it has been argued
that patents are of greater advantage to the monopolist who has the
resources to develop, patent and ‘keep on the shelf processes that might
enable a potential competitor to challenge his position.^
The monopolist has another possible advantage over the competitive firm
in the process of innovation in that funds for research and development are
more readily available to him. Tax laws that permit him to write off
business expenses may make research and development a relatively cheap
endeavour. Suppose that this year a monopolist expects to make £2 million
profit on which he have to pay taxes of approximately j^l million. If
will
he spends the £2 million on research and development, however, he will
show no profits for tax purposes this year, and will save ^1 million in taxes.
In effect, he can get £2 million worth of research for only ^1 million. His
successes will lead to future profits, and will strengthen his position as a
monopolist. Of course, in later years he will have to pay taxes on the profits
he earns.
There is thus no reason why cost conditions should be predicted to be the
same for a monopolistic industry and a competitive one. Although it has
been argued that the monopolist may be in the stronger position when
it comes to the incentive and means to innovate, the reverse may also be

argued that a firm has a stronger incentive to innovate if it is in a com-

:

petitive industry. A monopolist who does not innovate may be missing larger
profits, but at least he will have some profits. But, so goes the argument, if

the competitive firm does not innovate, its competitors may do so, and it
may find itself in a position in which it cannot even keep up with its com-
petitors, thus incurring bankruptcy and liquidation. It must be noted, how-
ever, that if the risk of being overtaken is smaller than the risk of innovating
the firm will still take it. Also one should note that, the easier it is to copy
an innovation, the less is the incentive to innovate and the less the loss from
not innovating when if it is very easy to copy any innova-
others do. Thus,
tion then the incentive to innovate for attack or for defensive reasons will
be very weak.
If you are tempted to choose one side or the other in this debate,
remember that, although theory is useful in posing alternative possibilities,
estimating the effects of monopoly and competition on efficiency and
innovation is an empirical matter that can be achieved only by an appeal

I This charge was a basic element in the case against the United Shoe Machinery Corpora-
tion in 1957. This company had a virtual monopoly on shoe-making machinery in the US. It
leased its machines to shoe manufacturers. The company had a huge array of patents, which,

it was charged, it simply held to prevent competitors from developing their own machines.
384 THE INTERMEDIATE THEORY OF SUPPL'V

to rcil-vvorld observations This appeal can be made only after the question
has been stated m a manner sufliaently precise so that it can be answered

An Example of the Decline and Fall of a Patent Monopoly

Ball-Point Pens'

We can illustrate much of the prcvtousdiscussion with the rather spectacular

case of the beha\iour of the economy in response to the introduction of a
revolutionary new technique for writing, thcball-pomt pen In 1945, Milton
Reynolds acquired a patent on a new type of pen that used a ball bearing
in place of a conventional point, he formed the Reynolds International Pen
Company, capitalised at $26,000 and began production on 6 October 1945
The Reynolds pen \% as first introduced with a good deal of fanfare by the

New York department store, Gimbels, who guaranteed that the pen would
write for two years without refilling The price was set at S12 50 (the

maximum price allowed by the s^artime Office of Price Administration

which was the body with which the American central authontics sought to
control prices during and after the Second \Vorld War) Gimbels sold
1 0,000 pens on 29 October 9 15, the first day they were on sale In the early
1

stages of production, the cost of production was estimated to be around

80f per pen

The Revnolds International Pen Compiny quickly expanded production
By early 1916, it employed more than 800 people in its factory and was
producing 30,000 pens per day By March 1916 it had S3 million in the
bank Demand was intense Macy’s, Gimbels’ traditional department store
introduced an imported ball-point pen from South America Its price
ri\al,
was S19 98 (production costs unknown)
The heavy sales quickly elicited a response from other pen manu
facturersE\ersharp introduced its first model in April, at 515 In July
1946, the business magazine, Fortune, reported that Schaeffer was planning
to put out a pen at S15, and E\crsharp announced its phns to produce a
‘retractable’ model priced at S25 Reynolds introduced a new model, but
kept the price at S12 50 Costs were estimated at 60^ per pen
The first signs of trouble emerged The Ball-point Pen Company of
Hollywood (disregarding a patent infringement suit) put a S9 95 model on
the market, and a manufacturer named David Kahn announced plans to

introduce a pen selling for less than S3 Fortune reported fear of an impend
mg price war m view of the growing number of manufacturers and the low

cost of production In October, Reynolds introduced a new model, priced

at $3 85, that cost about 30/ lo produce

! A and highly readable account of some aspects of the ball point case
fascinating «
Thomas Whiteside s 'Where are they Now* , The Nod Yvrker, 17 February 1951 Our discussion
leans heavily on Mr Whiteside » arlicJe
 MONOPOLY VERSUS COMPETITION 385

By Christmas, 1946, approximately 100 manufacturers were in pro-

duction, some of them selling pens for as little as S2.98. By February' 1947,
Gimbels was selling a bail-point pen made by the Continental Pen Company
for 98f. Reynold’s introduced a new model priced to sell at SI. 69, but
Gimbels sold it for 88/ in a price war with Macy’s. Reynolds felt betrayed
by Gimbels. Reynolds introduced a new model listed at 98/. By this time
bail-point pens had become economy items rather than luxury items, but
still were highly profitable.
In mid- 1948, ball-point pens were selling for as
little as 39/, and costing

about 10/ to produce. In 1951, prices of 25/ were common. In 1966, there
was a wide variety of models and prices, ranging from 1 1/ to S2.95, and the
market appeared stable, orderly, and only moderately profitable. Ball-
point pens were no passing fad, as every reader of this book knows. Their
introduction has fundamentally changed the WTiting-implement industry' in
America and in the world.
The example has interested observers in many fields.
ball-point pen
Lawyers have been concerned about the ease with which patent rights were
circumvented. Psychologists have noted the enormous appeal of a new pro-
duct even at prices that seemed very high. Advertising men have regarded
it as a classic case of clever promotion.
From the point of view of economic theory, it illustrates several things;

1 That a monopoly (in this case a patent monopoly) can in the short
run charge prices not remotely equal to costs and earn enormous profits.

2 That entry of new firms (even in the face of obstacles) rvill often occur
in response to high profits.
3That where it does occur, entry will in time drive prices down to a
level more nearly equal to the costs of production and distribution.
4 That the lag between an original monopoly and its subsequent erosion
by entry be long enough that the profits to the innovator,
may nevertheless
some of the imitators, may be very large indeed. (It is estimated
as well as to
that Reynolds earned profits as high as $500,000 in a single month — or about
20 times its original investment.)

Different observers might stress different aspects of the evidence sketched

in this example. Some would see it as an example of poiver of competition
in stimulating production of a desired commodity and ultimately making it
available at low cost to the consumer. Others would see it as evidence of the
ability of monopoly to exploit the gullible public. Still others would see it
asan example of the great incentive that capitaUsm provides to the successful
innovator (or promoter) to find and introduce a new product. It is probable
that all would be, to a degree, right.
13
386 THE INTERMEDIATE THEORY OF SUPPLY

THE RANGE OF CHOICE MONOPOLY VERSUS COMPETI

TION BETWEEN OLIGOPOLISTS
sometimes argued that one of the virtues of competition among several
It IS

producers is that it presents the consumer with a wide range of differentiated

commodities while complete monopoly with only one producer tends to

wards uniformity of product We shall not ask here to what extent a variety
of products IS desirable, but will merely ask if we would in fact expect
competition to produce more diversity of product than monopoly

An Example from Radio and Television

A very interesting case in which competition tends to produce a nearl)

uniform product while monopoly tends to produce widely differentiated

ones has been studied in detail
‘ This
is the case of radio and television The

argument runs somewhat as follows consider a case in which there arc two
potential radio audiences, one group, comprising 80 per cent of the total

audience, wishes to hear pop music the other group, comprising 20 per
cent wishes to listen to a concert of chamber music , and in which each
individual radio station seeks to maximise ns own listening audience If

there is only one station, it produce pop music If a second competing

will
station IS now opened up, its most profitable policy will be to produce a
similar pop*music programme on the grounds that half of the large audience
IS still better than all of tlie small one A third station would also prefer a

third of the large audience to all of the small one In fact five stations would
be needed before it would be profitable for any one station to produce a
programme of chamber music Thus competition between two or three
stations would tend to produce two or three almost identical pop music
programmes, each competing lor its share of the large audience A mono
poly controlling two stations would not, however, pursue this policy The
policy to maximise its total listening audience would be to produce pop
music on one channel and chamber music on the other The monopoly
might spend more money on preparing the programme for the larger
ViV!.\
,
vptwi TOwney Vo ptodott vmdax iptogiramme on
Its second channel - the optimal pohey for its second channel would be
to

go after the other 20 per cent of potential listeners so that, between the two
channels, the monopoly would have the largest possible audience In each
case, the individual firm tries to maximise its own listening audience, but
when there are two competing stations they both go after the same large
1 P O Steiner Monopoly and competition in T V some policy issues , The Mfuhisti^
School (1961) Program Patterns and IWerencei and the Workability of Competition in
Radio Broadcasting QuarUrlf JounuU
^
Eetmmtes May 1952
 MONOPOLY VERSUS COMPETITION 387

audience, ignoring the minority group, while, when there are tivo stations
owned by one monopoly, they go after both audiences, one for each station.
Under these circumstances, competition produces a uniformity of product
which ignores the desires of the minority, while monopoly produces a varied
product catering for the desires of both the majority and the minority group.
The theoty' was tested against the experience of the British radio which
is a three-station monopoly, and British television which was at that time,
based on competition between two stations, each taking as its criterion of
success its own listening audience. It was found that the three stations of the
monopolised radio produced very little similarity between the products
offered atany one time, while the two stations of British television produced
almost identical products for a great deal of the time. Thus at a randomly
selected time of the day the radio listener was
have two or three likely to
varied possibilities open to him, while the television likely to be Uewer was
forced to choose between two almost identical programmes.

The Principle of Minimum Differentiation

In some ways this radio example is a special case^ and we may now' consider
the problem in somewhat more general and abstract terms.
Let us consider first the simple case of a product with only one inde-
pendent characteristic w'hich can measure on a scale from — 10 to -MO.
w'e
This is illustrated in Figure 28.4. This might, for example, be a soap powder
in W'hich harshness was associated with cleansing pow'er and mildness with
lack of it; -|-10 might indicate a soap w'hich had great cleansing power but
removed the skin from the unfortunate housewife’s hands, and — 10 a soap
which was positively beneficial to the hands but w'hich would not remov'e the
merest speck of grease.^ Let us assume that firm A has settled its product
on the scale at —2. If firm B now' w'ishes to produce a competing product,
w'hat will be its optimal policy? It might go to an extreme, producing a
soap W'hich had strong cleansing power but rvas also rather harsh on the
hands, going out as far as, say, -t- 8 on the scale (indicated by B' in Figure
28.4). Now' consumers w'hose tastes lead them to prefer something betw'een
the Uvo products would have to decide beUveen them and, presumably,
they would choose the product w'hich came closest to satisfying their tastes.
Firm A would get all customers who preferred a product ranging from — 10
to -f 3, while B w'ould get all those who preferred a product in the range -i-3

to 4-10. Now' let us assume that, having decided to make a product with
1 The case is certainly sufficient to refute the proposition that some degree of competition
always produces more product differentiation than does monopoly. One must be very careful
not to dismiss refutations as merely special cases.
2 This product has nvo characteristics, cleansing power and harshness of effect on skin, but
they are not indepmdenl of each other; we have assumed that they vary directly with each other.
388 THE INTERMEDIATE THEORY OF SUPPLY

more cleansing power and more harshness than the competitors product
the producers ofB go onl> a little waj in this direction, just enough to make
the difTerence noticeable but not enough to cause a great gulf between the
twoproducls Letussa> they go to zero on our scale (see B' in Figure 28 4)
Now product B should get all customers who would like a product rating
between —1 and +10 Clearl> the optimal polic) for the producers of B is
to place their product to the right of Aon the scale, but onlj just enough for
the difference to be noticeable This fact, that the optimal policy is often to
make >our product different enough from that of your competitor’s product
for the difference to be noticeable but no more so, is sometimes referred to
as the PRINCIPLE of minimum differentiation

An example of oligopolistic complexity If we now add a third com

peting firm, any simple competitn c strategy svill lead to an unstable process

A B‘ B

1 1 I .
, ,

~IO -S 0 *i

Fig 28 4 The difterenuation of a product with one independent charactenstic

If, for example, A is at — 1 and B at zero then the optimal strategy for the
new firm produce a product rated at + But now firm B will haic
C is to 1

only 5 per cent of the market' (i e , those customers preferring a product

between — ^ and + and it %viH pay ii to move outside of either A or C
Say B goes to + 2 Now C has only a small pin of the market and it will
pay It to go to —2 putting A in ihc position of being bracketed bv the other
two so that It goes to +3 (or to —3) The student can be left to prove for
himself that if each firm moves in sequence and mikes the move which will
give It the biggest share of the market without worrying about what its
competitors might then do, an unstable situation will develop so that the
process ofmove and counter-move will not lead to an equilibnum position
In this example the addition ofa third firm removes the tendency towards
minimum differentiation and replaces it by a tendency towards perpetual
^
oscillation

to
1 ^^e are assuming thaC the custoram are dtstnbuted evenly along the «ale according
their tastes

2 As soon as we drop the assumption that each firm docs not look beyond its next move, a

host of inieresung possibibties are opened np The student should consider several possible
wh**
strategies (1) what happens if each firm looks ahead to its competitors next move’ (2)
happens if two firms look beyond their next move and one firm does not* i3) what happen* if
one firm looks ahead’ (4) what happens if two firms collude’ It is very interesting to note that
the two colluding firms can take the lionssbarcofthc market by locaUng at +5and —5 thus
ng the third firm unable to retaliate but that the third firm can, by locating anywhere
 MONOPOLY VERSUS COMPETITION 389

The Principle of Minimum Differentiation Once Again

If, however, we now add more characteristics to our product we shall find
that a stable tendency towards minimum differentiation will reassert itself.
Let us assume that the product has two independent characteristics, each of
which can be rated on a scale of — 10 to +10 as illustrated in Figure 28.5.
For example, the X axis
might indicate harshness and cleansing power as
before, while the Y
might indicate some quite different characteristic,
axis
say colour, which could range over the spectrum from red to purple. Now,
even if we have a fairly large number of different products, the best strategy

y
-+10

-+5

OB oA
Fig 28.5 The differentiation of • * A^ 0 1 "I" >
-10 -5 *>£ “ +5 +10
a product with two independent
OD oc
characteristics.

- --5

OF"
- --10

for the producers of each will be to keep their product fairly near to the
centre of our chart. Assume, for example, that there are four firms. A, B, C
and D, all differentiated as indicated in the figure. A fifth firm, E, now
decides to enter the market, producing a soap that rates minus X and
minus Y. If the firm goes to extremes on either X or Y, locating at E' or E",
it will not gain as many customers as when it comes quite close to firm D,
locating at, say, E. Again the optimal strategies will tend to produce a series
of products clustered about the origin, all only slightly differentiated from
each other. Thus consumers who wish a product near the centre of both
scales will have many similar products from which to choose, while con-
sumers who wish a product which is near the extreme of either or both
scales will find no product which comes close to meeting their desires.

between —4 and +4 determine how this share is split up between the two colluding partners.

All of this is a simple Illustration of the general point made inChapter 26 that the outcome
of an oligopolistic situation is influenced by the strategies adopted by the competing firms.
390 THE INTERMEDIATE THEORY OF SUPPLY

MONOPOLY VERSUS COMPETITION A CONCLUSION

We have been concerned in this chapter with what could be said about
monopoly and competition on the basis of positive economic theory On
many crucial points we have no accepted theory at all, and, on other points,
‘

existing theory has been imdcquaiely tested It is obviously necessary to

keep an open mind on the subject and to admit that, on the basis of existing
theory, it is impossible to make out an ovcnvhelming case either for or
against monopoly as compared with competition Everyone will have his
own guess, hunch, or prejudice on the subject, often based on bits of persona!
experience But from the point of view of economic science, wc are inter
ested in carefully documented, objective evidence, and on these grounds a
great deal remains to be discovered, even at a most elementary level, about
the comparison of the effects of monopoly with those of competition

1 S<e the hit of queitions on page 225 A casual reading of (he newspapers will raise many
questions in the field of the theory of the firm which cannot be handled by any exisungtheorv
CHAPTER 29

CRITICISMS AND TESTS OF

THE THEORY OF SUPPLY

In previous chapters we have deduced a number of testable implications of

the theory of supply. The theory is itself tested every time one of these
implications is confronted with facts, in such a ^vay that the discover)' of
conflicting et'idence is possible. In many cases, if a particular prediction
tvere refuted, this would necessitate only a minor change in the basic theor)',
so that it was once again consistent with the facts. We might, for example,
discover that firms did not always close dotvn when they were unable to
cover their variable costs of production. This is a refutation of the theory
of short-run profit maximisation as we have presented it; but only minor
changes might be required to make the theor)' consistent tvith these ne^v
facts.In other cases, however, empirical observations have been alleged that
strike at the very core of the theory of the firm. If these refutations tvere
substantiated it would be necessar)' either to make very drastic amendments

in, abandon completel)' the theory' as \s'e kno'w it.

or to
The theory of the firm is based on a number of critical assumptions. If
any of these do not conform to reality it is likely that the theoiy' twll make
incorrect predictions. The following are the three assumptions that critics
have thought most vulnerable and on which heav)' attacks have been
concentrated.
1 The assumption that the decision-takers have access to and use the
information that economic theory' assumes they use (see page 246).
2 The assumption that we can abstract from the indit'iduals who make
the decisions for the firm and the kind of organisation in which they work
(see page 226).
3 The assumption that decision-takers strive to maximise profits (see
page 227).
'
These assumptions are interrelated, for all concern the profit-maximisir-"
392 THE INTERMEDIATE THEORY OF SUPPLY
behaviour of the firm We shall, however, treat the criticisms of them
separately Before we do so, we shall digress for a moment to consider
different approaches to evaluating the theory We do so because in some
actual cases the data that have been used do not provide evidence relevant
to our theory

APPROACHES TO TESTING THE THEORY

Each of the following three approaches has often been used in criticising
and testing the theory of the firm
1 Formulate an alternative {and competing) theory that predicts different results

Given an alternative theory, one can discover the areas m which the two
theories make conflicting predictions and choose between the two theories
on the basis of which comes closer to predicting what actually is observed
to happen We might hypothesize, for example, that firms choose to maxi-
mise their sales rather than their profits, and we would then have two
competing theories We could then derive conflicting predictions from the
two theories and could confront these with the evidence This is a satisfactory
way of choosing between two theories
2 Observe decision-makers to see tf they behaie aj the theory predicts they will

behave We might observe, for example, how

a certain executive makes a
certain decision what records he consults, what questions he asks, and so
on Or we might create a laboratory situation and give ‘subjects’ a chance
to make decisions, then record and analyse their decisions
Although this approach may give rise to new theories, it docs not by itself
provide a test of an existing theory, since it does not tell us whether the
procedure actually employed by the decision taker really makes any difier-
*
ence in his decision
If, example, an executive systematically discusses proposed price
for
changes with his sales manager and his lawyer, but rarely with his cost
accountant, it may suggest the hypothesis that demand and anti monopoly
loom larger in his mind than cost conditions, but it does not
considerations
demonstrate that these things play a more important role than cost in
pnemg decisions ^
3 Ask decision takers how they take decisions Another approach to testing

1 One executive was observed to telephone hts wife and enquire whether there was a letter
from their married daughter before every meeting of hii investment board A theory that
Ignores such behaviour is not deficient in explaining that executive s investment decisions

2 The executive may be an expert on cost conditions, or he may need less time to acquaint
himself with cost data than nith demand data, or hu cost accountant may provide him i-i'h
lucid memos, whereas his sales manager can only communicate orally There is a host of other
lines
 CRITICISMS AND TESTS OF THE THEORY OF SUPPLY 393

the assumption of profit maximisation is merely to ask the businessman ‘Do ;

you seek to maximise profits ?’ Such an approach has irom time to time been
tried and the student wll not be too surprised to learn that, when asked if his
sole motive ivas to make as much money as possible, the businessman replied
that it was and that he sought to charge a fair price, to make only a
not,
reasonable profit, and generally to conduct his affairs in a manner con-
ducive to the social good. Asking people what they do and why they do it
may well pro\Tde some interesting hunches and suggest hypotheses about
behaviour for further testing. If you have always taken it for granted that
people do a certain thing and enquiry shows that everyone denies it, then
this may make you suspicious and lead you to check your ideas further. But
it can never show that your original idea was wrong. Consider what the

denials might mean: (1) the people were lying; they did try to do the thing
assumed but would not admit to it; (2) the people told what they thought
was the truth, but they were not aware of their own motives and actions;
(3) the people were correct in saying that they did not tiy to do it. Now
how are judge which of these possibilities is the correct one? One
we to
needs only a nodding acquaintance with elementary psycholog)' to realise
that we are not likely to discover veiy' much about human modvation by
asking a person what motivates him. Generally, he will have either no idea
at all, or else only a pleasantly acceptable rationalisation.
Direct questioning at best (assuming the subject tries to be scrupulously
honest) tells us what the person questioned thinks he is doing. Such in-
formation may can never refute an hypothesis about
be interesting, but it

what the person actually is doing. To refute such an hypothesis, we must

observe what he does, not ask him what he does.
The technique of direct questioning is, nevertheless, not without value.
It can definitely refute an hypothesis about w'hat people think they do. It

can make us sceptical about an hypothesis about w’hat people actually do,
and so lead us to make direct tests of this hypothesis. And it can suggest the
formulation of some new' hypothesis about w'hat people actually do. What it
cannot do is to provide evidence either in favour of, or against, an hypothesis about
what, in fact, people do.

CRITICISM 1; FIRMS DO NOT HAVE ADEQUATE

INFORMATION
One group of critics says that profit-maximising theoiy' w'ill prove in-
adequate because businessmen, however hard they may tiy', cannot reach
decisions the way the theory predicts. This criticism has a number of aspects
some of them very crude and some quite sophisticated.
One of the crudest forms of this criticism is based on the observ'ation that

13
394 THE INTERMEDIATE THEORY OF SUPPLY

businessmen do not calculate in the manner assumed by the theory Some-

times businessmen arc interviewed and it is discovered (apparently to the
surprise of the interviewer) that most businessmen have never heard of the
terms marginal cost and marginal revenue It is then argued that since the

theory assumes that businessmen equate marginal cost and marginal

revenue, and since empirical observations show that businessmen have never
heard of marginal cost and mai^inal revenue, the theory is refuted, because
businessmen cannot employ concepts of which they arc ignorant
The observation that businessmen are not aware of the concepts of
marginal cost and marginal revenue would refute the theory that business-
men make decisions by calculating marginal values and consciously
equating them But it does not refute the theory that businessmen make
decisions m
such a way as to maximise profits The economic theorist uses
and marginal revenue to dis-
the mathematical concepts of marginal cost
cover what will happen as longas, by one means or another - by guess,

hunch, clairvoyance, luck or good judgment - the businessman docs

approximately succeed in maximising his profits The constructs of the
theory of the firm are purely logical tools employed by the economist to
discover the consequences of certain behaviour patterns They are not

meant to be a description of how the businessman reaches his decisions If

he IS maximising then the tools of economic theory allow us to

his profits,

predict how he will react to certain changes - e g the introduction of a

,

tax - and lAis prediction is independent of the thought process by which the businiss

man actually reaches kis decision

An equally crude criticism stems from the observation that businessmen
do not calculate down to single units with such a nice degree of accuracy as
IS assumed in the theory In the verbal presentation of the theory to
elementary students, it is usually stated that the businessman will carry on
producing until the cost of produang the very last unit is just equal to the
revenue gained from its sale This is merely a verbal statement of the
mathematical conditions for the maximisation of profits The observation
that businessmen do not calculate down to single units, is not of itself

relevant as a test of the theory The marginal analysis allows us to predict

how the businessman will respond to certain changes in the data, if he is

maximising his profits he will be observed to respond in this wa> even

though he calculates m
a much cruder fashion than does the mathematician
More sophisticated critics point out that the information available to the

I It ihould be noted m
passing that the naiveti of such attacks is often surpassed by
the

naivete of the defences offeied tn iclum For cscamplt Benham defended profit maximising
theory against this particular attack by saying that the sue of the unit is undefined it might
^be thousands of tons or thousands of packages, when this w realised it is seen not to be silly to
jme that the businessman equates the cost and revenue of the marginal unit Both the attack
 CRITICISMS AND TESTS OF THE THEORY OF SUPPLY 395

producer is simply not adequate to permit him to reach the decisions that
the economist predicts he will make. This argument generally takes one of
three forms: that the businessman is the victim of his accountants, and
bases decisionson accounting concepts, which differ from economic ones;
that the natural lag between accumulatingand processing data is such that
important decisions must be made on fragmentary and partially out-of-date
information; and that, because acquiring full economic information is
costly, firms cannot afford to acquire as much information as economists
assume them to have.*

The hypothesis of full-cost pricing: Out of these lines of criticism

has come the hypothesis of full-cost pricing, which was originally suggested
by businessmen’s answers to questions on how they set prices.
This hypothesis explicitly denies that businessmen will charge the price
that will maximise their profits. According to the full-cost hypothesis,
businessmen use available data to compute full costs per unit (variable costs
plus overhead) and add to this a conventional markup; price is set at this
figure and sales are determined by what the market will absorb at that
price.^ The full-cost hypothesis portrays the businessman as a rather con-
servative creature, a prisoner of his habits and his accounting records,
instead of the alert profit-seeker of traditional theory.^ Thus firms are
conceived of as having a perfectly elastic supply curve set at the level of
‘full-costs’. The student should be able to verify for himself that the full-

cost theory leads to different predictions than does profit-maximising theory.

Although full-cost theory has occasioned considerable heated argument,
no generally accepted authoritative test of it exists. Insofar as the theory is
made to rest on the inadequacy of accounting records, it has been effectively

and defence suggest that marginal theory is meant as a description of how the businessman
takes his decisions. Strictly speaking Benham’s defence is wrong on its own grounds; the size
of the unit of calculation is defined in marginal theory, it is an infinitesimally small unit.
1 The growing importance of business consultants and of economic-research departments
within firms suggests that firms may not always have been successful in maximising profits and
that they are making improve performance in this direction. The evidence
serious efforts to
from such consultants suggests that the businessman aided by his accountant has often been
led to take many decisions which are not profit-maximising ones.
2 The theory may also be regarded as an attack on the desire of businessmen to maximise
profits. A paper by R.L.Hall and C.J. Hitch, ‘Price Theory and Business Behaviour’, Oxford

Economic Papers, May 1939, generated a long debate that is well summarised in R.A. Gordon,
‘Short-Period Price Determination in Theory and Practice’, American Economic Review, June
1948.
3 Many economists believe that the full-cost theorists discovered the rule of thumb by which
day-to-day decisions are made within the firm, but that the critical decision of what the
markup should be is made periodically by management at a high level with profit-maximisation
as an important objective.
396 THE INTERMEDIATE THEORY OF SUPPLY

refuted by showing that modem accounting procedures do not limit firms

to the use of a\ erage costs
^
But the belief that the theory is appropnate for
some firms cannot be refuted cither by showing that firms are not forced
to be full costers by modern accounting methods or by showing that some
firms do not choose to follow full cost methods
The fact that some economists continue to take the theory of full cost
pricing seriously, whereas others dismiss it as not even worth testing is

cogent argument for subjecting every theory to empirical tests no matter

how little confidence one has in it So long as theories are judged on the
basis of casual and private observation, there will be disagreement If how
ever the choice between the two theories is as clear as most economists seem
to think, a carefully documented set of tests should settle the matter once
and for ail
The procedure enough determine as many as possible
for testing is clear
of the predictions made by both the full cost and the profit maximising
theories take those cases in which the two theories predict different re
actions to the same event (e g , that a rise m demand will lead to a nse m
price profit maximisation that a nse m demand will leave pnee
unchanged - full cost) ,
and confront these predictions with empmeal evi

dence to see if one is contradicted by the facts and the other consistent with
them Such a task is of course, easier outlined than accomplished but it
IS interesting in view of the heated controversy over this theory, that so few

attempts have been made to test it systematically^

CRITICISM 2 DECISIONS DEPEND ON WHO MAKES

THEM
A major attack on the theory of the firm as we have developed it comes
from a group of economists whose central concern is with what is called

ORGANISATION THEORY In terms borrowed from social psychology, they

^

argue that m
big organisations decisions are made after much discussion by
groups and committees and that the structure of the process affects the
substance of the decisions Their central conviction is that different decisions
will result from differcut kinds, of ocgawsatious, even if all else is unchanged

Recent Developments in Cost Account ng and

the
1 See for example James S Eartcy
Marginal Analysis Journal of PohUtal Ecenemj June 1955
Rostas
2 Such evidence as there IS seems togoagainstthefuU cost theory See particularly L
Productivity Prices and Distnbution in Selected Bntish Industries MESR Occasional Popir
stay
{Cambr dge University Press 1948) The evidence is quite strong that markups do not
constant over time
3 The interested reader is referred to R M Cyert and J C March A Behavioral Thiers ej
V Firm (Prentice Hall 1963)
 CRITICISMS AND TESTS OF THE THEORY OF SUPPLY 397

Although it has proved easier for organisation theorists to express their

central point of view than to formulate specific testable hypotheses, they
have formulated a number of the latter. One is that a large and diffuse or-
ganisation finds it necessary to develop standard operating procedures to
help it in making decisions. These decision rules arise as a compromise
among competing points of view and, once adopted, are changed only re-
luctantly. One prediction following from this h>'pothesis is that the proce-
dures used may be nonoptimal and will persist for long periods of time,
despite changes in conditions affecting the firm. For either reason, profits
will not be maximised. Another prediction is that this procedure v\t11 lead
large firms to adopt conservative policies and avoid large risks. Smaller
firms will take bigger risks.
Hypotheses like these are very hard to
test. Proponents of organisation

theory' feel that the evidence supportsthem; critics feel that they are un-
demonstrated. It is hard to avoid the view that at the present time the
evidence is inconclusive. This does not mean that it is necessary' to reject
the attack of the organisation theorists; rather that it is necessary' to reseiv'e
judgment until a sharper statement of conflicting predictions is made and
until such predictions are more fully tested.

CRITICISM 3: FIRMS DO NOT SEEK TO MAXIMISE

SHORT-RUN PROFITS
A third kind of criticism, often made by the same critics, as follows; is

businessmen do not seek maximise profits at

to must
all. Of course, they
make some profits or they ivill go out of business. But once some minimum
level of profits has been achieved, they pursue totally different goals. The
actions necessary' to achieve these other goals are substantially' different
from those necessary' to achieve profit-maximisation. For this reason, all the
deductions based on the assumption of profit-maximisation will be at vari-
ance with the facts, except in those cases in which the actions necessary to
achieve the goals actually' being pursued happen, by chance, to coincide
with the actions necessary' for profit-maximisation.
This is stronger than say'ing that firms are prevented from achieving
maximum profits by the lack of information available to businessmen or by'

the organisation’s structure. This hypothesis states that businessmen do not

wish to maximise profits.

Satisficing Organisation theorists have criticised the profit-maximising

;

model and have suggested an alternative that they' call satisficing. Professor
Herbert Simon says, ‘We must expect the firm’s goals to be not maximising
profits but attaining a certain level or rate of profit, holding a certain share
398 THE INTERMEDIATE THEORY OF SUPPLY

of the market or a certain level of sales According to this theory, firms

will strive very hard to achieve certain minimum (or ‘target’) levels of profits
but, having achieved them, they will not strive aggressively to further
improve their position This means that the firm could come to rest m a
large number of situations rather than in only one unique situation (the
profit maximising one) In the language of economic theory, we say that
equilibrium is not unique

To test the satisficing theory, one must specify the ‘targets’ of the firm

This has not been done carefully enough to permit us to specify the precise
areas of conflict between satisficing and profit maximising theory Until we
know what predictions of the two theories are in conflict with each
precisely
other, wc do not know to what extent the theories differ and thus we cannot
test them to see which is more consistent without empirical observations

The satisficing theory is potentially an important alternative to the profit

maximising theory Perhaps Simon exaggerates when he says, ‘The satis
fiemg model vitiates all the conclusions about resource allocation that are
derivable from the maximising model when perfect competition is assumed
but very probably it does lead to difTerenccs in expected behaviour One
case that cued by proponents of satisficing is that, immediately after the
is

Second World War in the US, Canada and the UK, prices of new cars
were lower than pnees of used cars As viewed by the proponents of satis
ficing, manufacturers of automobiles were satisfied with (even possibly
embarrassed by) their high profits and were content not to take advantage
of the excess demand to increase profits As is so often the case, however,
there are other possible explanations of this behaviour, including fear of
anti-monopoiy action It cannot, however, be explained by the simple

theory that the sole motive of the firm is to maximise its short run profits

Sales-maximisation Another theory recently put forward is that firms

seek to maximise not their profits but their sales revenue Firms, it is assumed,

wish to be as large as possible and, faced with a choice between profits and
sales, would choose to increase their sales rather than their profits

The basic argument offered to make this theory seem plausible is a

modern version of the separation of management and ownership In the
giant company, the managers need to make some minimum level of profits
to keep the shareholders satisfied, after that they are free to seek growth
unhampered by profit considerations This is a sensible policy on the part

1 H A Simon Theories of Decision Making in Economics Amincm Economic Rtinew Jun*

1959 page 263

2 Ibid page 265
3 See VV J Baumol Business Bekastettr lo/w aatf CretcM (Macmillan 1959)
4 See Chapter 18 The hypothesis suggests that managers have the power to pursue ihcir
own goals rather than the goals of the stockholders
 CRITICISMS AND TESTS OF THE THEORY OF SUPPLY 399

of management, so the argument runs, because salary, power and prestige

all vary with the size of a firm as well as with its profits; generally, the
manager of a large, normally profitable company will earn a salary con-
siderably higher than that earned by the manager of a small but highly
profitable company. Thus we may assume that firms seek to maximise their
sales revenue, subject to some minimum-profit constraint. This theory is a
genuine alternative to the theory of profit-maximisation, because it leads to
different predictions.* It has not yet been carefully tested.

Long-run profit-maximisation ; In order to take account of various

criticisms of the assumption of profit-maximisation, some economists modify
profit-maximisation to mean ‘long-run profit-maximisation’. In this view,
for example, sales-maximisation is long-run profit-maximisation. Sales are
the key to growth, and growth to future profits. Thus, maximising current
is the correct long-run policy for maximising profits. In much the same
sales
way, the ‘long-run’ approach can be used to explain other facts that are
disturbing to believers in short-run profit-maximisadon. Why, for example,
did automobile companies in 1946 keep prices down? They did so in order
to keep the ‘good will’ of automobile buyers, a policy worthwhile in terms
of long-run profits. Or, for another example, a firm may be right to avoid
risky ventures, even if they promise large short-run profits, because the
surest way to long-run profits is to survive in the short run. Long-run pro-
fitability requires survival, and survival requires caution.
Is long-run profit-maximisation the answer? There is little doubt that
when the theory of short-run maximisation in a perfectly competitive market
is applied to a market in which each producer sets his own price, extensive
modifications in the theory must be made. In particular, prices can and do
change daily in perfectly competitive markets, but they cannot and do not
in markets in which the producer sets his own price. Thus, for these latter
markets, we cannot say that the producer maximises profits from day to day
as costs and demands change. We must specify a realistic period over which
he makes his decision. This might lead us to say: ‘The producer does not
change price from day to day to maximise profits in the short run; rather
he makes decisions over a long period and seeks to maximise his profits in
the long run’. It is, however, exceptionally difficult to give such a long-run
theory any testable content. If we are not careful, we may find ourselves
rationalising, whenev^er we find a firm not maximising profits, by saying

1 For example; a firm with substantial monopoly power will tend to charge a price where
the elasticity of demand is unity (sales-maximising theory) or where it is greater than unity

(profit-maximising theory). Observations of industries where elasticity of demand tends to be

consistently above unity, but in which the firms continue to earn profits above the required
minimum, cast doubt upon the sales-maximising hypothesis.
400 THE INTERMEDIATE THEORY OF SUPPLY

merely that it was maximising over some other tme period than the one we
were considering Unless we work out our theory carefully and include in
It a means of identifying the long nm
period over which profits are supposed
to be maximised, we shall have a universal untestabk alibi for all refutations

of profit maximising theorj If this happens, then our theory becomes con
sistent with absolutely any behaviour on the part of businessmen and
becomes, as a result, totally uninteresting

CRITICISM 4 A CRITICISM BASED ON THE OBSERVED

SHAPE OF SHORT RUN COST CURVE
In Chapter 20, we argued that the marginal cost curve would be U shaped
falling at first as a more effiaent combination of the fixed and variable factor
became possible, and then ruing as diminishing returns set m Empincal
evidence has several times suggested that many marginal cost curves were
flat up to the capacity level of output, so that each extra unit would cost as

much produce as did the previous unit, until the plant was operated at
to
capacity, after which costs would begin to rise
The economic theorist should ask three questions when faced with such
evidence about the shape of the marginal cost curves 1 Is the evidence
reliable ’ 2 \\ hat part of my
it refute ’ and 3 Does this upset
theory does
any important predictions that have previously relied on’ We shall not
I

consider question here, but see what follows if we assume the answer to
I

be yes A careful look at the theory of costs shows that the declining part of
the marginal cost curve occurs if the fixed factor is perfectly indivisible The
argument (see pages 268-72) runs in terms of having too low a ratio of the

variable factor to the fixed factor As production is increased more of the

variable factor is used and a better combination with the fixed factor

achieved This argument clearly implies that all of the fixed factor must be
used all of the time If the fixed factor is divisible so that part of it may go
unemployed, then there is no need as production is decreased, to depart
from the optimum ratio of the quantity of the vanable factor actually em
ployed to the quantity of the fixed factor actually employed Thus costs *

would be constant up to the point at which all the fixed factor was used
Be>ond this point, production could only be increased by combining more
of the vanable factor with the constant (total) amount of the fixed factor

Under these circumstances costs would be nsmg Consider as a simple

example a factory’ of ten sewing machines in a shed, each
which consists
with a capacity of 20 units per day when each is operated for 8 hours by
one girl If 200 units per day are rcqmred, then all ten machines are operated
on a normal shift If demand falls to 180 units then one girl is laid off But
1 See note I page 272
 CRITICISMS AND TESTS OF THE THEORY OF SUPPLY 401

there is no need to combine the nine girls with ten machines. Clearly
one machine can be ‘laid off’ as well and, in production, the ratio of
labour/machines is not varied. Clearly, production can go from 1 to 200
wdthout any change in factor proportions. In this case we would expect the
factory to have constant marginal costs up to 200 units, and only then to
encounter rising costs, as production was expanded, by means of ov'ertime
and other methods of combining more labour with the ten machines.
Thus the first answer that the theorist gives is that constant marginal costs
do not refute any part of the theory of costs, providing the fixed factor is
divisible. Constant marginal costs with an indivisible fixed factor would,
ho\vever, refute the theor>' of costs which predicts that marginal and average costs
must vary when factor proportions vary.
To deal Avith the final question, the economist places the new cost curve
(flat up and then rising) into his theoretical models in place of
to capacity
the former U-shaped one. Here the reader can be left the exercise of show-
ing that few if any of the predictions of the theory of the firm and industry'
(such as those developed in Chapters 27 and 28) are affected if a flat section
of the marginal and average short-run cost curves is substituted for the
declining sections.
We conclude, therefore, that the obser\’ations, if true, are not radically
upsetting to the traditional theory of the firm.

THE GENERAL DEMAND AND SUPPLY THEORY

OF A MARKET ECONOMY
In Chapter 5 we noted that there was evidence that markets for agricultural
goods and other primary products did function as assumed in the theory' of
demand and supply. We also stated that the generalisation of this demand-
and-supply theory' to a theory of the w'hole economy was a rather speculative
leap in the dark. We have now studied the theory of production sufficiently
to realise that we cannot in fact apply our simple demand-and-supply
theory to the whole economy. This theory is in fact a theory of competitive
markets, i.e., markets in which there is a large number of buyers and sellers.
Most manufactured goods, however, are produced under conditions of
oligopoly in which there are a very few firms. These firms may or may not
compete actively with each other, and which they do compete
in cases in
they are quite likely to keep prices fixed and to compete in terms of a host

of other factors such as serxdces, delivery dates, quality, and special features
of the product.
There are a number of factors to be noted about such a situation. First,

price often not the key variable which equates demand and supply; price
is

is often held quite rigid over long periods of time, while the key variables
402 THE INTERMEDIATE THEORY OP SUPPLY

become ones listed above We have as yet no agreed

factors such as the
theory about these other factors , certainly our theory of^f« docs not help us
in studying them Second, since most of the situations within manufacturing
are oligopolistic,we do not know how to expect the firms to react to such
influences as changes in demand, profits or taxes The simple competitive
theory of price and of the allocation of resources docs not tell us how re
sources are allocated in an oligopolistic economy We cannot even assert of
such an economy, that a nsc in demand will lead to a rise in price, or that an
increase m profits in some industry will lead to an increase in the resources
allocated to production in that industry (It may be so but there is nothing
in our thcoiy that predicts that it is necessarily so )
still wish to hold that our demand and supply theory does describe
If we
the working of free market societies of the sort actually found in today’s
world, then we must fall back on a somewhat vague formulation We might
say that the compctiuve theory will describe more or less what happens m
all markets because there some degree of competition everywhere and
is

especially in the long run, the actual economy will respond m somewhat the
same way as would a perfectly competitive one The trouble with such a
theory is that it is is difficult to see what the theory
rather too vague It
actually tells us We might wonder what observations would conflict with
It If the answer is that the theory could not be refuted because it is too

vague, then, by the same token the theory is too vague to be of any real use
for It tells no conceivable observation will
us nothing about the world if

refute it, then the theory is compatible with any conceivable observation
we might make in the economy
These comments are not meant to imply that the theory developed so far
is completely useless Some sectors ol the economy do come close to
fulfilling

the conditions assumed in the iheoncs of perfect competition, monopolistic

competition and monopoly For these sectors, our simple theories will be
successful in predicting the outcome of many of the changes m which we
may be interested Many examples of especially useful predictions have been
given in the preceding chapters What wc are casting doubt on here is the
idea thatwe have a theory that is satisfactory m explaining and predicting
behaviour throughout the entire economy To do this wc will need a sue
cessfuf breakthrough m the theory of oligopoly probably afong the
finer

suggested in Chapter 26
 PART 5

DISTRIBUTION
CHAPTER 30

A GENERAL VIEW OF THE

THEORY OF DISTRIBUTION

In Chapter 4 we raised a number of general problems that confront all

economies. We now come to the question of what determines the distribu-
tion of the national product amongst the various individuals and classes
that compose the society. In this chapter we shall give a general survey of
the theory of distribution which will be developed in detail in the rest of
Part V.

DISTRIBUTION; THE FACTUAL BACKGROUND

Income many forms wages and salaries, rental income from property,
takes :

interestand profits, to name the major ones. Table 30.1 shows the distri-
bution of income in the United Kingdom by major types, for 1964. This

Table 30.1

FUNCTIONAL DISTRIBUTION OF INCOME IN THE

UNITED KINGDOM, 1964

Percentage of
Type of income Millions of pounds
total

Wages and salaries 64

Rent, dividends, interest 11

Proprietors’ income 2,220 9

Retained profits 3,139 12
Other 853 4

Total 25,329 100

j
406 DISTRIBUTION

SO called functional distribution of income is the income paid to

factors of production, rather than the income received by persons who pro
vide the factor services A single individual may receive incomes of several
sorts - from his labour, from the use of his savings, from the rental of
property he owns and from his investments m shares of stock The amount
of his income is merely the sum of his income from each source
Much of the theory of distribution that we discuss in this part is concerned
with the functional distribution of income But economists are also interested
m the equality or inequality in the distnbution of income among persons
and families Many economic policies are designed to modify income

distribution

Table 30 2

INCOME OF BRITISH PERSONS 1960

Percentage of

Income m £j ^

earning persons

0-400 !
39
400-800 47
800-1,500 10
1,500-2,000 1 45
2 000-3,000 08
3,000-5,000 1

05
5 000 upwards 0 25

Source Annual abstract of Siatistics Board

of Inbnd Revenue

The basic facts about the distnbution of income arc given m Tables 30 2
and 30 3 Table 30 2 shows the distribution of British families by income
level Everyone knows that some families are richer than others and that
some families arc \erv poor indeed Median personal income in 1960 was
about £600, but more than a quarter of British people earned less than
£400 Table 30 3 focuses on this inequality the 20 per cent of the popu
the
lation at the bottom of the income scale receive only 6 per cent of
nation’s income the 20 per cent at the top receive 33 per cent of it
Econo

mists want to know why

It is tempting to give superhaal explanations of differences in
income
People often say A man is paid what he’s worth But the economist must ’

His wife, his

ask ‘"Worth what to whom^’, *What gives him his valuc^
people
friends and his employer give different answers Sometimes
may all
/ ‘Men earn according to their ability ’ But note that incomes are distn
 A GENERAL VIEW OF THE THEORY OF DISTRIBUTION 407

buted in a very much more unequal fashion than any measured index of
ability, be it I.Q,, physical strength, typing skill or the quality of the books
one writes. In what sense is John Lennon 20 times as able as the promising
new pop singer? He gets paid 20 times as much. In what sense is a lorry
driver more able than a schoolteacher? In what sense is a football player
more able than a wrestler?
If answerscouched in terms of worth and ability are easily refuted, so are
answers like a matter of luck’, or ‘It’s just the system’. We are con-
‘It’s all

cerned now to discover whether or not the theories of economics provide

explanations of the distribution of income that are more satisfactory than

Table 30.3
INEQ.UALITY IN INCOME DISTRIBUTION, 1963
1

Group in population Percentage of income

Lowest 20 per cent 6

Lowest 40 per cent !
27
Lowest 60 per cent ,
43
Lowest 80 per cent 66
Lowest 90 per cent 79
Lowest 100 per cent 100
!1

the ones mentioned above. The predictions of economics, and the hypotheses
they come from, rest upon theories we have already studied. In this chapter,
we present a general survey of the theory of distribution, which we shall
develop in detail in subsequent chapters.

FACTOR PRICES AND FACTOR INCOMES

The traditional theory of distribution states that distribution is simply a
special case of price theory. The income of any factor of production (and
hence the amount of the national product that it is able to command)
depends on the price that is paid for the factor and the amount that is used.
If we wish to build up a theory' of distribution we thus need a theory of
factor prices and quantities. Such a theory is a special case of the theory of
price little that is not already familiar.
and involves
The free-market price of any commodity is determined by demand and
supply. In Figure 30.1 we draw the demand and supply curves for some
factor of production. The equilibrium price is Oa and quantity Ob.
The
408 DISTRIBUTION

total income earned by the factor is ^Oaeb If we now assume that the
prices of all other factors of production, the prices of all goods, and the level

of national income are given and constant, then fluctuations m Oacb will
be associated with fluctuations in the relative earnings {compared to other
factors) and in the share of national income going to the factor Assume, for
example, that the demand curve for the factor in question rises from D to
Di Now the money price of the factor rises from Oa to Od, and the relative
price rises from OajF to Orf/F where f is the (given) price of some other
factor The total earnings rise from Oacb to Odje and, if the total income in
the whole economy remains constant at Y then the share of income going
to this factor rises from OacblY to OdfelY Thus the problem of distribution
\n a free market reduces to the question of the determinants of the demand
and supply of factors of production There is then the problem of what is

Fig 30 I

the effcci of the various departures from a free market caused by monopo
listic organisations, government action, unions, etc

THE DEMAND FOR A FACTOR IS A DERIVED DEMAND

In considering the demand for a factor of production, the most important
thing to remember is that the purchasers of factors (land, labour, raw
materials, machines, etc ), do not want these factors for themselves but onl)
because they can help to produce other commodities that are wanted by
consumers Consider a factor of production, pig feed, whose only use is to
help m
the production of pork If there is no consumer demand for
pork,

then there will be no demand for the factor, if there is a very large demand
for pork, then there will be a large demand for the factor If a rise m the
nnee of pork causes a great reduction in its demand, then there will be a
?reat reduction m the demand for the factor, if a nse in the price of
pork
fall
s only a small fall off in its demand then there will only be a small
 :

A GENERAL VIEW OF THE THEORY OF DISTRIBUTION 409

off in the demand for the factor. Clearly, the demand for a factor of produc-
tiondepends on the demand for the consumption goods which it helps to
make. We say the demand for the factor is derived from the demand for the
consumption good, or, in the terminology of the economist, the demand for
a factor of production is a derived demand.

THE ELASTICITY OF DEMAND FOR A FACTOR

We may now how the demand for a factor will change as its price
ask
changes. Will the demand fall greatly if price rises or will it hardly fall at
all ? The answer to this question is implicit in the idea of derived demand

and we may state it in terms of three hypotheses.

1 The ELASTICITY OF demand for a factor varies directly with

THE ELASTICITY OF DEMAND FOR THE FINAL PRODUCT; Consider what
happens if there is a rise in the price of a factor of production. This raises
the cost of producing the final product, and will cause an increase in the
price of the product. If the price rise causes a large reduction in demand,
i.e., the demand is ver>' elastic, then there will be a large reduction in the
quantity of the factor now needed. If, on the other hand, the price rise
causes only a small reduction in the demand for the final good (i.e., the
demand is inelastic) then there will only be a small reduction in the quantity
of the factor now required. We may summarise as follows: (i) the rise in the
price of the factor causes a rise in the price of the final good; (ii) the rise in
good causes a reduction in the quantity demanded, the
the price of the final
amount of the reduction depending on the elasticity of demand for the final
good; (iii) the reduction in demand, and hence in production, causes a
reduction in the demand for the factor, the amount depending on the
amount of the reduction in demand for the final good. Thus, the more
elastic is the demand for a final good the more elastic will be the demand for
the factors that go to make it.

2 The smaller the cost of a given factor as a proportion

is

OF THE total COST OF THE FINAL GOOD THE MORE INELASTIC WILL
BE THE DEMAND FOR THE FACTOR: This, as has often been pointed out,
illustrates the importance of being unimportant. Consider two factors
factor A makes up 50 per cent of the total cost of commodity X while factor
B makes up only 10 per cent. Now a 10 per cent increase in the price of
factor A would raise the cost of producing X
by 5 per cent, while a 10 per
cent increase in the price of factor B would raise the cost of by only 1 per X
cent. Thus a10 per cent increase in factor A’s price would occasion a larger
increase in the price of X, hence a larger reduction in demand, first for X
 A GENERAL VIEW OF THE THEORY OF DISTRIBUTION 411

the demand and we must now consider the supply. For the
for a factor,
case of supply we have two
separate problems: the supply of some factor
to the whole economy and the supply to a particular industry. First, we
may ask what causes variations in the total supply of a factor of production
to the whole economy? This is a very difficult question, in part because the
several factors of production are so different from one another. Labour in
particular, very different from other factors of production. We are a
is

society of men, not of machines. The supply of engineers is very much

affected by the educational policies of the nation’s universities. The over-all
supply of labour depends not only on the size and age distribution of the
population, but also on those customs and institutions of a society that
determine when children should leave school and enter the labour force
and when retirement should take place. The supply of coal is determined
by a wholly different set of considerations, and the supply of machines by
yet a different one. (We discuss the determinants of the total supply of
particular factors in Chapter 32.)
The adjustments of supply to changes in the conditions of particular
industries are easier to deal with. The supply of a factor to one particular industry
is predicted to depend on the price paid by that industry relative to the price paid by
other industries that operate in the same factor market.
Supplies of the factor can be expected to move between industries,

leaving those offering lower earnings and moving to those offering higher
earnings. Similar movements of factors may occur between geographical
areas, although in the case of labour, as we shall see, the movement ipay be
very slow.
The hypothesis that supply will be an increasing function of the price of
the factor leads to a rising supply cur\'e, such as the one depicted in
Figure 30.1.

DISTRIBUTION THEORY AND RESOURCE ALLOCATION

Factor prices play an important role in the distribution of the national
product. In a competitive factor market, factor prices will be determined
by the joint forces of demand and supply. Since the demand for factors is a
derived demand, the theor)' of factor prices provides a direct link between
the allocation of resources (factors) and the demand for goods and services.
The link depends upon the movement of factors between occupations and
industries (and areas of the country as well) in response to differences in
factor prices. theory, which is primarily a theory about the
Distribution
relative earnings of factors in different uses, is a key part of the theory of
the allocation of scarce factors between alternative uses.*
Chapter 5 even if it is for the second
1 At this stage it would be a very good idea to re-read

or third time.
410 DISTRIBUTION

and then for A, than would a 10 per cent increase in the pnee of factor B
In general, the smaller is the proportion of total costs of producing A
that
are made up by the cost of one factor, the more inelastic will be the demand
for that factor

3 The demand for a factor will be more elastic the easier

IT ISTO SUBSTITUTE SOME OTHER FACTOR FOR IT IN PRODUCTION ISTlCn
the price of some factor rises, that factor becomes more expensne relati\e
to all other factors There will be a tendency for firms to save on costs by
using less of the now more expensive factor and more of the other facton
How easy it is to substitute one for another depends on the technical con
ditions of production It is important to emphasise that it is very easy to
underestimate the degree to which factors can be substituted one for another
in the production of some commodity It is fairly obvious that a bushel of

wheat can be produced by combining land either with a lot of labour and a
little capital, or with a little labour and a lot of capital When, he comes to

manufactured goods, however, the student often tends to think m terms of

using inputs in pretty well fixed proportions A bit of casual observation of

an industry over time will show just how factor proportions can be vanedto
produce a given product There is, for example, the case m which glass and
steel turn out to be very good substitutes for each other One would neier

guess this by considering their physical qualities m

general, but in the case
of car manufacture one can be substituted for the other over a wide range
merely by varying the dimensions of the windows

This analysis gives some idea of the determinants of the elasticity of demand
for a factor of production It is easy enough to understand and to commit
to memory these general principles It is also equally easy to forget to apply
them as soon as one is faced with a practical application How often docs
one hear something like the following A high proportion of the UK, s im
them
ports are raw materials these imports are absolute nectssiUes for without
factones would grind to a halt, therefore the demand for these imports will
be completely inelastic ^ The principles ofdenvcd demand have, ofcoune,
been forgotten in making such a statement Consider the effect of an in
crease m
the price of an imported raw material which is a major cost in
the

production ofa commodity with a fairly elastic demand Clearly the assump- ,

tion, even as a first approximation, that the elasticity of demand for every
imported raw material is zero is likely to be misleading

SUPPLY OF FACTORS
Ue have said that the traditional answer to the question of distribution
is

that It IS all a matter of demand and supply We have already considered

 A GENERAL VIEW OF THE THEORY OF DISTRIBUTION 411

the demand and we must now consider the supply. For the
for a factor,
case of supply we have two
separate problems; the supply of some factor
to the whole economy and the supply to a particular industry. First, we
may ask what causes variations in the total supply of a factor of production
to the whole economy? This is a very difficult question, in part because the
several factors of production are so different from one another. Labour in
particular, is very different from other factors of production. We are a
society of men, not of machines. The supply of engineers is very much
affectedby the educational policies of the nation’s universities. The over-all
supply of labour depends not only on the size and age distribution of the
population, but also on those customs and institutions of a society that
determine when children should leave school and enter the labour force
and when redrement should take place. The supply of coal is determined
by a wholly different set of considerations, and the supply of machines by
yet a different one. (We discuss the determinants of the total supply of
particular factors in Chapter 32.)
The adjustments of supply to changes in the conditions of particular
industries are easier to deal with. The supply of a factor to one particular industry
is predicted to depend on the price paid by that industry relative to the price paid by
other industries that operate in the same factor market.
Supplies of the factor can be expected to move between industries,
leaving those offering lower earnings and moving to those offering higher
earnings. Similar movements of factors may occur between geographical
areas, although in the case of labour, as we shall see, the movement ipay be
very slow.
hypothesis that supply will be an increasing function of the price of
The
the factor leads to a rising supply curv'e, such as the one depicted in
Figure 30.1.

DISTRIBUTION THEORY AND RESOURCE ALLOCATION

Factor prices play an important role in the distribution of the national
product. In a competitive factor market, factor prices will be determined
by the joint forces of demand and supply. Since the demand for factors is a
derived demand, the theory of factor prices provides a direct link between
the allocation of resources (factors) and the demand for goods and services.
The link depends upon the movement of factors between occupations and
industries (and areas of the countr)' as well) in response to differences in
factor theory, which is primarily a theory' about the
prices. Distribution
relative earnings of factors in different uses, is a key part of the theory of
the allocation of scarce factors between alternative uses.
Chapter 5 e\’en if it is for the second
1 At this stage it would be a very good idea to re-read

or third time.
412 DISTRIBUTION

A DIGRESSION ON DISTINCTIONS OF KIND AND OF

DEGREE
It IS \cry common to think in terms of distinctions of kind - things are
iilherblack or white, either all one thing or all the other, rather than in terms
of distinctions of degree - things arc various shades of grey, a bit more of this
or a bit less of that
The example of the UK’s impiorts given at the end of the section on the
demand for factors of production is a case in point The argument is that
imports are absolute necessities, a 10 per cent reduction in imports would
spell disaster,and because of this they will, of course, have perfectly in
elastic demands Another example is provided by the common habit of

thinking m terms of the dichotomy necessities, which have perfectly

inelastic demands, and luxuries, which have highly elastic demands Every
shred of empirical evidence ever collected refutes this crude notion De
mand elasticities vary throughout the whole possible range with very few to
be found at either extreme Elasticity is to a great extent dependent on the
degree of substitutability of om commodity for another
^

Another common fallacy of this type is to think in terms of economies

which are either socialist, in the sense that everything is decided by the
central authority, or free enterprise, in the sense that everything is left to

free competitive markets Of course neither the completely controlled nor

the completely free economy has ever existed Economies differ one from
the other according to the degree of control exerted by the central authonties
Furthermore, this degree vanes widely from market to market within one
country For example, it is probably true that the degree of government
control of some agricultural products is less in the USSR than it is in the

USA"
Another example of this same tendency to think m terms of extremes
is

provided by the discussions of the importance of the Suez Canal at the time
of Its nationalisation by President Nasser of Egypt m
1956 It was commonly
said that the Suez Canal was Bntain’s lifeline Without it we would suffer
slow strangulation, our industries would grind to a halt and the end would
be at hand It is seldom, when presented with some view, that the economist

can reply with such confidence as he can in this case Utter nonsense' In the
of
event, we were deprived of use of the canal for many months and none
per
these dire consequences ensued The cost of certain imports rose a few
industrial
cent, as higher costs of transport round the Cape were incurred ,

production was hardly affected, and, presented today with an output senes
for almost any product, an economist would be hard put to locate the Su«

1 See page 233

2 See page 79
 A GENERAL VIEW OF THE THEORY OF DISTRIBUTION 413

crisisby means of changes in the series. Seldom has a popular view been so
immediately tested and so conclusively refuted.
One of the dangers of such an extreme view is that it encourages the
response ‘The Canal must be kept open at any price.' Such a view encourages
:

extreme behawour without any calculation of the consequences. Whenever

the student sees something advocated as necessary whatever the price, all his
critical faculties should be aroused. This phrase is almost always a cover-up

for sloppy thinking and unwillingness actually to calculate the price. There
are few objectives which anyone feels are worth hav'ing at any price (what
if the price is destruction of 50 per cent of the world’s population or im-

poverishment of everyone?). In fact, for almost every objective one can

think of, there are some costs which would make the objective worth pursu-
ing and some (higher) costs which would make it silly to pursue the
objective.
CHAPTER 31

THE DEMAND FOR FACTORS:

MARGINAL PRODUCTIVITY
THEORY^

This chapter is principally concerned with establishing the prediction that

the demand cur\e for a factor is dowmsard-sloping The reason we wish to
derive this prediction rather than treating it as an onginal hypothesis is that

we wish to preserve the link that demed demand provides between pncing
of factors and the pncing of products We want therefore to link our theory
of the behaviour of the firm to our theory of distribution

PROFIT-MAXIMISATION AN ALTERNATIVE STATEMENT

In Part IV we considered (he conditions for maximising profits in the short
run Some factor was fixed (usually capital) and some other factor was
allowed to vary, and we saw that the profit-maximising firm would increase
Its output to the level at which the hst unit produced added just as much
to cost as it did to revenue or, in technical language, until marginal cost

equalled marginal revenue Another nay of staling exactly Ike same thing is to lay

that the firm ittll increase production up to the point at uhich the last unit of the

lanable factor employed adds just as much to reienue as it does to cost Just as it is

true that all profit-maximising firms, whether they are selling under con
ditions of perfect competition, monopolistic competition or monopol>, pro-
duce to the point at which marginal cost equals marginal revenue, so is
It true that all profit maximising firms will hire units of the variable factor
up to the point at which the maiginal cojt<<Tthc factor (i e the addition ,

to the totil cost resulting from the employment of one more unit) equals the
Thu The student who is vsith econom cs
1 IS a difficult chapter not intending to go further
need not read all of It He must however go to the point of seeing why the price ofa ftciQt
ot
will be equated with its marginal revenue product Throughout this chapter it w assumed
imphcity that there is only one variable (actor
 ::

THE DEMAND FOR FACTORS 415

marginal revenue produced by the factor. Since rve have already used the
term marginal revenue to refer to the change in revenue resulting from the
sale of an additional unit of production, we shall use another term,
MARGINAL REVENUE PRODUCT to refer to the addition to revenue resulting
from the sale of the product contributed by an additional unit of the variable
factor. It is true, therefore, of all profit-maximising firms that in equilibrium

Marginal cost of the = marginal revenue product

variable factor of that factor (1)

If the firm is unable to influence the price of the variable factor by buying
more or less of it (i.e., the firm buys its factors in a perfect market) then the
marginal cost of the factor is merely its price.

Fig 31.1 A marginal revenue

product curve for a factor.

0 10 20 30 40 50 60 70 80 PO 100
Number of workers

The cost, for man on the payroll is the

example, of obtaining an extra
extra wage that must be paid for that man. For firms that take factor prices
as given we may now state the condition of ( 1 ) above in the follosving form

Price of the factor = marginal revenue product of the factor (2)

THE DEMAND FOR A FACTOR

Condition allows us to derive the demand cun'e for a factor as soon as
(2)
%ve have its marginal revenue product cur\'e. As an example we assume in
Figure 31.1 that we have the marginal revenue product curve for some
factor showing us bow much would be added to revenue by employing one
more unit of the factor for each level of total employment of the factor.
Condition (2) states that the profit-maximising firm udll employ additional
units of the factor up which the marginal revenue product
to the point at

equals the price of the factor. example, the price were

If, for 1,200 per
year, then it \vould be most profitable to employ 50 workers. (There is no
point in emplopng a fifty-first since he would add only ^^1,180 to revenue
410 DISTRIBUTION

but ;CI,200 to costs, and hence he would reduce total profits by £20 The
)
curve m
Figure 31 2 shows the quantity of labour employed at each pnee
of labour Such a curve can be derived from Figure 31 1 by picking various
pnees of the vanable factor and reading off the amount used from the
marginal revenue product curve in just the way desenbed abo\e for the

price of £1,200
Note that this curve is identical with the marginal revenue product curve
in Figure 31 1 The curve m Figure 31 2 relates the pnee of the vanable
factor to the quantity employed and hence it is the demand curve for the

vanable factor This is the key proposition in marginal productivity theory

The marginal revenue product curve of a factor is the

demand curve for that factor

This proposition is true for all profit-maximismg firms, whether they

produce under conditions of perfect compcution, imperfect competition or

Fig 31 2 A firm s demand curve for

a factor of production

monopoly '
Any firm that is maximising its profits will be in a position

m which the prices of the variable factors arc equal to the factors’ respectn e

marginal revenue products This will be so whether or not the firm has
ever heard of variable factors, fixed (actors, marginal revenue products and
such concepts, since this is nothing more than another way of stating the
necessary conditions that profits should be maximised *
firms
1 Since we have denved this from condiuon (2) it follows that it is true as long as the
in
purchasi their factors m a perfect market The proposition is untrue if the firm is
able to
the
fluence the price of the factor by varying the amount that U purchases and to discoef
firm s behaviour m this case we must ose condition (I)
accept e
2 The rest of this chapter can be omitted provided the student is prepared to
favoura e
downward sloping demand curve for a factor as a basic hypothesis for which much
empirical evidence exists
 THE DEMAND FOR FACTORS 417

THE DERIVATION OF THE MARGINAL REVENUE

PRODUCT CURVE
The marginal revenue product of the factor is defined as the addition to
total revenue resulting from the employment of an additional unit of the
variable factor. This may be broken up into two components, a physical
component and a value one, and we must now consider how each of these
varies as the quantity of the factor varies.
Assume, for simplicity of exposition, that we have only one variable
factor, all other factors being fixed. As we vaiq^ the quantity of this factor,
output will vary. The hypothesis of variable proportions that we developed
in Chapter 17 predicts that as we go on adding more and more units of
the variable factor to the given quantity of the fixed factor, the optimal

Fig 31.3 The marginal

physical product curve.

0 10 20 30 40 50 60 70 80 90 100

Number of workers

proportion will be increasingly departed from, and the extra output pro-
duced by successive increments of the variable factor will decline. The
hypothesis is also called the hypothesis of diminishing marginal product and
it is illustrated in Figure 31.3.
Now to convert this curve into a marginal revenue product curve, we
need to know the value of the extra physical product. The marginal physical
product depends solely on the technical conditions of production, but the
value to the firm of this extra product depends on the price, and we must
consider this point in some detail.

A SINGLE PERFECTLY COMPETITIVE FIRM WITH THE OUT-

PUT OF ALL OTHER FIRMS CONSTANT
We start by considering the demand curve for a factor of production of a
single perfectly competitive firm on the assumption that all other firms
418 DISTRIBUTION

keep their output constant VanaDons in this one firm’s output will thus
lease the price of the product unaffected, and we can regard the market
price as exogenously determined and constant The value to the firm of an
extra unit of output is clearly the market price of the product If, for

example, the price is -0., then the marginal revenue product of a factor is

thenumber of additional units of output produced by the factor multiplied

by 0. The marginal revenue product curve is now derned from the
marginal physical product curve simply by multiplying the marginal ph>si
cal product corresponding to each unit of the variable factor by the (fixed)
market price Figure 31 1 shows the marginal revenue product curve corre
spending to the marginal physical product curve of Figure 31 3 and a
market price of (^2 per unit for the product The curve m Figure 3 I is the 1

firm’s demand curve for the variable factor, on the assumption that all

other firms keep their output constant

When the price of the product is constant condition (2) above becomes

J*nce of the factor =* marginal physical product of the factor

muHiplud by the price of the product (3)

A SINGLE PERFECT! Y COMPETITIVE FIRM WITH THE OUTPUT OF ALL

OTHER FIRMS VARYING Wc have two basic relations, the marginal ph>si
cal product curve which depends solely on the technical conditions of
production, and the marginal revenue product curve which is the firm s
demand curie for labour In the previous section we moved from the former
curve to the latter one by assuming that the price of the product was fixed
so that marginal revenue product equalled marginal physical product times
pricewhich was a constant This procedure is valid only if all other firms
keep their output fixed If the price of the variable factor changes, however,
we should expect all firms to vary their production If, for example, the
price of the factor falls, then all firms will hire more of the vanable factor
and the resulting rise m output will cause a fall m the market price of the
commodity This fall m price will cause the marginal revenue productivity
one
of the factor the to total revenue by the employment of
amount added
more unit of the factor - to be less than it would have been if the price of
the product had remained unchanged We must now derive a marginal
revenue product curve on the assumption that, when factor pnees change,
all firms in the industry vary their output in order to maximise their profit
We assume that we know the maigmal physical product curves for all
firms m the industry, and also the demand curve for the product produce
by the industry, and our problem is to denxe a single firm s demand curve
for the vanable Factor on the assumption that all firms in the industry are
to vary their outputs m response to changes in factor prices
 THE DEMAND FOR FACTORS 419

1 Assume some particular price of the factor and find the equilibrium
price for the product. This is done in the manner of Chapter 22 once the

factor price is known, the marginal physical product curves can be trans-
lated into marginal cost curv'es; these cost cur\’^es summed, giwng
are then
us an industry supply curve which, together with the demand curve, deter-
mines the equilibrium price of the product.
2 Next take the marginal physical product curve of the firm in which we
are interested, and multiply each quantity by the market price determined
in (1) above. This gives a marginal revenue product curve on the assump-
tion that market price remains constant as output is varied. This is the cun'e
derived in Figure 31.1, and reproduced in Figure 31.4. Locate the point A

3
Fig 31.4 Derivation of the
firm’s demand curve for a factor
on the assumption that all
firms change their output so
that the price of the product
changes when the price of the
factor changes.

in Figure 31.4 corresponding to the existing price of the factor and the
quantity actually being employed.

This MRP
curve is the firm’s demand curve for the
variable factor, on the assumption that the price of
the commodity is fixed; its slope depends solely on
the technical conditions of production, i,e., on the
slope of the marginal physical product curve.

Now consider a lower .price of the factor, say Oy, instead of Ox. Our
firm, in an effort to maximise profits, will hire more labour. But so will all
other firms and, as a result, the price of the product will fall. As a result the
curv'e showing marginal physical product {MPP) multiplied by existing
market price shifts inwards towards the origin. Thus the firm moves towards
equilibrium in two ways ;
(i) by hiring more labour and (ii) by having its
curve shov\dng MPP times market price shift inwards. A possible equilibrium
is illustrated by point B. The lower price of the product gives rise to a new
420 DISTRIBUTION

curve showing MPP times market price and the new quantity of labour
hired is Os instead of Or We repeat the procedure for each possible pnce
of labour and generate a set of pomts like A and B We then join up these
points and obtain a demand curve for labour allowing for the pnce changes
in the final product This curve is steeper than any of the fixed pnce de
mand curves How much steeper this curve is depends upon how much the
pnce of the product falls as all firms expand output, i c ,
on the elastiaty
of the market demand for the product

A PERFECTLY COMPETITIVE INDUSTRY S DEMAND

FOR A FACTOR
In order to derive the industry’s demand curve for the factor we could
merely aggregate the demand curves developed for individual firms under

Fig 31 5 Denvauon of the

demand curve for a factor

on the part of a perfectly

competitive industry

Quantity

curve m the
(3) abo\c Alternatively we could denve the industry demand
following way Take the AfPP curve for each firm Assume some specific
pnce of the variable factor Derive a marginal cost curve for each firm in

the manner of Chapter 20 Sum these to obtain an industry supply curve

Now repeat the process for each possible price of the variable factor This
gives rise to a whole family of industry supply curves each corresponding to
a particular price of the vanablc factor Such a family of curves is illustrated
in Figure 31 5 (Sjo corresponding to a factor pnce of £10, to a factor

price of £8, etc ) Assume some particular factor pnce, say £6 By using the
market demand curve and the supply curve S^, the cquihbnum pnce and
quantity can be derived Now draw a horizontal line, RQj through the point
of intersection of and DD
Points of intersection of RQ, and the vanous
supply curves tell how production would vary as factor prices vaned, if the
market price did not change Points of intersection of the actual demand
curve, DD, and the various supply curves show how production would vary
if market pnees are allowed to vary Since we now know the amount oi
 THE DEMAND FOR FACTORS 421

production- corresponding to each particular price of the variable factor,

we know the amount of the variable factor that must be employed. Plotting
this amount against its price gives the industry' demand curve for the factor.
Clearly, the variations in production are less when prices are allowed to
change than when they are fixed. We conclude that,' for given changes in
factor prices, the variations in production, and hence the elasticity of de-
mand for the variable factor, are less when prices vary than when they are
assumed be constant. The student should now draw for himself a diagram
to
similar to Figure 31.5 but mth a much steeper demand curve than the one
in the printed figure. He will then be able to show' for himself that the
elasticity of demand for a factor is lower, the low'er is the elasticity of de-
mand for the product.
The reader should also be able to illustrate proposition (2) on page 409
by relating the magnitude of the shifts of the supply curves, to the share of
total costs accounted for by the cost of the factor w'hose price changes.

SUMMARY OF INFLUENCES AFFECTING THE DEMAND

FOR A FACTOR OF PRODUCTION
The demand for a factor has been show'n to depend on two main influences,

the technical conditions of production and the demand for the product that
the factor produces. The technical conditions are the marginal productivity
of the factor, the ease with w'hich one factor may be substituted for another
and the importance of the factor in the total cost of production. The depend-
ence on demand is simple the more elastic is the demand for a product the
;

more be the demand for the factors that make the product. An
elastic svill
industr)'’s demand for a factor of production is dow'nw'ard-sloping because,
as more of the factor is used, its marginal physical product declines and also
the price of the product declines.
CHAPTER 32

THE SUPPLY OF FACTORS

What determines the total supply of labour to the UK economy’ What de-
termines the supply of labour available to a new factory opening up ma
smallish tovm in the Midlands’ In considenng the supply of factors we
make a sharp division between the total supply to the whole economy and
the supply to a particular firm or industry

THE TOTAL SUPPLY OF FACTORS

Are the total supplies of factors fixed, in any meaningful sense’ Many
people think they are, and, at first glance, this is intuitively plausible After
all, there is only so much land in the world, or m England, or in London
There an upper limit to the number of workers, there is only so much
is

coal, oil, copper and iron ore in the earth These considerations do indeed
put absolute maximums upon the supplies of any Factor But, in virtually
every case, we are not near these upper limits, and the problem of changes
in the total effective supply of land, or labour, or natural resources, or
capital deserves discussion

Labour
By the total supply of labour we mean the total number of hours of work
that the population is willing to supply This quantity, which is often called

the supply of effort, is the size of the population,

a function of three things
the proportion of the population willing to work, and the number of hours
worked by each individual What are the determinants of each of these
\ ary m size, and these variations may be influenced
things ’ Populations to

some extent by economic factors There is some evidence that the birth rate
is higher in good times than in bad
You may have found it hard to get into
unuersity, m part because you were bom in the postwar baby boom Much
 THE SUPPLY OF FACTORS 423

of the variation in population is, however, not explained successfully by

economics, and, for Western countries, we must accept that the total popu-
lation varies for reasons that are at the moment largely unexplained.* The
proportion of the population entering the labour market, the labour force, as
it is called, varies considerably in response to variations in the demand for

labour. Generally, a rise in the demand for labour, and an accompanying

earnings of labour, will lead to an increase in the proportion of
rise in the
the population willing to work. Women and elderly people are inclined to
enter the labour force when the demand for labour is high. The dramatic
increase in the proportions of married women and persons over 65 who
were employed during the Second World War is a case in point. Variations
in the number of hours people are willing to work have resulted in a substan-
tial reduction in the supply of labour over a long period of time. Generally,
a rise in real wages, such as has occurred in most Western countries over the
last two centuries, leads people to consume more goods and also to consume

more leisure. This means that people will be willing to work fewer hours per
week, a fact that, unless offset by a rise either in total population or in the
proportion of the population in the labour force, will lead to a decline in
the supply of labour. Workers are in the position of trading their leisure for
goods; by giving up leisure by working), they obtain money and,
(i.e.,

hence, goods. A rise in the wage rate means that there is a change in the
relative price of goods and leisure. Goods become cheaper relative to
leisure, since each hour worked results in more goods than before, and each
hour of leisure consumed is at the cost of more goods forgone.
This is illustrated in Figure 32.1. Leisure is measured on the vertical axis
and the money value of goods consumed on the horizontal axis. Each indi-
vidual starts with 24 hours of his own time. If the wage rate is lOr per hour,
he can have 24 hours of leisure and no goods, or f,\2 worth of goods and
no leisure (much less any sleep), or any combination of goods and leisure
indicated by points on budget line A. Assume, first, that he chooses the posi-
tion indicated by point x, so that he consumes 14 hours of leisure and trades
the other 10 (at lOr per hour) for worth of goods. Now assume that the
wage rate doubles to, say, ;(^1 per hour. He can now have any combination
of goods and leisure indicated by points on budget line B. If he continues
to work for 10 hours per day, he now gets ,(^10 worth of goods, but there is
nothing to stop him from moving to a point above and to the right of x,
in which Ccise he can have more goods and more leisure. If, for example, he
moves to the position indicated by y, he will have an extra two hours of
leisure and an extra worth of goods. On the other hand, the extra in-
come that can be obtained per unit of leisure sacrificed might make him more
1 In some underdeveloped areas, the population varies directly with the food supply and
the quantity and quality of medical services.
 '

424 DISTRIBUTION

Willing to giveup leisure to gel goods He might, for example, move to point
2 and work one more hour, getting 1 worth of goods
'

Thus our standard theory makes no prediction about the effect of an

increase in the wages on the supply of effort We must now wonder what
actually happens m the world do increases in wages cause people to work
fewer hours or more hours ^ Attempts to increase the supply of effort
offered by coal miners in post-war Britain have consistently been thwarted
for the miners have wished to take out part of their increased real income
by consuming more leisure They can do this without departing from the
agreed upon work-week merely by increasing their absenteeism Most of the
long run evidence confirms that as real earnings per hour rise, people wish

Ftg 32 / The choice between

income and leisure

Goods to pounds sieslmg

to reduce the number of hours they work This is of course concerned with
the supply of effort to the whole economy, there is plenty of evidence that
a rise in earnings in one industry will increase the supply of effort to that

industry by attracting workers from other industries

that the
Taxes and the supply or effort It is a very common belief
by
level of taxes found in today’s world tends to reduce the supply of effort
about
reducing the incentive for people to work Of course people complain
result of an
1 The movement from his ongina] position on to hu new position on B is the
income effect and a substitution effect When the wage rate rises, the substitution effect
ta
to increase the supply of labour because givii^ up leisure to get goods is now a more pro
increase
occupation than before The income effect worhs to decrease the supply of labour (i e
including
the consumption of leisure) because the person can consume more of everything
will wts
leisure Whether the nse m
wages causes a nseor a (all in the number of hours people
to woefe depends on the relative strengths of these two effects
 THE SUPPLY OF FACTORS 425

high taxes - they have done so as long as taxes have existed - and of course
they say that it is not worthwhile working because of crushing tax burdens.
This, however, is not real evidence about how people actually do behave,
and, as social scientists, we should want .some systematic, objective evidence
on this subject. Such evidence as does exist suggests that the common belief
that high taxes reduce the supply of effort is not correct. A study of this
problem was made by an American professor, G.F. Break.* He
in Britain
studied a high income and highly tax-sensitive group, but found no general
disincentive effect of the high tax rates (including high marginal rates) over
the period studied. Also the long-run historical evidence of all countries
demonstrates that, as people get richer, i.e., the reward per hour’s effort
rises, they wish to rvork less rather than more hours. We must conclude that

such evidence as exists goes against the commonly held view that a lowering
of the existing level of taxes, and hence an increase in the net reward per
hour worked, would increase the supply of effort in the economy.

Land
If by land we mean the total area of dry land, then the supply of land in a
countr)' is, in this definition, pretty well fixed. A rise in the earnings of land
cannot result in much of an increase in the supply of land, unless we can
drain land that is at present covered with water. The traditional assumption
in economics is that the supply of land is absolutely inelastic. However, if by
land we understand all the fertile land available for cultivation, then the
supply of land is subject to large fluctuations. Considerable care and effort
is required to sustain the productive power of land and, if the return to
land is low, the fertility of the land may
be destroyed within a short period
of time. On the other hand, a high return to land may make worthwhile
irrigation, drainage, and fertilisation schemes that can greatly increase the
supply of arable land.^
There is no value in debating which is ‘real’ land : the total land area or
the total supply of arable land. The magnitude we are interested in depends
on the problem at hand. For most problems in economics, however, it is

the total supply of cultivable land that is relevant. If we are interested, for
example, in the effect of land taxes on the prices of agricultural goods, then

1 George F. Break, ‘Income Taxes and Incentives To Work: An Empirical Study, Great
Britain, 1955-56’. American Economic Review, September, 1957.
2 It is common practice, following David Ricardo, an English economist of the early
nineteenth century, to define land as the original and inexhaustible powers of the soil. Ricardo wrote
before the phenomenon of dust bowls, which turn large tracts of land into barren deserts, was
widely known, and before men were aware that the deserts of North Africa had once been
fertile areas. Clearly, as we know today, nothing of the fertility of land is inexhaustible.

14 *
426 DISTRIBUTION

It 13 of no help to be told that the total land area of a country is fixed what
,

wc need to know is the cflect of such taxes on the supply of cultivable land
The total b) no means perfectly inelastic,
supply of cultivable land is it can
be expanded greatly by irrigation and other forms of reclamation and it can
be contracted drastically and rapidly - as many farmers have found to their
sorrow - by neglecting the principles of soil conservation

Natural resources Men often worry - usually too late - about exhaust
ing natural resources For example, the proved oil reserves in the world
will hardly last another twenty years at current rates of production But
this was also true in 1920' And in 1935* The apparent paradox is resolved
by the fact that every year wc discover about as much oil as wc produce
As long as oil remains valuable, it pa>$ to find more of it If the cost of
finding oil becomes too high, the supplies that it is economically feasible

to find will indeed be exhausted - at those paces But, m that event,

prices will surely rise (unless wc discover a cheaper fuel or a cheaper

method of finding and producing od) But even if technology remains un
changed, an increase in the price of oil of as much as 35 per cent will make
It economically worthwhile to process the vast quantities of heretofore
unexploited shale oil

The same imminent exhaustion vvas told about iron ore as is now
stor> of
told aboutAmericans were repeatedly vvamed that the great Mesabi
oil

Range in Minnesota would be exhausted by 1975 It was cfTectivcly ex

hausted by 1965' And yet Americans sull produce steel, mainl) because
they have discovered ways to use low-grade iron ores, once thought worth
lessthan nothing (because they had to be dug through to get to the ncher
ores) The known supplies of coal appear to be inexhaustible at the current
rate of use
Although natural resources can indeed be exhausted, the problem of
actual exhaustion does not arise as often as one might think B> ingenuity
by substitution of other factors, and by other means, resources have generally
been expanded as needed Perhaps the best example of this is the experience
with rubber during the Second World War Cut off from all the world s
world
supplies of natural rubber, the chemical industnes of the Western
produced synthetic rubber in a period much shorter than would have been
required to grow new rubber trees' To be sure, the pace was higher and
the quality less good

But the success stories should not make the reader over confident Man s
the
ability to find more of any given resource, or to find a substitute to do
job, is impressive in case after case But he has had reverses as well The
species
destruction of the forests of Great Bntain and the extinction of some
of fish and wildlife should serve as warnings Pollution of the water and the
 THE SUPPLY OF FACTORS 427

atmosphere goes on, and could become disastrous. As with land, natural
resources are neither inexhaustible nor rigidly limited.

Capital

Capital is a man-made factor of production. The supply of capital in a

country' consists of the stock of existing machines, plant, equipment, etc.
This capital is used up in the course of production, and the supply is thus

diminished by the amount that wears out each year. On the other hand,
the stock of capital increases each year as a result of the production of new
capital goods, the expenditure on which is called investment expendi-
ture. New' machines replace ones that wear out (although the new ones
will rarely be physically idendcal with the machines they are ‘replacing’).
The total amount of machines produced for all purposes is called gross
INVESTMENT. Machines that are not replacing w’om-out ones, and that
therefore represent net additions to the stock of capital, are called net
INVESTMENT.
The supply of capital has been obser\'ed to increase considerably over
time in all modem countries. The volume of net investment determines the
rate of increase of the capital stock. There is considerable evidence that net
additions to the stock of capital var>' considerably over the trade cycle,
being low in periods of slump and high in periods of boom. Taking the long
view, however, and ignoring cj'clical fluctuations, there has been a" fairly
steady tendency for the stock of capital to increase over a very long period
of time. The theory of investment, which we shall develop in subsequent
sections of this book, is thus a theory of changes in the stock of capital.
The only limit to the possible growth of the capital stock is the willingness
of people to divert resources from the production of consumption goods to
the production of inv'estment goods.

THE SUPPLY OF FACTORS TO PARTICULAR USES

The question of what determines the allocation of factors of production
among various possible uses is a very general one. Even if all factors had
only one use, it w'ould still be necessary to allocate them among competing
firms in the same industry. As it is, factors have many uses a given piece;

of land can be used to grow' a variety of crops, and it can also be subdivided
for a housing development. A
machinist from Coventry can work in a
variety of automobile plants, or in a dozen other industries, or even in the
physics laboratories at Cambridge. Factors must be allocated among dif-

ferent industries and they must also be allocated among different firms in
the same industry.
u’n ntITRtBt TIDS

owners of ihor factors art mainl> ctiucenml with maVms; « much

If the
monc> n» tlicj nn the) will most ihnr firtor to tlial use u which it earns
the most monc> tins mosemrnt out of one use into another will continue
,

until ihecamm^of i factor in eachnfitssinoui possible usenre the same

Since owners of ficion take other thinj^ I>esicln mone> into account
tlimijt like risk coiuemencr and a roikI climate ficton ssill be moud
amoni* uses until tbere is no net adv in(ai<e in fiirtlier movement allowing
for Ixitli tlie monel ir) mil nonmonetif> idaaniatjes of differeni uses

1 iir jtai oTjirsis nr t t>t At srT AisvASTAnr Hie hypothesis of equal

net advaiitJRC may lx" ilatesi as follows

Owners of factors u itl choose that use of their factors

that produces the {•reatest net adv-antage to them
selves Net adv-antage includes both pecuniary and
nonpecuniar) elements

Ilni hvpothrsii phva tlie same role in the iheon of distnbution as the

hypothesis that firms seek to maximise profits plava m the theorv of produc
non ITie siemfie mt prethction that it leash to is that factsm of production
Will lie allocated among Mtioiis uses in such a way that Ihev receive the

same net return in each use Tins Inpsithesis is howeser iinsilisficiora ai

It stands
Tlie troidile witli itthat unless ssr can measure nonmonetary ad
is

vantage it is irrefutable Suppose we «l>sene that a mechanic is working in

Ixmdon for a vcir less than he could make m Ncwcaitlc Is this eva
dcnce agamst the hypothesis or does it merely mean that the tionmoneurv
Iwnefils of living in Ixindon or of not living Newcastle) are worth lo

him ’ A moment s thoiiglii will make it clear that anv conceivable observa
lion could be rationalised to fit the liypolhrsis John would rallier starve

as a bimsicr than work for his father )

To restore the hypothesis to the |>omJ where it is possible to imagine

gathering evidence winch conflicts wadi it and llms to the point where it
can be useful to us we must do one of tsvas things cither we must defnc

no & mr-iwrakAt wtv aVic oonmoncaary bencYfts tAiti we "len v?j^

tant to choices or we must make an assumption about the relative stability

of monetary and nonmonetary advantages TIic first alternative is generally

regarded as impossible unless weassunve that the hypothesis is correct The
o
second alternative to make an assumption about the relative stability
monetary and nonmonetary advantages - is more feasible If for example
we assume that the differences in nonmonetary advantages between two
uses of a factor remain constant over lime we cm predict that variations
 THE SUPPLY OF FACTORS 429

in monetary advantages will widen or narrow the gap and that some re-
sources will flow in response to the change.
It is not necessary, of course, to make the strong assumption that non-
pecuniary advantages are constant. Instead, we can assume that they
change, but more slowly than pecuniary ones. In this case, we can still
extract predictions about behaviour.^ This weaker assumption leads us to
the following fundamental prediction

Any change in the relative size of the rate of pay be-

tween two uses of a factor will lead owners of factors
to increase the quantity they wish to supply to the use
in which the relative pay has increased, and decrease
the quantity they wish to supply to the use in which it
has decreased-

This prediction implies a rising supply curve for a factor in any particular
use.Such a supply curve (like all supply curves) can shift in response to
changes in other variables. One of these is the size of the nonmonetary
benefits.

Factor Mobility (Elasticity of Supply)

Does our hypothesis that factor-owners respond to changes in what they can
earn from their factors mean that even small changes in offered rates of
remuneration lead to large movements of factors, and thus to ver)' elastic
factor supplies? The answer depends upon factor mobility, the ease with
which factors can move between uses.
Factor mobility, or immobility, is an important aspect of how well
resources respond to the signals that indicate where factors are wanted. If
a factor is highly mobile in the sense that the alert owners of this factor will
quickly shift from use A to use B in response to a small change in factor price,
then supply will be highly elastic. If, on the other hand, factor-owners are
‘locked in’ to some use and cannot respond quickly, the supply will tend to
be inelastic, even though owners may genuinely wish to take advantage of
the higher prices offered elsewhere. Factor mobility is dependent on the
speed with which factors will respond. The barriers to mobility vary sub-
stantially from factor to factor.

Mobility of land; Land, which is physically the least mobile of all

factors, is paradoxically one of the most mobile in an economic sense.

1* It would be sufficient to assume that pecuniary and nonpecuniary advantages vary inde-
pendently of each other. See Chapter 3. But this is an assumption that is likely to prove
incorrect.
430 DISTRIBUTION

Consider agricultural land Within a year at most, one crop can be harvested
and a totally different crop planted A
ferm on the outskirts of a growng
citycan be sold for subdivision and development on very short notice Once
land IS built upon, as urban land usually is, its mobility is much reduced

One can convert a site on which a hotel has been built into an office*
building site but it takes a very large differential m the value of land use to
make it worthwhile because the hotel must be torn down

Mobility of capital Most capital equipment, once constructed, is

pretty immobile A great deal of machinery is utterly specific, once built it

must either be used for the purpose for which it was designed, or else not

used at all This

of course, not true of all pieces of capital equipment a
is,

shed, for example, may be used lor a large number of purposes but much
capital equipment is extremely immobile among uses dunng its phjsical

lifetime 1 1 is the immobility of capital equipment that makes exit of firms

from declining industries a slow and difficult process
In popular discussion, sums of money are often referred to as capital

Money on resources A firm or a household that has

represents a claim
accumulated money savings can spend these on anything that it desires It
can buy beer or machines and by so doing, will direct the nation’s resources
to the production of beer or machinery AUo, the firm or household can lend
Its money to other firms or households and thereby allow the borrowers to

determine what the nation’s resources will be used to produce

Money itself is not a factor of production But it represents a claim to the
goods and services produced by factors, and the way in which the claim is
exercised influences the allocation of factors of production among alter
native users Money itself is highly mobile An owner who planned to use
It one purpose can change his mind at the last moment and use it for
for
another The long term mobility of capital equipment comes about through
the mobility of money During the life of a piece of capital, the firm may
make allowances for dcpreaation that allows the firm to replace the capital
good when it wears out If conditions of demand and cost have not changed,
the firmmay spend money to replace the worn-out piece of equipment with
an identical one The firm may, however, do many other things wuh its
funds it may buy a newly designed machine to produce the same goods, n
may buy machines to produce totally different goods, or it may lend the
money to some other firm for the latter’s uses In this way, the long run
allocation of a country’s stock of capital among various uses changes

Mobility of labour Labour is unique as m one important way

a factor
nonmonetary considerations play a lai^er role in the mobilitv of labour than
in the mobility of any other factor, because the resource involved and the
 THE SUPPLY OF FACTORS 431

owner of it are inseparable. People may be both satisfied ndth or frustrated

by the kind of work they do, where they do it, the people they do it with,
and the social status of their occupations.
Many organisations, private and public, adopt policies that affect their
personnel and impede their mobility as workers. When labour unions
negotiate seniority rights for their members, they protect the old employee
from being laid off in a cutback of production, but they also make him very
reluctant to change jobs. When an employer provides his employees wth a
pension plan, the employees may not want to forfeit this benefit by changing
jobs. When the governments of individual states in the US provide com-
pensation to unemployed residents, these residents may be reluctant to
leave the state, even to find work.
Although labour is in a substantial degree influenced by nonwage con-
siderations, we should still expect, according to the fundamental hypothesis
of factor supply, that labour would respond to changes in the wage struc-
ture. But we must be careful. If doctors’ earnings go up and farmers’ earn-
ings go down, would farmers move into medicine ? No, for it is not easy for
a farmer to become a doctor (or for that matter for a doctor to become a
farmer). The very visible barriers between various occupations have led
some economists to speak of labour as consisting of a series of noncompeting
groups and to explain differential pay between occupations as the conse-
quence. Since, as we all know, the sons of many farmers become doctors,
the notion of real walls between various occupations does not seem a very
adequate explanation of the immobility of labour. And not all occupational
shiftsare as difficult as that between medicine and farming.
The key to understanding labour mobility is time. In the short run, it is
very difficult for people to shift occupations. It is not hard for a secretary to
shift from one company to another, and it is not hard to persuade her to

take a job in London instead of in Colchester, but it will be difficult for her
to become an model in a short period of time. There are
editor or a fashion
two considerations here; ability and training. Lack of either will stratify
some people and make certain kinds of mobility difficult for them.
Over long periods, labour mobility between occupations is very great
indeed. In assessing the mobility of labour, it is important to remember that
the labour force is not static. At one end, young people enter the labour
force from school, and at the other end, exit from retirement or death. The
turnover in the labour force, due to these causes, is something on the order
of 3 or 4 per cent per year. Thus, even if no one ever changed johs, it would
be possible to re-allocate 3 or 4 per cent of the labour force annually merely
by redirecting new entrants to jobs other than the ones left vacant by per-
sons leaving the labour force. Over a period of 20 years, a totally different
occupational distribution could appear without a single individual ever
 i

432 DISTRIBUTION

changing his job The role of education m adapting people to needed jobs
IS very great Since children spend much of their first 17 ^ears in school n
IS possible to achieve large increases m the supply of any desired labour sliil

uithin a decade or so Sputnik made the world aware of the notable succea
of the Soviet Union in increasing its supply of scientists in a very short
period of lime
Various studies have been made to determine the amount of mobility
shown by labour in mov ing from job to job and place to place In times of
heavy- depression mobility from place to place is very low Labour is under
standably reluctant to move from areas with say 20 per cent uncmploy
ment into areas with say 10 per cent unemployment even though tie

chances of finding a job in the latter areas may be higher than m the former
ones '
In periods of more or less full employment there appears to be some
evidence that differentials m wages between areas and occupations do reflect

relative scarcities and that labour does tend to some extent to move from
low wage sectors of the economy to high wage ones there seems to be even
stronger evidence however that labour is more attracted by the chance of
obtaining a job than by the wage rate actually paid for that job
By way of contrast studies of labour mobility over the generations or
social stratification as the sociologists call it indicate impressive mobility The
data show in a nutshell that while it pays to have a successful father (sue

cessful m the sense of being a member of one of the occupations at the high

end of the scale) it is not necessary to becoming successful one s self Edu
cation IS the key to an indmdi5al s earning a high income and by the sami
token to expanding supplies of factors in those occupations for which there
IS the greatest demand for services

To sum up, labour is much more mobile in the long

run than in the short run Over a given time period, it
IS more mobile between jobs in the same location and
occupation than between different locations (where
movement of the family is a deterrent) or between
different occupations (where lack of skills is a
deterrent)

Man made barriers to labour mobility Among the many institu

tions that man has introduced that limit factor mobility we have already

mentioned private pension schemes eligibility requirements for unemplo)

ment compensation and senionty provisions in collective bargaining agree
ments
There arc other barriers as well Incensing is required in dorens of trade*
1 Scf H \V Rob nson The Rnpon eoTLabour to Eeonom e Inccniivn r\ Oififit S

1 Iht Pr t \Stekan cd T \V Iwn and P \\ S And ewj (Clarendon Pre?i 1951)

 THE SUPPLY OF FACTORS 433

and professions. Barbers, electricians, doctors and, in some places, even

pedlars must have licences. There is, of course, a perfectly legitimate reason
for requiring licences in cases in which the public must be protected against
the incompetent or the quack or the nuisance. But licensing can also have
the effect of limiting supply. The fact that in countries without national
health services medicine is often one of the highest paid occupations (aver-
age income of doctors in the United States was about ^^8,000 per year in
1965) and that doctors have long been in short supply is a result of the
difficulties in getting into medical schools, the long internship and residency
requirements, the rules concerning certification, and so forth. It is at least
possible that in these countries doctors’ earnings are high because the
barriers to entry into the profession prevent even long-run increases in the
proportion of the population being admitted to medical practice. Whether
such barriers as exist are required by the standards of the profession or are

designed to keep the supply limited (and the earnings high) is a matter
open to debate. Unions may impose barriers to labour mobility. The
‘closed shop’, for example, which requires all employees of a plant or a
trade to be a member of a particular union, gives unions the power to limit
the supply of labour that they represent. Racial prejudice, and other
arbitrary attitudes, also limit the mobility of labour.
How important barriers of each of these kinds are, singly and in aggre-
gate, is an empirical matter and the evidence is not conclusive in every case
that we have considered.
CHAPTER 33

THE PRICING OF FACTORS OF

PRODUCTION IN COMPETITIVE
MARKETS

We have now developed theones both of the demand for, and the supply of
factors of production This is all we need to develop a theory of the pncing
m a free market given that factor prices are free to vary, paces
of factors
and quantities employed will tend to the point at which supply equals
demand Furthermore, shifts in either the demand for, or the supply of
factors will have the effecu on prices, quantities and factor incomes pre*
dieted by normal price theory
The theory ofjaetor prices is an absolutely general one If one is concerned with
labour, one should interpret factor prices to mean wages, if one is thinking
about land, factor prices should be interpreted to mean rent, and so on
In this chapter we assume that factors are bought and sold on a competitive

market by a large number of buyers and sellers In Chapter 34, we shall

introduce monopolistic elements into the markets for factors

MONEY REAL AND RELATIVE PRICES

Changes both in relative factor prices and in the share of national income
going to a factor are likely to be brought about through changes m the
money prices of a factor, but the mere knowledge that the money price of a
given factor has risen tells us very little about what has happened cither to
the position of the factor relative to another factor, or relative to the whole
national income If we wish to enquire into relative pnees we need to know
how the money pnees of at least two factors have changed, for only then
will we be able to say that the price of one factor has increased or decreased
relative to the price of another
factor If we wish to know what has happened
to the share of the
national income going to some factor, then we must know
 PRICING OF FACTORS OF PRODUCTION 435
itsmoney income and the money value of total national income. Only if
money national income and all other money prices are constant can varia-
tions in the money prices of one factor give us direct information
about the
factor s relative price and its share in the national income. In partial equi-
librium analysis, we assume that total income and all other prices remain
constant. In this case, therefore, a variation in the money price of
a factor
causes a simultaneous variation both in its relatwe price {compared to any
other single factor) and in the share of the national income going to the
factor.

RELATIVE FACTOR PRICES UNDER COMPETITIVE

CONDITIONS
If there was only one factor of production all the units of which were identi-
cal then the prices of all the units of this factor would tend towards the

Number of factors Number of factors

i ii

Fig 33.1 The distribution of 7,000 workers between two industries with different
demand curves.

same level. Units of the factor would tend to move from low-price occupa-
tions to high-price ones. The would diminish in
supply of the factor
occupations in which prices were low and the resulting shortage would tend
to force the price up the supply of the factor would increase in occupations
;

in which prices were high, and the resulting surplus tvould force the price
down. The movement would continue until there were no further incentive
to transfer, i.e., until the price of the factor was the same in all its uses. This
equality would be established whatever the states of demand in the various
industries. This proposition is illustrated by example in Figure 33.1 which
436 DISTRIBUTION

shows that, given the states of demand, most of the factor would flow into
the second industry until prices were equalised
Causes of differences m
factor prices are of two sorts, dynamic or dis
equilibrium ones, and static or equilibrium ones The dynamic differences
are ones associated with changing circumstances, such as the rise of one
industry and the decline of another Such differentials set up movements in
factors that will themselves act to remove the variations The differences m
pnees may persist for a long time, but there is a tendency for them to be
reduced and, m
equilibrium they will be eliminated Equihbnum differ
ences m
pnees, on the other hand, arc ones that would persist in a state of
equilibrium without there being any tendency for them to be removed by
the competitive forces of the market

Dynamic Differentials and Factor Mobility

dynamic differentials If there were a nse in the demand for
First consider
product A and a fall in demand for product B, there would be an increase
m the (derived) demand for factors in industry A and a decrease in the
(derived) demand for factors m industry B Factor pnees would go up in

A and down m m
B This is an example ofa dynamic change relative pnees
for the changes themselves will cause factors to move from industry B to
industry A, and this movement will cause the price differentials to lessen
and eventually to disappear How long this process takes depends on how
easily factors move from one industry to the other, i e on factor mobility
,

Some factors, particularly labour, may be relatively immobile in the short

run and thus dynamic differentials can last for a long time even if there are

no barriers to movement The factors which affect labour mobility and

which thus determine the duration of dynamic differentials have been dn
cussed m more detail in Chapter 32

Equilibrium Differentials

Equilibrium differentials in factor pnees are related to differences m the

factors themselves (c
g land of different fertilities, and labour of different
,

abilities) or to different nonmonetary advantages of different factor employ

ments Ce/enj/JanJuf a job with high nonmonetary rewards will have a lower
equihbnum wage rate than a job with low nonmonetary rewards Thus it
jobs
IS pay people in academic and research
often possible, for example, to
lessthan they would be able to earn in the world of commerce and industry
because there are substantial nonmonetary advantages attached to the
former type of job compared with the latter If labour were paid the same
m both jobs then it would move out of industry and into academic employ
 PRICING OF FACTORS OF PRODUCTION 437
ment. Excess demanci for labour in industry and excess supply in universities
would then cause industrial wages to rise relative to academic ones until
the movement of labour ceased.
Equilibrium differentials may also be caused by differences between
factors. There is, for example, a shortage of persons able and wiUing to do
skilled jobs. Thus, in equilibrium, the skilled worker earns more than the
unskilled worker; no movement from unskilled to skilled jobs eliminates
this differential, because it is impossible for most unskilled workers to be-
come skilled ones. It is important to realise that the high pay of the skilled
man relative to the unskilled one merely reflects relative demand and supply
conditions for these two types of labour. There is nothing in the nature of

competitive markets that ensures that the skilled worker always gets high
pay just because he is skilled. If, for example, the demand for skilled workers
fell off so much even though the supply was small, there was a glut of
that,
such workers, their wages would come down. On the other hand, if there
were a change in education so that unskilled workers could now acquire
skills, the wages of skilled workers would fall relative to those of unskilled

workers. History is replete with examples of particular groups of skilled

workers who have lost their privileged position when there was a change
in the supply of their services. Most middle-class people
both feel that it is

unjust and incomprehensible Second World "War,

that, since the lorry
drivers and coal miners have made much more money than many relatively
highly educated office workers. Whatever the justice of the matter, it is
certainly not incomprehensible. A rise in the supply of office workers rela-
tive to the demand for their services and a decline in the supply of lorry

drivers and coal miners relative to the demand for their ser\dces will,
according to the normal workings of the market, raise the earnings of lorry
drivers and coal miners relative to those of office workers. If there are sub-
stantial nonmonetary^ benefits to being an office worker rather than a coal
miner or a lorry driver then the earnings differentials will not set up a flow
of labour out of the latter occupations into the former one and the differen-
tial will persist (i.e., it will be an equilibrium one).

TRANSFER EARNINGS AND ECONOMIC RENT

Some differentials in factor prices wll, as we have seen, persist even in

equilibrium. In order further to study differences in factor earnings be-

tween alternative uses we need to distinguish between the amount that a
factor must earn in its present use in order to prevent it from transferring to
another use which is called the factor’s transfer earnings and any
excess that the factor actually earns over this amount which is called the
factor’s ECONOMIC rent. The meaning of rent in this context differs greatly
438 DISTRIBUTION

from the meaning of rent in the everyday usage of the term rent, and y,c
must first try to understand how this rather specialised usage of the term
rent grew up

The history of the concept of economic rent In the early part of

the nineteenth century, when English economics was in its infancy, ihere
was a controversy about the high pnee of corn (the generic term for a!!
grams) Some people held that com had a high price because the landlords
were charging very high rents to the farmers who were growing com on
their land It was held that, in order to meet these high rents, the pnee that
farmers charged for their com had to be raised to a very high level Thus it
was argued, the price of corn was high because the rents of agncultura! land
were high The second group, which included David Ricardo, one of the
great figures of English Classical economics, held that the situation was
exactly the reverseThe price of com was high because there was a shortage
of com caused by the Napoleonic Wars Because corn had a high pnee there
was keen competition amongst farmers to obtain land and this competition
bid up the rents of corn land If the pnee of corn were to fall so that corn
growing became less profitable, then the demand for land would fall, and the
price paid for the use of land, i e , rent, would fall as well Thus, this school
held that the rent of corn land was high because the price of corn was high
and not vice versa The modem student of economics will recognise m the

Ricardian argument the idea of denved demand Landlords, Ricardo was

saying, cannot just charge any pnee they want for land, the prices they
well
get will depend on demand and supply The supply of land is pretty
fixed and the demand depends on the pnee of com The higher is the pnee
of corn the more profitable will it be to grow com, the higher
will be the

demand for com land and the higher will be the pnee paid for its use Thus
the rent of corn land depends on the price of com
The argument was elaborated by considering land to have only one use,
the growing of corn The supply of land was given and virtually unchange
prefer
able, I e land was in perfectly inelastic supply and landowners would
,

to rent out their land for some return rather than to leave it idle. Nothing
had to be paid to prevent land from transferring to a use other than growing
corn because it had no other use, and no self-respecting landowner woul
leave his land idle as long as he could obtain some return, no matter how
small, by renting it out Therefore, so went the argument, all of the
pay

ment to land, i e , rent, is a surplus over and above what is necessary to keep
on the
It
of land, the price will depend
in Its present use Given the fixed supply
demand for land which is itself a function of the pnee of com
Rent, which was the term for the payment for the use of land, thus
became the term for a surplus payment to a factor over and above what
 PRICING OF FACTORS OF PRODUCTION 439

was necessary to keepin its present use. Subsequently two facts were
it

realised. First, factors ofproduction other than land often earn a surplus
over and above what is necessary to keep them in their present use. Movie
stars, for example, are in very short and pretty well fixed supply, and their

possible earnings in other occupations are probably quite moderate; but,

because there is a huge demand for their services, they may receive pay-

ments greatly in excess of what is needed to keep them from transferring to

other occupations. Secondly, it was realised that land itself often has many
alternative uses and, from the point of view of any one use, part of the pay-
ment made to land would necessarily have to be paid to keep the land in its
present use. Thus it appeared that all factors of production were pretty
much the same part of the payment made to them would be a payment
;

necessary to keep them from transferring to other uses, and part a surplus
over and above what was necessary to keep this factor in its present use.
This surplus came to be called economic rent.

Two MEANINGS OF THE TERM RENT: The term economic rent is a most
unfortunate one. The adjective economic is often dropped and the econo-
mist often speaks of rent when he means economic rent, thus causing a con-
fusion between the concept of a surplus over and above transfer earnings
and the payment made to landlords for the hiring of land and buildings.
When a tenant refers to his rent he is referring to what he pays his landlord,

much of which is a transfer earning necessary to prevent the land and build-
ing in question from being transferred to some other use. It is usually clear
from the context whether the term rent is being used to refer to a surplus
over and above transfer cost (the economist’s usage) or to the total pay-
ment made from tenant to landlord (the everyday usage) but it is important
to guard against confusing the two concepts.

The modern distinction between transfer earnings and economic

rent; Transfer earnings may be defined as any payment which must be
made to a factor to keep it in its present use, i.e., a payment necessary to

prevent the factor from transferring to some alternative employment. Eco-

nomic rent may be defined as any payment made to a factor over and above
that necessary to keep it in its present use; economic rent, therefore, is the
difference between the factor’s actual earnings and its transfer earnings.

In most cases the actual earnings of a factor of production will be a

composite of transfer earnings and economic rent; it is possible, however, to
imagine limiting cases in which all earnings are either transfer earnings or
economic rent. Consider some individual firm or some industry faced with
a perfectly elastic supply curve of a factor of production it will be able to ;

obtain all that it wants at the going price but, if it does not pay this price, it
440 DISTRIBUTION

wiU not obtain any quantity of the factor In such a case, which is illustrated
in Figure 33 2, the whole price paid to the factor represents transfer earn
mgs, the amount that is actually paid must be paid to prevent the factor
from transferring to another use

Ftg 33 2 All of the income earned

by the factor « tiansfcr tannngs

Quantity

Now consider the case of a factor that is fixed m supply and has only one
use Assume that this factor is put on the market by its owners and sold for

whatever it will fetch, on the grounds that some income is better than

nothing at all The whole supply is owned by thousands of different owners

so there is no point in any one of them withholding his own (small) supply

Fig 33 3 All of the income earned

by the factor u an economic rent

from the market in an effort to raise the price Such a factor will be m
perfectly inelastic supply the amount offered for sale will be the
same
whatever the price This case is illustrated in Figure 33 3 The whole of th^
pnee that is paid to the factor is an economic rent because, if a lower pnee
were paid, the factor would not transfer to an alternative use It might be
 .

PRICING OF FACTORS OF PRODUCTION 441

thought that, in such a case as this, the factor would not command any
price, but this is not the case. The price, as in all other free-market cases, is
determined by demand and supply. The fixed quantity available is, in the
example illustrated in Figure 33.3, the amount Oa, while, if the price were
zero, the amount demanded would be Ob. Thus at a price of zero there
would be excess demand for the factor, and competition amongst buyers
would force the price upwards until it reached Op and the excess demand
disappeared. The equilibrium price paid to the factor is Op, the quantity
employed is Oa and, hence, total factor income is indicated by the area of
the rectangle Opra.

Fig 33.4 Some of the income earned by the factor is a transfer earning and the
remainder is an economic rent.

Figure 33.4 shows an upward-sloping supply curve of a factor of produc-

tion.Given the demand curve, D, the equilibrium price would be Op and
the equilibrium employment Og; total factor earnings would be Opzg. If
Og units of the factor are to be attracted into the industry, and if a single
price must be paid, then it is necessary to pay the price Op. However, all
but the Ogth, which might, say, be the 15,000th) would be
last unit (i.e., the
prepared to remain in the industry for a price less than Op. In fact, Oa of
these units would be prepared to remain if the price were as low as Or If the
price rose from Or an additional ab units would be attracted into the
to Os,
st units, an additional be units would
industry; if the price rose by a further
enter the industry. Clearly, for any unit that we care to pick, the point on
the supply curve corresponding to it shows the minimum price that must
442 DISTRIBUTION

be paid in order to keep that unit in the industry {i e , its transfer earnings)
Equally clearly, if the supply curw slopes upward, all previous units have
lower transfer earnings Consider, for example, the Oifth unit (possibl) the
10,000th unit) The transfer earnings of this unit are Ou, if any less is paid,
this unit will not be supplied, and if any more is paid, it is a rent - a pa)
ment what is necessary to keep the factor in its
in excess of present use
We could repeat the same ai^ument for every point, we could show, for

example, that the transfer earnings of the Oath unit are Or, of the 04th unit
Os and of the Octh unit Ot Repetition of this argument for every unit from
the first to the O^th shows that the total transfer earnings of the Og units is
the white area below the supply curve Since the total payment made is the
rectangle Opzg, it follows that the economic rent earned by the factor is the
shaded area aboie the supply curve and below the line pz
The following example illustrates why a rising supply curves involves
rents if universities increase the salaries paid to professors of economics in

order to attract additional economists into university teaching (and awa)

from industry and government), those economists who are pmuaded to
shift into university teaching will be receiving only transfer earnings But
those economists who were already content to be university professors will

find their salaries have increased as well, and this increase will be a rent to

them
Figures 33 2, 3 and 4 suggest the following important conclusion

The more elastic the supply curve, the less the Binovmt
of the payment to factors that is a rent and the more
that IS a transfer earning

Kinds of Transfers

How much of a given payment to a factor is an economic rent and how much
isa transfer earning depends on what sort of transfer we are considenng
a
If we consider the transfer of a factor from one firm to another within
elastic,
single industry for which the supply of the factor should be highly
eanungs If the
then pretty well all of the factor's earnings will be transfer
firm in question did not pay the factor the going price, then the factor
would transfer to another firm in the same industry If we are considenng
the transfer of a factor from one industry to another, then part of the pa)
ment may be a transfer earning and part an economic rent, since mobilit)
will be less and thus the supply curve less elastic IVe cannot point to a given
factor, a labourer, say, and assert that of his income of ,000,
ua
transfer earning and ;{|200 a rent, for it all depends on what transfer wc arc
denng
 :

PRICING OF FACTORS OF PRODUCTION 443

Economic Rent and Transfer Earnings in the Payment to Labour

Labour is able to move from job
to job and something must be paid to keep
a given unit of labour in its present use; what has to be paid for this reason
is a transfer earning to labour. How much has to be paid to keep labour in

its present use depends upon what

the use is. Let us say, for example, that
carpenters receive working a normal 8-hour day. Then a single small
for
construction firm uill have to pay £Z per day or it will not obtain the ser-
vices of any carpenters. To that one firm the whole £Z is a transfer payment;
if the whole £Z were not paid, carpenters would not remain with that firm.

But consider what would happen if the construction industry encountered

difficulties so that all construction firms were forced to reduce the wages
offered to carpenters. In this case carpenters could not move to other con-
struction firms to get more money. If they do not like thewages offered,
they will have to move to another industr>'. If the best they can do else-
w'here is £2 lOr per day, then they will not begin to leave the construction
industry until wages in that industry' fall below £2 lOr. In this case the
transfer earnings of carpenters in construction are £2 lOr and, when they
were receiving £Z (presumably because there w'as a heavy demand for their
services), the additional lOr was an economic rent from the point of view of
the construction industry'. Now' consider what w'ould happen if there was a
decline in the demand for carpenters in all industries. The only thing left

to do, if one did not like the wages,would be to move to another occupation
in other words one must ceasebe a carpenter. If no one w'as induced to
to
do this until the wage fell to £2, then £2 would be the transfer earnings for
carpenters in general. The ivage of £2 must be paid to persuade people to
be carpenters at all. The important general moral to this story' is as follows;

how much of any given payment and how' much

represents transfer earnings
economic rent depends on what sort of a transfer we are considering; one
from firm to firm, one from industry to industry, one from occupation to

occupation, or possibly one from area to area.

Some very highly specialised tj'pes of labour are in very inelastic supply.
This is particularly true of persons such as singer^ and actors who have a
special talent which cannot be duplicated, whatever the training. The
earnings that such persons receive are mostly in the nature of an economic
rent: they enjoy their occupations and w'ould pursue them for very' much
less than the high remuneration that they actually receive. Their
high

rew'ards result fact that they are in very scarce supply relative to the
from the
demand for their services. If the demand for their seri'ices w'ere to rise their
earnings would rise permanently, while if the demand fell their earnings
would fall permanently.

An important policy implication of the distinction between rents and

444 DISTRIBUTION

transfer earnings concerns the eflectiveness of wage increases on supply For

example, if the Central Authonties fed that it is in the national interest to
have more physicists, should they subsidise physicists’ salanes’ In the long
run, such a policymay well have an effect upon supply It may influence
school boys uncertain about whether to become engineers or physicists to
become physicists But it will also mean that a great deal of money will have
to be spent on extra payments to people who are already physicists And
these payments will be economic rents, since existing physicists have demon
strated that they arc prepared to be physicists at their old salanes Although
some may have been considering transferring to another occupation such
movements arc not common ' An alternative policy, which may produce
more physicists per pound, is to subsidise scholarships and studentships for
students who will tram to become physicists This policy operates only on
those on the margin of entenng or leaving the occupation and does not
provide extra rent for those already in it The Department of Scientific and
Industrial Research m Britain and the National Science Foundation in (he

US do preciselyIn the same way, football teams and movie studios

this

provide training schools If the supply curve is quite inelastic, an increase

in the quantity supplied may be more easily achieved by shifting the supply
curve to the right through policies such as those outlined above rather than
by moving along it by raising the market pnee of the factor

Rent and Transfer Earnings in the Return to Capital

If a piece of capitalequipment has several uses then the analysis of the last
section can be repeated for the case of the machine Many machines, how
In
ever, once they are constructed, are utterly specific, having only one use
this case, any income that is made from the operation of the machine
is in

the nature of a rent Assume, for example, that when a machine was m
stalled It was expected to earn £500 per annum m excess of all its operatuiS
can
costs If the demand for the product now falls off so that the machine
only earn £200, it will still be worth while operating it rather than throwing
It away In fact, assuming the machine to have only one use, it will
pay to

keep It m
operation rather than scrap it, as long as it yields any return
at

say
all over its operating costs ^ Thus, if the machine does yield a return ol

The problem phi's fsis

1 i> diflerent where loternational mobility is high - as it is with

One Bnta n
of the reasons for the considerable migration of trained physicists of all ages from
to the United States is very much higher monetary rewards to be earned the US compaf'
m
with the UK Clearly many physicaU are being paid less than theit Internationa
transfer earnings and the result u a steadynet flow from Britain to America
356 t a
2 This 15 just another way of stating the proposition given in Chapter 27 page
variable costs
It pays a firm to continue m operation in the short run as long as it can cover its

of production
 PRICING OF FACTORS OF PRODUCTION 445

£500 per annum in any one year, we can say that all of the return is an
economic rent because the machine would still have been allocated to its
present use - it has no other - as long as it yielded even £1 above its operat-
ing costs. Thus, once the machine has been installed, any net income that it earns
is a rent (i.e., a payment not necessary to keep it in its present use). How-

ever, the machine will in time wear out and it will not be replaced unless
it is expected to earn a return over its lifetime sufficient to make it a good

investment for its owner. Thus, over the long run, some of the revenue
earned by the machine is a transfer earning; if the payment is not made, a

machine will not continue to be allocated to that use in the long run. In any
one year, however, the income earned can sink to zero without affecting
the allocation of existing machines to different uses in the economy.
In this case, whether a payment made to a factor is an economic rent or
a transfer earning depends on the time span under consideration. In the
short run all of the income of a machine is in the nature of a rent, while in
the long run some of it is in the nature of a transfer payment. Factor pay-
ments which are economic rents in the short run and transfer payments in
the long run are called quASi-RENTS.

Economic Rent and Transfer Earnings in the Payment to Land

The formal analysis is identical to that given in the case of labour. How
much of the payment made to a given piece of land is a transfer payment
depends upon the nature of the transfer. Consider, first, the case of an
individual wheat farmer. He must pay the going price of land in order to
prevent the land from being of other wheat farmers. From
transferred to the use

his point of view, therefore, the whole of the payment that he makes is a
transfer payment. Now consider a particular industr)'^ which uses land. In
order to secure land for, say, wheat production, it will be necessary to offer

at least as much as the land could earn when put From the
to other uses.

point of view of the wheat industry, that part of the payment made for land
which is equal to what it could earn in its next most remunerative use is a
transfer payment. If that much is not paid, then the land will in fact be
transferred to the alternative use. however, land particularly suitable for
If,

wheat growing is scarce relative to the demand for it, then the actual pay-
ment for the use of this land may be above the transfer payment; any addi-
tional payment is an economic rent. Assume, for example, that the maxi-
mum payment that can be by farmers wishing to use land for the
offered
crop next most remunerative to wheat is £10 per acre. Now wheat farmers

must offer £10 per acre to secure the land. But it may be that the price of
wheat is such that the profits of wheat growing are very high when only
£10 per acre is paid for the land. A large number of farmers tvill wish to
446 DISTRIBUTION

hire land at this price in order to grow wheat the demand for wheat land
will exceed supply Competition will bid up the rent ofTcred until the
its

demand is equal to the supply The rental finally established might bt say
£\Z per acre In this case, from the point of view of the wheat industry as a
whole, ^10 is transfer earnings and the remaining is economic rent If

the price of wheat falls and, as a result, the demand for wheat land falls

then the rent paid for wheat land will also fall This will continue until the
rent offered falls to ;(^10 and then land will begin to be transferred out of
wheat into other uses Thij transfer wnll continue and the supply of wheat
will diminish until the price of wheat rises sufiicicntly so that the remaining

wheat farmers can offer to pay the £\(i transfer cost to keep the land out of
other uses
Land is very mobile between agricultural uses because the location is

usually of little importance In the case of urban uses, however, location of

the land is critical and, from this point of view, land is of course completely
immobile If there is a shortage of land in central London, such land as is

available will command a high pnee but, no matter what the price paid

this will not cause land in rural areas to move into central London The
very high payments made to urban land are in the nature of economic rents
The land is scarce relative to demand for it and it commands a pnee very
much above what it could earn in agricultural uses The payment which it

receives is thus well in excess of what is necessary to prevent it from trans

ferrmg from urban back to agricultural uses From the point of view, how
ever, of one particular type of urban use, high rents are a transfer payment
which must be paid to keep the land from transferring to other urban uses
Motion picture cinemas for example, provide but a small portion of the
total demand for land in central London if there were no cinemas at
all
,

rentals of land would be about what they are now Thus the cinema indus
whole
try faces a perfectly elastic supply of land m central London and the
of the price thatit pays for its land is a transfer payment which
must be
^
paid to keep the land from transferring to other uses

A DIGRESSION ON THE VALUE OF ASSETS

The above a discussion of the relawan between
analysis leads naturally to
applies
the earning asset and its market price The discussion
power of an
to any can cam a revenue whether it is a factor of production
asset that
(such as land) or a stock or a bond Consider first an example of two pieces
of land of equal size one growing wheat and returning, after costs a pro
h
1 Thus the old examination quest on D scuu ihe view that the price of cinema seats is

m central London because the pnce ofland u high should be answered m the affirmai ve no

the negative as examiners often seemed to expect

 PRICING OF FACTORS OF PRODUCTION 447

per annum; the other growing oats and returning, after all costs,
a profit of ;(^500 per annum. Assuming that the second piece of land is as
good as the first, the person farming it would be likely to transfer it from
oats to wheat production. Assume, however, that this cannot be done or
that the first piece of land is suitable only for growing wheat and that the

second piece is suitable only for growing oats. Under these circumstances,
the first piece of land is clearly worth twice as much as the second piece of
land. If, for example, the two pieces were offered for sale at prices of ,^10,000
and ^7,500, no one would buy the second piece in preference to the first
one. Ten thousand pounds invested in buying the first piece of land yields
£1,000 per year, whereas three-quarters as much money, i.e., ,^7,500, in-
vested in the second piece yields only half as much income, i.e., ;{^500 per
year. Clearly, the first piece of land is a better investment than the second
one, and no one would be prepared buy the second piece. If, however,
to
the first piece were offered for10,000 and the second piece for ;^5,000,
then both pieces would represent equally good investments. The rate of
return per pound invested would then be the same for eaeh piece of land.
When people buy an asset — a piece of land, a government bond, a com-
mon stock - they are investing their money in return for some expected
yield. Self-interest on the part of investors ensures that the relative prices of
various assets will reflect their relative earning powers. If one asset brings
an income twice as big as that brought by another asset, ceteris paribus, the
one asset will command twice the price of that commanded by the other
the case, then a pound invested in either asset will bring the
asset. If this is

same rate of return. Were this not the case, a pound invested in one asset
would bring a higher return than a pound invested in the other and all in-
vestors would prefer one asset to the other.
This comparison between assets can best be made by expressing the return
on each asset as a percentage of the purchase price. In the above example,
the annual earnings are 10 per cent of the purchase price; anyone buying
either piece of land would have an annual income equal to 10 per cent of
his investment.
Now, assume that the general rate of return that can be made on invest-
ments of a given degree of risk is X per cent. In this case, the price of any
single asset of a comparable degree of risk will be determined in the market,
so that it also yields a return of X per cent. If the price were lower, everyone
would rush to buy~this asset rather than the others that yield X per cent; if
the price were higher, no one would wish to buy this particular asset when
all others yielded X
per cent.
Assume, example, that the going rate of return is 5 per cent. A par-
for
ticular asset yielding a net income of £200 per year indefinitely would
command a price of ;^4,000. If the price were less, say, ,^2,000, the earnings
448 DISTRIBUTION

would be in excess of 5 per cent, in fact, they would be 10 per cent and
everyone would rush to buy this asset The competition among potential
purchasers would push up the pnce If the price were higher than £4,000,
say £^8,000, then the yield would be less than 5 per cent - m fact, 2 5 per
cent in this case - and no one would wish to purchase the asset and lU pnce
would have to fall

The price of an asset is directly related to its earning

capacity. If there is competition among purchasers of
assets, asset prices will tend to be set at levels that
yield the same races of return on each asset.

The above discussion of the pneing of assets can be applied to one prob
lem that is extremely important m subsequent parts of this book A bond
ISevidence of debt It usually takes the form of a piece of paper recording a
promise to repay the sum of money ongmally borrowed (the principal) at
some future date, and to pay a stated sum of money each year in the inttiim
(the interest on the loan) A stylised picture of such a bond is presented

here

bond at a
Assume that I loan £\ ,000 to this company by purchasing this

time when the market rate of interest is 5 per cent Now assume that a few
years later the market rate of interest rises to 10 per cent so that firms issuing
new bonds similar to the one illustrated will have to offer £100 per y^r
by way of interest payment Now, if I should wish to sell my existing
£li0w
ond, paying for
£50 interest per year, I will be unable to obtain £1,000
 PRICING OP FACTORS OF PRODUCTION 449

Since investors can purchase a new bond yielding 10 per cent, no one will
purchase my bonds for 1,000, as this will yield them only 5 per cent. Thus
a purchaser would offer me a price sufficiently below 1,000 so that he
would receive 10 per cent on his investment.

A rise in the current rate of interest entails a fall in

the market price of bonds, the
fall being
all existing
sufficient so that apurchaser of an existing bond will
receive the same return on his investment as he could
receive by buying a new bond. Thus the current rate
of interest on new loans and the market price of exist-
ing bonds vary inversely with each other.

Land rents, land values and land taxes; The relation between the
earnings of land and the value of land is just a special case of how the value

of assets is determined. Quite clearly, the larger the yield that can be
obtained for any given piece of land, the higher the rent (in the everyday
sense of the word) that can be obtained for the land’s services, and the
higher the price that can be obtained from selling the land. The sale value
of a piece of land is therefore directly related to the rent that can be asked
for the land ; both of these reflect the same thing, earning power. This is all

very obvious but it is surprisingly easy to forget it in practice. Consider

two separate questions (1) What would be the effects of a 10 per cent tax
;

on all rents paid for the use of land?* and (2) What would be the effects of
a 10 per cent tax on the value of all land in a country ? Most students would
have little trouble with the first question. They would answer somewhat as
follows: since the tax is on all rents the relative profitability of
levied
different uses will be unaffected and thus there will be no allocative effects.
Land will not be forced out of use because land which is very unprofitable
will command little rent and so pay little tax. Thus there will be no change
in the supply of goods which are produced with the aid of land and, since
there is no change in supply, there can be no change in prices. Thus the tax
cannot be passed on to consumers. Farmers will be willing to pay just as much
(and no more) as they would have offered previously for the use of land.
Agricultural prices and rents will be unchanged and the whole of the tax
will be borne by the landlord. The net rents earned by landlords will fall
by 10 per cent and the sale value of land will fall correspondingly.
But now consider the second question. A tax on land values will have an
identical effect to those analysed above! There is no way in which land-
lords can pass the tax on to tenants. The tax will be borne wholly by land-
lords and the same landlords will pay the heaviest taxes as in the case of a

1 Assuming that persons using their own land would have a market rental assessed by the
tax department and the- tax then levied on this assessed rental.

13
450 DISTRIBUTION

tax on because the lands which earn the highest rents will also
rents, this is

be those with the highest values Presented only with the second question,
however, it is easy to forget that earnings and sale values are just twova^t
of expressing the same thing and that the results of a tax on land i-alues

would be the same as a tax on rents

Itshould be noted that the general argument that the tax on rents (or on
land values) is borne wholly by landlords depends on the condition that
the fall m
the price of land will not induce landlords to withdraw land from
the market Under competitive conditions, taxes arc passed on when each
individual producer is induced by the tax to reduce the quantity of the
product which he offers for sale There will thus be a reduction m total

supply, which will cause the price to rise Of course, a tax placed only on
land used to grow wheat would cause landlords to transfer land from wheat
production to other uses, the supply of wheat would be reduced, the pnee
of wheat would rise, and the tax would be passed on, at least in part, to
consumers The supply of land to wheat production is variable, but if the
total supply of land to all uses does not respond to the tax on rentals, then
the tax cannot be shifted by landlords We have already warned against
the assumption that the total supply of arable land cannot be reduced in
supply Some taxes on land might discourage long run conservation policies
so that the total supply of cultivable land would dimmish If this were truci
the prices of agncultural commodities would nsc and part of the effect of the
tax would be passed on to the consumers of agricultural produce

THE SINGLE-TAX MOVEMENT

Taxation of land values has had enormous appeal both intellectually and
politically The peak of its appeal was 80 years ago The ‘single-tax move-
ment’ of which Henry George (1839 1897), an American printer,' was the
guiding genius, caught the imagination of hundreds of thousands of people
on both sides of the Atlantic The policy was to tax away the ‘unearned
increment’ that accrued to landowners What was the appeal ’
The total supply of land in a country tends to be pretty well fixed As
both population and incomes me, there will be an increase in the demand
for land Thu nsc in demand wiH cause an increase in the rents, and, since
values depend on what can be earned, it will also cause an increase m the
values of land ^ Thus the owners of land gam from the natural progress of
society without their having to contribute anything

1 George ran for Mayor of New York Oty in 1836, and very nearly won He campurnf'*
on the issue of the single tax His book Prtgms ond Peterl/ (Appleton 1880) is ptobaWy the
alltime best seller on an economic mue
«
2 See page 487 for a discussion of the sralidity of this prediction that was made bv t

Classical economists of the early mneteenih century

 PRICING OF FACTORS OF PRODUCTION 451

We cannot, on the basis of positive economics, make a moral judgment

on this situation. It is sufficient to point out that many
people have been
incensed at this ‘unearned increment’ and, watching the huge fortunes
accruing to landlords in a rapidly growing society, they have proposed
various measures for removing this ‘unearned wealth’ from the landowner.
It should be noted that the normative issue here is not necessarily one of
the ‘rich landowner’ versus the ‘poor worker’. Many landowners (farmers
and owners of family residences) may be relatively poor, whereas a milhon-
aire may invest in certain assets whose values do not rise automatically with
the progress of society.
The less more enduring appeal of taxes on land values arises
emotive, but
from the economic rent can be taxed away without affecting the
fact that
allocation of resources. Thus, for someone who does not wish to interfere
with the allocation resulting from the free play of the market, the taxation
of economic rent is attractive.' It should be noted, however, that economic
rent is not unique to landlords ; it accrues to the owners of any factor that
is fixed in supply and that faces a rising demand. If there is, for example, a
fixed supply of first-class opera singers in the country, then they will gain in
exactly the same way as landlords do when the society becomes richer and
the demand for operas increases without there being any corresponding
increase in the supply of singers.

I We must be careful not to forget the meaning of our terms. The statement in the text
refers to economic rent as described; it does not refer to the payment actually made by tenants to
landlords. What is called rent in the world is partly an economic rent and partly a return on
capital invested by the landowner. The policy implications of the statement in the text depend
on being able to identify economic rent in practice. At best this is a very difficult thing to do at
;

worst, it may be quite impossible.

 CHAPTER 34

LABOUR UNIONS, COLLEC-

TIVEBARGAINING AND THE
DETERMINATION OF WAGES

Why do engineers get pretty much the same pay

for thetame work no
matter where they are employed
in the UK’ Why do charwomen and
^nnies get very different rates of pay
m different parts of the countr)’
Why do coal miners, who work in a declining
industry, get higher rates of
pay t an equally skilled
workers in many expanding industries’ How does
a worker in a plant employing
5,000 men ‘ask for a nse’ ’ How does he let
his employer know that
he would be glad to trade so many pennies per
hour m wages for a better pension
scheme’ Why do strikes occur’
The competitive theory of factor price
determination, which was the
su ject of the last
chapter, yields extremely useful predictions about factor
prices factor movements, and the
distribution of income, many of which
ave ecn tested and
confirmed Indeed, for the pricing of many non
there is little need to discard
or to modify the competitnc
mo e uch of what we observe about labour
markets is also consistent
with the theory But
not all of it is
Labour is in many ways
the exceptional factor of production The forces
at govern a man s
working conditions and pay are vitally important to
imse and to his family
When employees and employers negotiate over
e price to be paid
to the factor of production called labour, they are
negotiating over something
that is vital to most households It is not sur-
nsing t at people
arc sometimes prepared to fight over such negotiauons
alters other than
matenal advantage enter into the relationship between
E 'Ver and employee, for it
is a relation between people to whom quab*
^ sue as loyalty,
fairness, appreciation andjustice are at least as important
s pro uctivity,
and attitudes about these qualities can vitally affect
orodurnvu,, ^
 BARGAINING AND DETERMINATION OF WAGES 453

Labour unions, employers’ associations and the institutions and customs

governing collective bargaining are features of the real world that have
developed in response to the exceptional conditions governing the bargain-
ing between free men about the terms on which one \\'ill work for another.

If these insdtutions are important determinants of what happens, it must

be because the nature of the institutions influences the wages and working
conditions that are finally agreed upon.

THEORETICAL MODELS OF FACTOR MARKETS'

It is often useful to approach the complexities of the world by examining
extreme cases. In this section, we shall look at four of these.

1 Many buyers, many sellers (competition): Suppose there are so

many employers many employees providing it that no
hiring labour and so
single employee or employer can exert any appreciable influence on the
wage This is a situation in which the competitive theory of wage de-
rate.
termination, outlined in the previous chapters and illustrated graphically
in Figure 34.1 (i), applies. The equilibrium wage is w,. At that point, the
demand for labourers and the supply of labourers are equal at Oq^.

2 One seller, many buyers (monopoly): Suppose there is a single

labour monopolist who controls the entire supply of some kind of labour.
The cost to him of supplying different quantities of labour is assumed to be
a rising function of the quantities supplied. The curve iS" in Figure 34.1(ii)
reflects the marginal cost to him of supplying different quantities of labour.
He is in precisely the position of the monopolist we studied in Chapter 23.^
Knowing that the demand cur\'e for labour slopes dowmward, he will re-
strict the quantity supplied to the point at which the marginal revenue from
the last unit of labour he provides is equal to the marginal cost of supplying

In terms of Figure 34.1 (ii), he chooses to supply the quantity 0^2

it.

wage rate 0w2- Wages are higher than in the competitive case, and the
quantity of labour supplied is less. (If you do not understand this result, you
should re-read Chapter 23.) The labour union, as we shall see, may be an
instrument for exercising monopoly in factor markets.

1 The following theories apply to any factor, but we shall discuss them in the context of
labour because it is labour that concerns us in this chapter.
2 This is not as arbitrary an assumption as it might first appear. Standard theory of house-
hold behaviour predicts that the supply curve will reflect the marginal cost to the workers of
the last unit of labour provided (since the household would go on providing labour until the
wage paid for the last unit supplied just compensated for the disutility - marginal cost - of
providing it).
 CHAPTER 34

LABOUR UNIONS, COLLEC-

TIVE BARGAINING AND THE
DETERMINATION OF WAGES

\V;n do engineers get much the same pay for the same work no
mitter where ihc> tre employed m the UK’ Why do charwomen and
nannies get \ery difTcrcnt rates of pay in different parts of the country’
Why do coal miners wlio work m a declining industry get higher rates of
pay than equally skilled workers m many expanding industries’ How does
a worker in a plant employing 5 000 men ask for a rise ’ How docs he let
his employer know that he would be glad to trade so many pennies per
hoin m wages for a better pension scheme ’Why do strikes occur’
The compctitnc theory of factor price determination which was the
svibjcct of the yields extremely useful predictions about factor
last cli iptcr
^*»
5 nccs, fictor mosements and the distribution of income, many of which
and confirmed Indeed for the pricing of many non
f'd

1 there is little need to discard or to modify the compctitwe

ich of what we obserac about labour markets is also consistent
c theory But not all of it is
our is m many w ays the exceptional factor of production The forces
govern a man s working conditions and pay are vitally important to
'f and to his family When employees and employers negotiate over
a'lct -prn.-c \a> k* -padi ao ahjc 'iacvor o1 pro&uction fvficfi ’la'Dour, ‘fiiey
negotiating over something that is vital to most households It is not sur
prising that people are sometimes prepared to fight over such negotiations
Matters other than material advantage enter into the relationship between
cmplovcr and employee for it is a relation between people to whom quali
tics such as loy alty, fairness appreciation and justice are at least as important
as productivity, and attitudes about these qualities can vitally affect
productivity
 BARGAINING AND DETERMINATION OF WAGES 453

Labour unions, employers’ associations and the institutions and customs

governing collective bargaining are features of the real world that have
developed in response to the exceptional conditions governing the bargain-
ing between free men about the terms on which one wU work for another.
If these institutions are important determinants of what happens, it must
be because the nature of the institutions influences the wages and working
conditions that are finally agreed upon.

THEORETICAL MODELS OF FACTOR MARKETS^

It is often useful to approach the complexities of the world by examining
extreme cases. In this section, we shall look at four of these.

1 Manybuyers, many sellers (competition!; Suppose there are so

many employers hiring labour and so many employees pro\'iding it that no
single employee or employer can exert any appreciable influence on the
wage rate. This is a situation in which the compedtive theory of wage de-
termination, oudined in the pre\'ious chapters and illustrated graphically
in Figure 34.1 (i), applies. The equilibrium wage is uq. At that point the
demand for labourers and the supply of labourers are equal at O^j.

2 One seller, many buyers monopoly) Suppose there is a single :

labour monopolist ^vho controls the entire supply of some kind of labour.
The cost to him of supplying different quantifies of labour is assumed to be
a rising function of the quantities supplied. The cur% e S in Figure 34.1(ii)
reflects the marginal cost to him of supplying different quantities of labour.
He is in precisely the position of the monopolist wc studied in Chapter 23."
Knoufing that the demand curve for labour slopes downward, he -^s-ill re-
strict the quantity supplied to the point at \\ hich the marginal revenue from

the last unit of labour he prowdes is equal to the marginal cost of supphfing
it. In terms of Figure 34.1(ii), he chooses to supply the quantity 0^2
wage rate Ow-,. Wages are higher than in the competitive case, and the
quantity of labour supplied is less. (If you do not understand this result, you
should re-read Chapter 23.) The labour union, as we shall see. may be an
instrument for exercising monopoly in factor markets.

1 The fo!lo\fing theories apply to any factor, but we shall discuss them in the conte-xt of

labour because it is labour that concerns us in this chapter.

2 This not as arbitrary an assumption as it might 6rst appear. Standard theory of house-
is

hold behas-iour predicts that the supply cun-c will reflect the marginal cost to the workers of
the last unit of labour provided (since the household would go on providing
labour until the

wage paid for the last unit supplied just compensated for the disutility - marginal cost - of

providing it).
454 DtSTBIBUTION

3 One BUYER, MANY SELLERS (monopsony) Supposc that a monopsoDist,

the sole employer of labour in a certain town, can offer any wage rate he
choosesThe labourers can cither work for him or seek employment else-
where The supply curve shows how much labour he will get at any given
wage In Figure 34 l(iii), D is the monopsonist’s demand for labour based
‘
upon his marginal revenue product curve

For any given quantity that is purchased, the supply curve shows the
pnee per unit that must be paid, to the monopsonist, this is the average cost

curve of the factor

total cost
Average cost of a factor = ;
number of units purchased
= price per unit of the factor

But the marginal cost of employing extra units of the factor must exceed the
average cost If, for example, 100 units are employed at lOr per hour, then
total cost IS ;(|50 and average cost per unit is lOf If 101 units are employed

and the factor price is driven up to 1 Or 6d, then total cost becomes 0/ Qd,
the average cost per labourer is lOr 6d, but the total cost has increased by
0 an6das
ing
Oj a result of hiring one more labourer
extra labourer will exceed the
The marginal cost of obtain-
wage paid, because the increased wage
rate necessary to attract him must also be paid to all the labourers already
employed Thus, 34 l(ui), we can draw a marginal cost curve for
in Figure
labour that will above the average cost curve The demand curve is the
lie

marginal revenue product curve and the curve just derived is the marginal
cost curve The proht-maxtmtsing monopsonist will equate marginal cost of
labour with its marginal revenue product, in other words, he will go on
hinng labour until the last unit increases total costs by as much as it increases
total revenue The equilibrium employment will be Oq^ and the wage rate
Owi From the above discussion we have denved the following prediction
Monopsonistic conditions in the factor market will
result in a lower level of employment and a lower
wage rate than would rule when the factor is pur-
chased under competitive conditions.
The reason for this is that the monopsonistic purchaser is aware that, by
trying to purchase more of the factor, he is driving up the price against
himself He will, therefore, stop short of the point that is reached when the
factor IS purchased by many different firms, no one of which can exert an
influence on its price
If we m
compare factor monopoly, Figure 34 1 (ii), with factor monopsony,
in Figure 34 !{in), we see that wages under monopoly will be higher than
1 If you are not clear about this important concept you should review pages 414-15
 BARGAINING AND DETERMINATION OF WAGES 455

under monopsony {OW2 compared to OW3). We can make no comparative

prediction about the level of employment. In either case, employment will
be less than under competitive conditions. (We have draivn the diagram so
that 0q2 is the same as Oq^, but by varying the shapes of the demand and
supply curves you can alter this result.) The wage rates differ in the two
cases because, in the monopoly case, the seller of labour is in a position to
restrict the supply to Oq2 and, thus, to receive the wage OW2 that purchasers

Wage

Wage

Fig 34.1 Four types of factor markets

(i) Many buyers, many sellers (competition).

(ii) One seller, many buyers (monopoly).

(iii) One buyer, many sellers (monopsony).
(iv) One buyer, one seller (bilateral monopoly).
456 DISTRIBUTION

Will pay quanUty of labour, in the monopsony case, the buyer

to hire that
will hire only quantity Oq^ of labour and the supply curve shows that he
can get this quantity by ofTenng the low wage OW 3 We predict that
Monopsonistic conditions in the factor market will
result in a lower wage rate than would rule under
monopolistic conditions

4 One buyer, one seleer (biiateral monopoly) In bilateral

monopoly a monopolist sells to a monopsonist The monopolist wants to
supply 0^2 at wage rate Qwj, The monopsonist wants to hire Oq^ at wage
rate Owj If both are stubborn 10 the point where neither will yield a penny,
no labour will be bought or sold but this is not a likely result Instead we
may expect the buyer and seller to bargain Suppose (to keep things simple)
they agree first on a quantity of labour to be used represented in Figure
34 1 (i\ ) as Oqi — Oq^ The supplier will not accept any wage below Ow^
The demander will not pay any wage above Ou^ But for any wage between
these limits each of them is better olT than if they did not strike a bargain
Our theory leads to the following prediction

The wage rate in bilateral monopoly will be some-

where in between the monopoly wage rate and the
monopsony wage rate, but the theory does not predict
where it will be
Given our present theory the wage rate is indeterminate within a
specified interval To predict more preciscl> where the wage rate will be set
would require a further theory Such a theory might be based upon such
things as the bargaining skill of the monopolist and the monopsonist, on the
resources of each to hold out until the other gives in on the presence of
third parties who might mediate or arbitrate the disagreement and so on
Such matters are left to a more advanced course

Which Market Form Best Describes the Labour Market Today?

Monopoly and monopsony in the sense of only one seller and only one buyer
act Ca'.cly cact tjiw. a not/ ftiw. vt'wht'n, hi wMaIv
to be the only employer to whom the workers may turn We shall, as we did
•with monopoly in the product markets speak of monopoly power and
monopsony power as matters of degree, and be particularly concerned with
relative strengths of buyers and sellers
The answer to the question of which pattern exists today is ‘All four
depending upon where you look In the UK 40 per cent of the labour force
unionised while m the US the figure IS only 25 pet cent Furthermore the
 BARGAINING AND DETERMINATION OF WAGES 457

extent of the economy covered by a single wage bargain varies from coun-
try to country: for countries where collective bargaining is the predominant
way of adjusting rates
of pay, the areas covered by the typical bargain differ
widely. In Scandinavia, Western Germany, the Netherlands and the
United Kingdom, for example, the typical area is that of the single industiy".
In France, on the other hand, collecti%'e bargaining is predominantly
regional,and in the United States more than four-fifths of the collective
agreements in existence in a recent year had been negotiated with single
employers.^ In some sectors of the economy, such as agriculture, there is no
labour organisation to speak of, and among white-collar ivorkers union
membership is low. The market for agricultural labour and for clerical
workers is reasonably close to that of the competitive model with many
and many buyers. In many of the seiA'ice trades large employers are
sellers

dealing with unorganised workers, and the situation is closest to that of the
monopsony model - with the important modification that there is a mini-
mum wage. In some industries, especially the building trades, textiles, and

certain parts of transport, unions are very strong, and employers are very
weak. The balance of power lies on the union side, and the situation is close
to that ofmonopoly. For a vast sector, including most of manufacturing,
strong unions are opposed by strong employers or employers’ associations
and the conditions of bilateral monopoly come close to applying.
Because there is so much variety in the form of labour markets, there is

no wage determination. A detailed

single pattern of discussion of patterns
of bargaining and wage determination must be left to a more specialised
course.

THE VOCABULARY OF LABOUR ECONOMICS

The vocabulary we use in this chapter comes partly from theoretical
economics (‘monopoly’, ‘monopsony’, etc.) and partly from labour eco-
nomics (‘employers’ association’, ‘collective bargaining’, etc.). These latter
terms are usually self-explanatory, but occasionally they are not, so that a
slight digression on terminology is required.

1 Potential .monopsonists: firms and employers’ associations: We

have encountered the term fiTm many times before. The large firm has some
degree of monopsony power by virtue of its size and the number of em-
ployees with which it deals. It recognises that its actions affect the wage rate,
especially the rates received by those kinds of labour that are in
some way

1 See E. H. Phelps-Brown, The Economics of Labour, Yale

University Press, New Haven, 1962.

15
458 DISTRIBUTION

which the firm operates This limitation will result

limited to the industry in
when one large firm’s need for this kind of labour represents a large fraction
of the total demand for it The British Coal Board exercises enormous
monopsony power in its dealings with miners, because it represents a lai^e
share of the demand for miners in the country
Employers' assoctations are groups of employers who band together for the
purpose of adopting a common policy in labour negotiations If all car
manufacturers offer the same terms, they exercise more monopsony power
than any one can exercise on its own The car worker who docs not wish to
accept the common conditions has no alternative but to seek work in an-
other industry, and he is likely to be much less Well-equipped for work ma
different industry

2 Potential monopolists unions No one bothers to define unions

any more, perhaps because every one knows what they are, or perhaps
because a union is so many things a sooal club, an educational instrument,
a political club, one more source of withholding money from a worker’s pay,
a bargaining agent for an individual worker, and, to some, a way of life
For the purposes of our discussion of labour markets, a union (or trade union,
is an association of individual workers that speaks for them
or labour union)
in negotiations with their employers
Unions today have two different principles of organisation the trade
(or craft) unions, in which workers with a common set of skills are joined in
a common association, no matter where or for whom they work, and indus-
trial unions in which all workers in a given plant or a given industry are

collected into a single union, whatever their skills

Industnal unions became common m the United States with the rise m
the 1930’s of the Congress of Industrial Organisations (CIO) Among the
prominent industrial unions in the US are the United Auto Workers and
the Steel Workers Unions The existence of these industrial unions means
that the great automobile and steel companies have only one union to deal
with A single agreement over wages, working conditions or union practices
is sufficient to change the situation throughout the entire firm
A single union covering an entire industry is less common in the UK
and in many other countries than it is m the US When the employer has to
deal with many unions, and twenty or more unions are not uncommon
within a single firm in the UK, agreement between labour and management
can be very hard to reach and sometimes jurisdictional disputes break out
over which union is to be responsible for which jobs An experienced observer
of the Bntish industnal scene has said ‘On occasion, new capital instru-
ments have been inadequately used and production has been held up by
disputes among rival unions about the types of workers to be employed on
 BARGAINING AND DETERMINATION OF WAGES 459
new operations or new materials. The shipbuilding, printing and building
industries have provided examples of this obstructionism.’^

3 Kinds of bargaining arrangements: open, closed and union

shops: In an open shop, a union represents its members, but does not have
exclusive bargaining jurisdiction for all the workers of its kind. Member-
ship in the union is not a condition of getting or keeping a job. Unions feel
that this represents an impossible situation. If, on the one hand, the em-
ployer accedes to union demands, the nonmembers achieve the benefits of
the union without paying dues or sharing the risks or responsibilities. If, on
the pther hand, the employer chooses to fight the union, he can run his plant
with the non-union members, thus weakening the power of the union
members in the fight.
In a closed shop, only union members may be employed and the union
controls membership however it sees fit. Employers sometimes regard
its

this as an unwarranted limitation on their right to choose their employees.

In a union shop, the employer may hire anyone he chooses, but every em-
ployee must join the union within a specified period.

4 Weapons of conflict: strikes, picket lines, boycotts, strike-

breakers, LOCK-OUTS, BLACK LISTS: The strike is the union’s ultimate
w'eapon. It consists in the concerted refusal to work of the members of the
union. It is the strike or the threat of a strike that backs up the union’s de-
mands in the bargaining process. Workers on strike are, of course, off the
payroll, and many unions set aside a portion of the dues collected to have a
fund for paying striking workers. Picket lines are made up of striking workers
who parade before the entrance to their plant or firm. One of their objec-
tives is to get public opinion on their side; another is to force a total shut-

down in the plant’s operation. Other union members will not often ‘cross’ a
picket line. This means that if bricklayers strike against a construction firm,
carpenters may not work on the project although they may have no griev-
ance against the firm, nor will any lorry driver deliver supplies to a picketed
site.Pickets represent an enormous increase in the bargaining power of a
small union. (Much of the bitterness against jurisdictional disputes arises
from the fact that an employer may be unable to settle with either union
without facing a picket line from the other union.) A boycott is an organised
attempt to persuade customers to refrain from purchasing the goods or
services of a firm whose employees are on strike.
The lock-out is the employer’s equivalent of the strike. By closing his plant,
1 G.C.AIlen, The Structure of British Industry, Longmans, London, 1961, page 189. Such
disputes probably receive a degree of publicity out of proportion to their actual harmful
effect

on the economy.
460 DISTRIBUTION

he locks out the workers until such tunc as the dispute is settled Slnke
breakers are workers who are brought in by management to operate the plant
while the union is on stnke A black list is an employers’ list of workers who
have been discharged for union acUviUes, and who are not supposed to be
given jobs by other employers

5 Collective bargaining This term is used to desenbe the whole

process by which unions and employers (or their representatives) arrive at
and enforce agreement

THE EVOLUTION OF THE MODERN UNION

When representatives of the engineering union sit down this year with repre-
sentatives of the employers to discuss wages, contributions to pension funds,
number and length of holidays and other issues, they arc engaged in what
has been termed mature collective bargatntng At the end of the negotiations,
the newspapers show the smiling representatives shaking hands, and each
of the members of the union then knows (he conditions under which
he will work Should an employee have a gnevance
for the next year or so
at any time, he reports it to his union representative (called a shop steward),
and a carefully designed procedure is set in motion to settle the dispute
Unionism today is both stable and accepted It was not always so As
recently as 30 yean ago unions were fighting for their lives, and union
organisen and members were risking theirs Why the change and how did
It come about
^

The Urge to Organise

Trade unionism had its ongm in the pitifully low standard of living of the

average nineteenth-century worker and his family Much of the explana-

tion of the low standard of living throughout the world lay in the small size
of the total national output relative to the population Even in the Wealthiest
countries, an absolutely equal division of national wealth among all families
in 1850 would have left them all m
poverty by our present standards
Poverty had existed for centuries It was accentuated, however, by the
twin processes of urbanisation and mdustciaUsation The man who was
moderately content working his land usually became restive and discon
tentedwhen he moved into a grimy, smoky, nineteenth century city, took
employment in a sweatshop or a factory, and settled with his family in a
crowded, unsanitary tenement ‘ The focus of his resentment was his
employer
1 Many of them moved because they had no choice having been dnven off their land by the
entlosuie movement
 bargaining and determination of wages 461

The employer set the wages, and the wages were low. He was often
arbitrar)'and seldom sympathetic, and was usually conspicuously better off
than his employees. The unhappy worker had, of course, the right of all
free men to quit his job — and star\'e. If he grumbled or protested, he could be
fired, and worse still black-listed, which meant no one else %vould hire him.

Stories of the workers’ verj' real suffering during the industrial revolution
could fill many volumes, but an-example will at least illustrate some of the
real horrors that lay behind the drive for change and reform.
In the cotton-spinning work, these creatures [the workers] are kept, fourteen hours
in each day, locked up, summer and w-inter, in a room heat from eighty to eighty two
degrees. The rules which they were subject to are such as no negroes [i.e., slaves] were
ever subjected to. The door of the place wherein they work, is locked, except half an
. . .

hour, at tea-time, the workpeople are not allowed to send for water to drink, in the
hot factory; even the rain water is locked up, by the master’s order. ... If any spinner
be found with his window open, he is to pay a fine of a shilling! ... for a large part of
the time, there is the abominable and pernicious stink of the gas to assist in the
murderous effects of the heat. the notorious fact is, that well constituted men
. . .

are rendered old and past labour at forty years of age, and that children are rendered
decrepit and deformed, and thousands upon thousands of them .slaughtered by con-
sumption [tuberculosis], before they arrive at the age of sixteen. . .
.*

Out of these conditions came the full range of radical political movements
from revolutionary socialism, which today we call Marxism or Communism,
to Fabian socialism, which tried to effect change gradually through existing
political systems. Out of them also came the union which was to some extent
a club providing some meagre protection for those workers who were unem-
ployed, disabled or retired, and to some extent a bargaining agent. Unions
were for a long time resisted by the full power of both the employers and
the central authorises.
The worker men acting together had more in-
perceived that ten or 100
fluence than one acting alone, and he dreamed of the day ^vhen all the
workers would stand solid^ against the employer. The union was the
organisation that would provdde a basis for confronting the monopsony
power of employers wdth the collective power of the workers. But it was
easier for the worker to see where the solution to his problems lay than to
achieve this solution. Organisations of workers would hurt the employer,

1 William Ctobbett, Political Register, vol. LII, 20 November 1824, as quoted by E.Rot-ston

Pike, Human Documents of the Industrial Revolution in Britain, Allen & Unwin, London, 1966,
pages 60-1. This fascinating book chrom'cles some of the most common horrors of the nine-
teenth-century industrial revolution. Every' student of society should spend at least one
evening brow'sing its pages. When he is finished he should have little trouble understanding the
militant reforming attitudes of early Marxists and trade unionists. Whatever he ma> think of
they were tiy'ing to
the solution offered by the Marxists he sviU hardly be able to doubt that
deal with, rather than to ignore as so many others did, a very real human problem.

2 The word solidarity figures importantly in the literature and songs of the labour movement.
462 DISTRIBUTION

and the employer did not sit by idly He too knew thatm union there was
strength ‘Agitators’ who tried to oi^anise other workers were fired and
black listed, and in some cases beaten and killed

Requirements of a Successful Union

In order to realise the dream, it was necessary, first, for the unions to gam
effective control over the supply of labour, and, second, to have the financial
resources to outlast the employer m a struggle of strength There was no
‘right’ to organise , the unions had to force the employers to do business with
them
The first unions were unions of the highly skilled, for obvious reasons
organise the unskilled or the semi-skilled and the employer could find re
placements for them But the skilled workers - the coopers, the boot-
makers, the shipwrights were another matter There were few of them,
and they controlled the access to their trade by controlling the conditions of
apprenticeship The original unions were m effect closed shops one had to
belong to the union to hold a job, and the union set the rules of admission
But control of supply was only half the battle Most employers detested
unions and regarded them as an infringement of their rights Employers
in general opposed unionisation and fought it with great bitterness
Such success as unions had was m
occupations both small and strategic
The costs of giving in to such a union were relatively small compared to the
costa of a protracted strike m such cases unions had their ups and
'
Even
downs When employment was full and business booming the cost of being
fired for joining aunion was not so great, for there were other jobs During
periods of depression and unemployment, however, the risks were greater
The individual worker knew that other unemployed members of his trade
would be there to take his job if he caused trouble Slowly over the nine-
teenth century unions grew in size and power but not at a steady pace A
clear cyclical swing in membership is observable with gams in periods of pros
perity and setbacks in periods of business depression By the intcrwar period
the union had established itself as an integral part of the industrial scene

Who Controls the Modern Union and to What Ends?

In most countries, unions have been growing in size and the relations be-
tween the rank and file members and the union leaders have necessarily
been becoming more and more tenuous In the UK in 1900 there were 1,323
I The pointimportant If a certain group of key worken accounts for only 5 per cent of
is

the total labour cost giving them a 10 percent increase in wages increases total costs by only
0 5 per cent See page 409 cm the relation between the dasticity of demand and the importance
of the factor in the total costs of production
 BARGAINING AND DETERMINATION OF WAGES 463

unions with a membership ofjust over 2 million workers while in 1958 there
were only 657, but total membership had risen to million workers.
Furthermore a mere seven unions accounted for almost half the total
membership and 17 unions for just over two-thirds of it.' This has led to
some divorce between ‘ownership’ and control in the unions and sometimes
to the seizing of control of a union by a minority group that did not reflect
the real views of the rank and file of the members. Such minority control
by people out of step with the rank and file is still, however, the exception
rather than the rule.
Unions, though democratic, do tend toward one-party democracy in the
vast majority of cases. Union leaders are true professionals whose main
business is to run the union, whereas the main business of the union member
is to earn his living in his job. The typical union member’s indifference is

understandable; he is paying dues that permit the union to pay generous

salaries to union leaders to look out for his interests - and as long as the
leadership ‘delivers’, all goes well. But delivers what, and to whom?
What are the goals for which the union leaders strive? Such goals are
many and we can mention here only a few of the most important ones to
give an idea of some of the problems involved.

Wages versus employment: How is an increase in the wage rate

secured ? The union leader has substantial power in determining the wage
rate, but he can only ignore at his peril the fact that the firm’s demand
curve for labour is probably downward-sloping, as in Figure 34.2. If the
wage rate is raised from Owi to 0 w 2 the level of employment will fall from
,

Ogi to 0^2 there will then be union members who would like to work a't
i

this wage, but who cannot find employment. The excess supply of labour

is 5 2 ? 3
'
The members of the union who remain employed will be much
-

better off, but those who lose their jobs mil not be. Should the union strive
to maximise the total earnings of the group that remains employed, or the
total employment of its present members?
This choice iS illustrated in Figure 34.2. If the union is forced to accept

the employer’s demand employment,

curve for labour as determining total

then the competitive wage Owi gives the highest possible volume of em-
ployment, Oqi- If the union wishes to maximise the total earnings of all
employed workers it would pay them to raise wages and restrict employ-
ment if the demand curve were inelastic at the competitive equilibrium.

The union can always income per head of employed population

raise real

by restricting employment to some point such as 0g2 in Figure 34.2.

In some situations it may be possible to avoid this choice between high
wages and employment by bargaining over both wages and employment.

1 Facts quoted by G.C. Allen, op. cit,, page 179.

464 DISTRIBUTION

The nse of the guaranteed-annual-wagc type of agreement the m

in US
recent years represents an attempt to dojust this This situation is illustrated
by point 2 in Figure 34 2 The demand curve shows for each wage rate the
amount of labour the employer would like to hire But he may prefer to
hire some other amount rather than to go without labour altogether When
wages are Owi and employment ts Oqi, the union might offer the em-
ployers the alternative of employing Oj| labour at a wage of Owj or of
facing a strike If the employer accepts the former alternative, then he will
move to point z, which is off his demand curve In this case, the union
raises the per capita wages and total real earnings of its members without
causing any reduction in employment The union’s success in pursuing

Fig 34 2 Alternative Union

Policies m a Labour Market

such a policy will probably depend on the size of the profits in the industry
(i e,on the extent to which the industry departs from the perfectly com
petitive equilibrium) and on the willingness of the employer to go along
with the union

Wages versus nonwage benefits Settlements of a contract negotiation

are often m the following form Tr per hour of which 9d was m wages, 2dm
increased contributions to the union’s welfare and pension fund, and the
remainder m other "fringe benefits”, including increased holidays and sick
leave ’
The most important figure to the employer is the size of the total
package - Ir per hour in this case for contributions to pension funds add
to his costs in just the same way as payments in the form of wages But there
are advantages to the employer in giving indirect, or fringe, benefits that
may lead him more than he otherwise would One advantage is
to grant
that the real cost of fringe benefits per hour is often difficult to estimate
Although the union may claim
it has got Ad worth, the employer may really

believe it will come to less A second advantage is that some forms of fringe
benefits (such as pension funds) tend to bmd the worker more closely to the
company, and thus decrease the ‘turnover rate’ among his employees
 BARGAINING AND DETERMINATION OF WAGES 465

Finally some forms of payments (such as pay to workers who are laid off
,

during seasonal slumps, etc.) may be easier to justify to stockholders or

to
other employers.
From
the employees point of view, fringe benefits also have some appeal.
Many of them are not taxable as current income. Pension funds and the
like allow employees to provide for their future and for that of
their
family.
Most important, for both union and employer, fringe benefits provide
scope for bargaining. A union official may have promised his members not
to take less than and the employer’s representatives may have assured
Ir 3(/,

their directors that theywould stand firm at 9d. In the negotiation, they
may agree on 8d in wage increases and Id in fringe benefits. Both negotiators
may claim success, whereas either would feel reluctant to accept a straight
Ir wage increase.
By and large, the
union’s choice between wages and other benefits comes
from the union’s leadership, not from the rank and file, but strong feelings
on the subject do influence policies. Union members will accept decisions
on these matters from their leaders that they would not accept if such
decisions originated with their employers.

Wages versus job security: Unions var>' greatly in the extent to which
they adopt a defensive or an offensive attitude to the labour market. The
people who were in their twenties during the great depression of the 1930’s
are now in the age group of 55-65 that dominates the leadership of unions,
management and government. Not surprisingly, the labouring members of
this age group have been strongly conditioned to a defensive attitude to
jobs. They lived through a period when unemployment never fell below
20 per cent of the labour force as a national average and when it was nearer
50 per cent in many of the hardest-hit areas. They saw people grow up,
marry and raise children on the dole. They saw young people who were
eager to work but were totally unable to find any form of employment
slowly have their spirit broken as they had to confess their apparent im-
potence to wives and children. They suffered the humility of being read
lessonson hard work, thrift and patience b>' a London-based, middle-class
bureaucracy that had itself never tasted the bitterness of returning home to
say once again ‘No one wants my work’ or of saying to a son ou can t stay'
on at school because we need the few shillings you might earn now as a
labourer’. When occasionally' they rose up in their misery' in a general strike

or a hunger march the troops were called against them.

In such a situation of mass unemployment the installation of a new
machine in a factory' that removed someone’s job condemned the job holder
to a possible indefinite future of existence on the dole. It is little wonder that
466 DISTRIBUTION

ncH machines and rationalisation that might reduce costs by removing

jobs were opposed bitterly and that job saving restrictive practices were
adhered to with tenacity In a penod of heavy unemployment techno-
logical change is likely to mean unemployment for those whose jobs are lost
by the change
The defensive attitude which was so understandable in the 1930 s sur-

vived (not surpnsingly) into the post-war period in which circumstances

werevery different The post-war penod was a time of full employment and
a time in which new jobs were available to replace old ones that were
destroyed by technological change In such a world the attitude of defending
existing jobs at all costs made much less sense than in the 1930’s, and,
furthermore, under the impact of a rate of growth of real income per head
of population of between 2 and 3 percent per year, it was doomed to failure
Growth nccessanly means change both in the techniques of production and
in the pattern ofdemand, and the defence of all existing jobs is doomed to
failure in such achanging world Slowly over the last 20 years the attitude
of many unions has changed from one of resisting technological change to
one of collaborating with it and of trying to reduce some of its costs to
individuals who are adversely affected by it The acceptance of such change
IS then used as an offensive base from which to launch a concentrated

attack for higher wnges and better working conditions We must never
forget, however, that as long as people are alive who suffered through the
thirties, the heritage of the mismanagement of that period will leave a set
of defensive attitudes m many which it is better to try to undentand than
merely to condemn

UNIONS AND THE GENERAL LEVEL OF WAGES

Union members and do raise wages But do they

believe that unions can
one of the most debated questions in all of
really ^ Surpnsingly, this is
economics, despite the fact that a great deal of empirical work has been
done on the subject We cannot settle the issue here, but we can discuss why
It anses at all

In the previous sections we have considered what can be done to raise

wages by a single union which u small tn relation to the whole economy This
Italicised clause is the key to being able to use demand and supply curves
The student must beware of generalising the conclusions to apply to the
economy as a whole If we pose the question can unions raise the real wage
rate of all workers in the country^ we must beware of drawing the con-
clusion that unions can raise wages only at the expense of a reduction m
overall employment We cannot use the earlier conclusion (see page 463)
 ;

BARGAINING AND DETERMINATION OF WAGES 467

because it was based on demand curves which are in turn based on assump-
tions of other things being equal. If unions raise the ^vage rate of a significant
part of the labour force they will cause incomes to change significantly; this
will cause demand curves for consumers’ goods to shift; this will
cause
outputs to var^^5 and this, in turn, will cause shifts in the demand curves for
labour. Unless we have a theor>' of how each of these changes is related to
the other, we cannot attempt an answer to this question. In fact there is no
generally accepted, well-worked-out theory' which would allow us to deal
with this important question. Here then is a real challenge to the student
interested in questions of labour and income to develop and test a theory
:

which would shed light on this important question.

A study of Figure 34.3 may however make us a bit sceptical of the propo-
sition that unions have had an enormous effect in shifting the distribution
of income towards labour. If labour as a whole is to gain relative to other
groups, its share of the total of national income must grow. The fact is,

however, that the share of income going to labour has not shown any
marked long-term tendency to increase as is shown in Figure 34.3.
Of course, as we have seen, not everyone is unionised. Do union members
do better than workers in unorganised industries? Again, the evidence is
inconclusive. Workers in many nonunion trades have achieved larger per-
centage increases in pay than union members since 1 930. But, of course, what
unions achieve affects labour markets and labourers in other industries as
well. The American economist Albert Rees, one of the leading students of
the influence of unionism on wages, concluded, T would say that perhaps a
third of the trade unions have raised the wages of their members by 15 to
20 per cent above what they might be in a nonunion situation, another
third by 5 per cent to 10 per cent, and the remaining third, not at all’.^
Since the Second World War, there has been a narrowing of wage differen-
tials for various grades of labour and the unions have had something to do

with this. There are two extreme views about the influence of unions. The
first is that unions have complete power over wage differentials; wage

differentials are, according to this view, a function of union choice. The

second view that unions exert no effective influence on wage differentials
is

in spite of collective bargaining, relative wages come out to be pretty well

what they would be in a freely competitive market. It is important to realise

that each of these views constitutes an hypothesis that can be subjected to

empirical test. That the conditions of demand and supply exert some influ-

ence on relative wages, independent of the wishes of unions, has been pretty
conclusively established. What has not really been established is how
much
independent influence unions are able to exert on the structure of relative
wages.
1 Albert Rees, Wage Inflation (National Industrial Conference Board, 1957).
 BARGAINING AND DETERMINATION OF WAGES 469

THE PLACE OF UNIONS IN THE MODERN ECONOMY

If we are uncertain about precisely how and how much unions have affected
the wages of their members, there is little dispute over some other facts.

Unions have clearly been a vehicle for improting both the working con-
ditions and the dignity of the working man. They have brought a ver\'
substantial equality of bargaining power to labour and management and
in many cases they have brought order and stability to the process of collec-
tive bargaining. Modem labour-management relations make the ‘class
struggle’ between the worker and the capitalist seem mild indeed, and dire
predictions of an eventual armed conflict no\s’ seem quaint and unreal.
All of the above notwithstanding there is a fairly strong feeling in today’s
world that unions often abuse their power. This belief leads to various policy
conclusions ranging from the relatively moderate one that some curbs
should be placed on union power to the extreme one that unions have com-
pletely outlived their usefulness and should be made illegal. IVhatever the
final outcome of the debate about any harmful effects that unions may
have by slowing down the institution of new growth-creating innovations
or by contributing to inflation through ‘immoderate’ wage demands there
can be no question of the need for unions in the modem world. The growth
of large firms has created a need for organisation on the labour side if only
to express the desires of labour (e.g., a preference between higher wages and
better working conditions) and to bargain with the employer over them.
It is quite unrealistic to think that every member of the labour force of
ICI or BMC could bargain indmdually with the management on these
matters. Also as oligopolistic industrial structures become increasingly com-
mon, employers obtain real power over the labour market. An opposing
power group on the side of the workers is the only guarantee that wages will
not be depressed as in the monopsony case uith which we began this chapter
(see Figure 34.1(iii) on page 455).
CHAPTER 35

INTEREST AND THE RETURN

ON CAPITAL

Why are when buying a house than when buying a

interest rates different
car’ Whydo bonds of different companies pay different rates of interest’
What IS the difference between micresi and profits’ Why do people borrow
to buy instead of saving to buy’ How do firms decide how much to borrow’
Are interest and profits elements in incomes only in a capitalist society or do
they exist in socialist economies too’ What role do they play in the economy’
What gives nse to them ’
In this chapter we are concerned with the payments made for the use of
money Such payments are called inlmst if made to the creditors of a firm,
but they are called dividends or profits if paid to the owners of a firm In dis-

cussing the functional distribution of income, a common distinction is made

between rents, profits, and interest payments, all of which may be, m part
or m whole, payments for the use of someone’s funds
as we shall use the term in this chapter, is a payment for the use
Inttrtsi,

of money An of 5 per cent per annum means that one must pay
interest rate

Ij for pound of money (or one )car The interest rate is the
the use of one
price of usingmone> for one vear, and then returning it One does not buy
money, one hires its services, and the interest rate is the price paid for those
services
If you borrow money from your bank to buy a house, a car, or a washing
machine, you will have to pay interest (or the use of this money You will
also have to arrange to pay the money back, cither in instalments or in a
lump sum The interest that you pay each year is likely to be somewhere
between 4 per cent and 20 per cent of the amount of money that you borrow
If a firmborrows money from the public by selling its bonds, it will have to
pay interest on this loan Most bonds state the rate of interest that will be
paid each year for as long as the loan is outstanding If a firm uses its own
 INTEREST AND THE RETURN ON CAPITAL 471

capital, the economist will impute an interest payment to that capital equal
to the opportunity cost of the funds.

WHY PAY FOR THE USE OF MONEY? THE DEMAND FOR

LOANABLE FUNDS
The demand for loanable funds comes from households, from firms, and
from the central authorities. Since each borrows for different reasons, we
shall consider each separately.

1 Households; People wish to borrow money for consumption and for

investment. They are prepared to borrow it at interest because they prefer
to have a certain quantity of goods now rather than to save now and buy

later. This is sometimes called impatience, but it is more than that. A young
couple that had to save the full price of a house would probably not have
enough money saved until they were ready to retire. Furthermore, many
people have an expectation of a rising level of income over their lifetime and
they are glad to buy now, when they are relatively poor, and pay later,
when they are wealthier. Time preference for present consumption very clearly
exists. Individuals, we obsen'e, spend part of their income to hire now money
they must pay back later. Total outstanding debt of UK households has
been rising steadily over the past decades. Loans on housing are the largest
single item in this stock of debt but hire purchase is also extremely impor-
tant, as also is the debt outstanding on normal monthly accounts held by
households at shops.' In addition, credit cards extend the amount of credit
outstanding. Some individuals borrow money for investment or speculative
purposes. Buying stock and bonds ‘on margin’ involves borrowing money
from a broker of interest) in order to buy securities that
(at a specified rate
the purchaser hopes will more than pay him back for the use of his funds
and the costs of the borrowed funds. It is generally believed (and there is
some evidence support the belief) that households’ demand curve for
to
funds is downward sloping; the amount of money households borrow in-
creases as the rate of interest falls.^
1Are charge accounts free ? Unlike loans from banks or finance companies, no specific
interest charge is made on them. But who pays for the ‘loan’ of funds for a month and for the
costs of record-keeping, collection and bad debts? The seller must either borrow the money to
meet his payments for raw materials, and so forth, tie up his own capital in them, or pass the
credit back to his supplier. Someone is advancing credit and thus incurring costs, and someone
must pay. The obvious choice is the customer, who pays in higher prices.
2 We make such a guarded statement because some of the motives for borrowing
reflect

the prices of cars (or stocks) to rise, one may

expectations about prices of goods. If one expects
rush to buy now even though interest rates are high. We shall assume for the remainder
of this
money rate of interest represent
chapter that the price level is constant, so that variations in the
real variations in the incomes of lenders.
472 DISTRIBUTION

2 Firms The demand for loanable funds comes from

biggest private
firms that borrow because they expect an investment m capital goods to
yield a return over the cost of the capital \Vhen capital is used, production
can be greater than when it is not used, even when allowance is made for
the labour and resources consumed in making the capital This extra return
IS called the productivity of capital
The demand for funds by businesses ts a derived demand, just as the demand for

labour is a denied demand Let us examine it more closely Consider first a

single investment opportunity Let us say a firm can invest £\ 000 in a
machine that will last for 10 years and will produce output that it can sell

for ;^300 per year for each of the 1 0 years The firm estimates the total costs
(for wages raw materials and depreciation), excluding only the oppor-
tunity cost of capital at ;{^180 per year It thus estimates a return on its

investment of £120 per year The productivity of this investment u said to

be 12 per cent per year (£120 is 12 per cent of £1,000 ) The firm would like
to borrow (i e demand) £1 000 of money capital at rates of less than 12
,

per cent m order to purchase this machine

At any one time, however, the firm has many investment opportunities,
some better than others a few may promise rates of return of 40 or 50
per cent, some may promise only 3 or 4 per cent, and some will involve
negative rates of return If the firm knew the whole set of its investment
opportunities, and the rate of return on each, it could prepare a schedule of
the amount of money it would invest at various rates of return Such a
schedule is given m Table 35 I The first two columns show the number
of pounds a firm can invest at each of a senes of different rates of return
The cumulative schedule of these amounts, shown m the third column, is
the marginal productmty of capital (MFC) schedule for this firm For example,
if the firm had £40,000 to invest, its MFC would be 10 per cent the best it

could do (with the investment opportunities listed) would be to invest

£5,000 m projects yielding 30 per cent, £20,000 iti projects yield 20 per cent,
and £15 000 in projects yielding 10 per cent The last or marginal pound
would earn a return of only 10 per cent ‘
These values of the marginal productivity schedule arc plotted on Figure
35 I For a firm with a larger number of investment opportunities, or for an
industry., or a whole economy., wc could n^lace the usdividual dots by a
curve that represents the whole cumulative listing of investment oppor-
tunities, as we have done m Figure 35 2 The curve in Figure 35 2 tells us,
for example, that Oa pounds of investment have an MFC of Or per cent
per year This means that each of the Oa pounds has a productivitj of Or
per cent or more
1 \ ou should chfck your understanding of this passage by verifying that the etrtagt return

on £40 000 would be 17 5 per cent The average return will be above the marginal return
 INTEREST AND THE RETURN ON CAPITAL 473

Table 35.1
HYPOTHETICAL SCHEDULE OF PRODUCIVITY OF
CAPITAL FOR A FIRM
i

Marginal productivity
Rate of Number of of capital
return pounds that {number of poiaids
{in per- can earn at this that can earn this
centages) rate of return rate or more)

30 5,000 5,000
20 20,000 25,000
10 15,000 40,000
5 30,000 70,000
1 150,000 220,000
i

Would it be worthwhile borrowing money in order to purchase a capital

good? would depend, of course, on the productivity of the particular
It
investment and the rate of interest that must be paid on the loan. If the
rate of interest very high, then there will be only a few projects for which
is

it will be profitable to borrow money - those that have a rate of produc-

2 30
o

Fig 35.1 2(D 20

oc

10 h

50 100 150 200

Thousands of pounds

tivity greater than the rate of interest. At a rate of interest of Oz, only Ob
pounds worth of capital would pay a rate of return over their cost. If the

rate of interest productive projects will yield a return over and

falls, less

above the interest charges on money borrowed to institute them. At very

low interest rates, it ivill be profitable to borrow money for all sorts of
projects that yield only a low return over other costs. Thus the MPC
474 DISTRIBUTION

schedule represents the demand for loanable funds by businesses ' Notice
that, although we are talking about the demand for money, its productivity
IS based upon the use of the real capital - the factory or the machine - that
It will purchase

3 Central authorities Governments borrow funds to build highways

and schools and to undertake long-term investment projects Governments
have two mam sources of revenue, taxes and borrowing, and most local
governments borrow to finance their long-term investment projects The
amount of money governments need will depend upon such factors as the
growth in population, defence commitments (vast sums were borrowed
during the Second World War), changes in the demand for roads, educa-

Fxg 35 2

MFC
*
Q J J
Quantity of capital

tion and the many other services provided by central authorities The
government’s choice between borrowing and taxing is, however, often
based on considerations other than the cost of funds As wC shall see in
Chapter 47, governments sometimes borrow m
order to run a deficit (spend-
ing more than they collect) for the purpose of combating depressions Thus
we cannot assume that the government’s demand for loanable funds is

determined primarily by interest rates

Many empirical studies have been made of the demand for funds and
they tend to show that insofar as it affects the demand, the lower the rate of
interest the higher is the demand for money Generally the studies also
suggest that the relation is rather weak, the demand being affected by

many other factors m addition to interest rates and that the demand is

1 The MPC
schedule j$ the demand cutvefbrcapitalonly if we assume that firms are profit
maximisers and that they are certain about the yield from each individual investment If they
are uncertain about the yield, as wdl almost always be the case, they may require a margin
over the expected return m
order to compensau them for the risks involved In this case the
demand curve for capital will be to Iheleftofthe MPC
schedule and in equilibrium the return
on capital will exceed the rate of interest
 INTEREST AND THE RETURN ON CAPITAL 475

rather inelastic, a ceteris-paribus fall in interest rates leading to only a small

change in the demand for loanable funds.

WHO LENDS MONEY? THE SUPPLY OF LOANABLE FUNDS

Money is offered for loan by firms and households that have
funds that
they do not require for current expenditure. If one has surplus funds on
hand, it pays to loan them out, providing that the rate of interest offered
more than covers any possible risk diat the loan might not be repaid. Funds
are provided by the savings both of households and firms. But this is only
part of the stor>'. Banks have a great deal to do with the supply of loanable
funds. Not only are they the channel by which a good part of the savings of
individuals and firms is loaned to those who wish to borrow, but, as we shall
see in Part X, banks actually can create and destroy money. Further, the
central authorities exercise a great deal of control over the supply of loan-
able funds.

'THE' INTEREST RATE; THE PRICE OF LOANABLE FUNDS

We h ave said that the rate of interest is the price paid for borrowing mo ney.
\Ve have also said that there is a downward-sloping demand curv’e of money
to borrow plotted against the rate of interest, and also that there is a supply
of money to borrow. I t would seem a very natural thing to say next that the
prife-of loanable funds is determined by the demand for, and the supply of,
such funds. However, a simple supply-and-demand analysis is not adequate,
for several reasons.
First, the rate of interest is not free to fl uctuate s ufficiently so as always to
bring .about an e quality between the d emand for loanable funds and the
supply. To take one example, banks fix the rates of interest that they charge
on loans with many considerations in mind. They are reluctant to change
these rates every time changes occur in the demand for money to borrow.
If there is an excess demand banks often ration
for loanable funds, then the
the available supply of funds among their customers according to such
criteria as the credit rating of the borrower, how long the banker has known
him, the size of business done, and the persuasiveness of the customer, rather
than by raising the rate of interest. Credit rationing is commonly found
in lending institutions in most Western countries.
Second, evidence suggests that although the demand for loanable funds
is influenced by the rate of interest, it is also influenced by many other
factors, such as the general level of demand in the economy and changes in
this demand. When we draw a demand curve, we assume that factors
affecting demand other than the rate of interest remain constant. In fact.
476 DISTRIBUTION

observed not to be the case the other things are observed to change
this IS

continuously and to exert a major influence on the demand for loanable

funds This means that our demand curve will be continuously shifting
about due to changes m
these other things and unless interest rates adjust
almost instantaneously they will be unable to keep the demand for and the
supply of loanable funds in equality
Third the central authorities affect both sides of the money market The
authorities are large borrowers and they are also the chief influence on the
size of the money supply Central authorities commonly use their influence
*
to affect the rate of interest for purposes of public policy
Whatever are the factors that determine the interest rate, there can be
no doubt that it has an important effect on the share of national income
going to lenders of funds and on the allocation of resources among com-
peting uses

INTEREST RATES AND THE INCOMES OF THE LENDERS

The great bulk of outstanding loans on which interest is paid represents
loans made m the past, whereas loans made within the current year repre-
sent a small proportion of total loans outstanding The interest rate payable
on 1 loan is usually (although not always) fixed at the time when the loan
IS originally made Thus fluctuations in the current price of loanable funds
affect the tncome only of those supphen of funds who are currently making
new loans The total income of all suppliers of loanable funds is arrived at
b> multiplying the volume of outstanding loans made m each past year by
the market rate of interest ruling in the same year, and adding up these
amounts
Fluctuations m interest rates affect the aealth
of all lenders of funds (In
order to appreciate the following discussion you should re read pages
446-49 at this time )
If the interest rate rises from 5 per cent to 6 per cent,
the owner of a i^ljOOO, 5 per cent bond will still receive his ^50 each year,
but the market value of his bond will ^
fall
The incomes of suppliers of funds who have purchased equities rather
than bonds will vary according to the dividend policies of the corporations
they own, and their wealth ivill vary with the pnees of their shares But
these are complicated matters in which man> more factors than the rate
1 Because there is a very large stock of government bonds outstanding the central authorities
can, by choosing to buy or sell such bonds drive their prices up or down (and thus push
interest rates down or up) They can also through regulatory powers affect the behaviour
of commercial banks See Chapter 49
2 Exactly depends upon the redemption date of the bond For a perpetual
how far it will fall

bond the price would fall to /^B33 33 (6 per cent of which is £50) if the bond is redeemable
tomorrow however its pr ce would fell below £l 000 by only a few pennies
 ^

INTEREST AND THE RETURN ON CAPITAL 477

of interest come to bear. There is no simple relation between the welfare of

lenders of funds and the rate of interest.

INTEREST RATES AND THE ALLOCATIVE MECHANISM:

SOCIALIST AND CAPITALIST ECONOMIES
Funds made available voluntarily for investment are scarce at most times.*
These funds represent command over resources that can be used for con-
structing capital goods. There is thus, as with all scarce factors, the problem
of allocating them between a large number of competing uses. A freely
fluctuating rate of interest determined by the demand for, and supply of,
loanable funds, would provide one way of doing this. If capital were scarce,
interest rates would be high, and if it were plentiful, rates would be low.
A high rate would mean that only projects with a high productivity would
obtain funds, whereas a low rate would allow low-productivity projects to
obtain funds. If the rate of interest were set so as to equate the demand for
and supply of funds, this would ensure that the projects yielding the highest
money return would always be the ones undertaken first. If there were not
enough of such projects to exhaust the supply of funds, then the supply
would exceed the demand, the rate would fall, and less remunerative pro-
jects could obtain funds. If, on the other hand, there were, at the current
rate of interest, more projects yielding a positive return over interest
charges than there were funds available, then demand for funds would
exceed supply, the rate of interest would rise, and the less remunerative
projects would be squeezed out. Allocating scarce funds by means of
interest rates thus means that it is always the most remunerative projects
that obtain the funds.
determined by the central authorities (instead of by
If the interest rate is

the free market), precisely thesame process would take place if the central
authorities at the same time controlled the supply of funds in such a way as
to equate demand and supply at that interest rate. Setting rates (with no
control over supply) is similar to controlling prices (discussed in Chapter
1 An artificially low rate of interest will create a problem of rationing and
1).

possibly a black market in funds. An artificially high rate of interest will

find loanable funds accumulating with no borrowers. To control the volume
of loans involves managing both their price and their quantity.
It is, of course, not necessary to letthe market allocate funds. In fact,
banks and other lending institutions do not usually allocate loanable funds
1 At other times, for example, during severe depressions, loanable funds may be in excess

supply.
2 Remunerative in this context is judged according to calculations based on the firm's

private costs and private revenues.

na DISTRIBUTION

m this manner, more often, they chaige a customary interest rate and then
ration the available supply of credit among their customers This means
that projects that cannot yield a return as high as the going interest rate
willnot get money, since no one will want to horrow money to undertake
them Thus there is some allocation according to relative profitability All
projects that yield a return in excess of the going rate of interest, hoi\evcr,
be able to make a profit, and if demand for funds exceeds supply, some
ivill

other allocation principle will have to be adopted by the lenders of funds

If, for example, lending institutions allocate most of their funds to their
oldest customers, and hardly any to new ones, tins will not necessarily mean
that the projects >ielding the highest returns Will get the money We may
call this the alloealion offmdi according to lenders' preferences ‘ Concern is some
times expressed that new and small businesses may he discriminated against
m such allocations
In a system in which the current rate of intcfcst does bear some relation
to the general scaccity or abundance of investment fundsv but where current
demand and supply are equated by credit raiioning, we have a mixture of
allocation by profitability and allocation by leaders’ preferences As loan'
able funds become abundant, the rate of interest will fall and more pro-
jects will be able to compete for the scarce capital, but the ones that are
successful out of those profitable enough to be able to compete will be
determined by lenders’ preferences This mixed system is the one found in
most Western countries
Another means of allocating scarce capital » by government edict The
central government could establish its own set of pnontics according as it

judged various investments to further the national interest, and it could then
allocate the scarce funds first to the top-pnonty project, then to the second
priority one, and so on down the list, until the funds were exhausted
A mixed system would be one in which the government allocated some of
the available capital to projects that be in the social interest,
it judged to

independent of their financial profitability, and let the remaining funds go

to the highest bidder and thus to the financially most rcmunerativ c projects
In centrally-planned economies, it is impossible for one central planning
body to make all the decisions about the allocation of investment funds
Usually, the central planners make decisions on broad matters of policy, and
much of the detailed - but often quite important - decision-taking is de
centralised Decision-takers require some guide as to what investments are
worth making and what investments are too unremunerative, given the
general supply of capital in the whole country, this guide often takes the
form of some sort of a rate of interest (often called by another name) If the
decentralised decision takers ate told that their investments must yield a
1 This important problem has already been discussed on psfcs 128-9
 INTEREST AND THE RETURN ON CAPITAL 479

return over and above an interest charge on the capital used, then this en-
more remunerative investments will take precedence over the
sures that the
less remunerative ones. If the state ensures that prices of goods and factors
of production reflect its own valuation of their relative importance, then

this investment procedure ensures that the most important projects will be
undertaken first. The difference between such a system and the Western
market system is, thus, not in the existence of an interest rate, for one exists
in both kinds of societies, but in the absence of a class of people who receive
the interest payment in return for making available their privately owned
capital, and in the way in which different projects are valued (according
to market prices or according to state-administered prices).

THE STRUCTURE OF INTEREST RATES

We have spoken heretofore of the interest rate, but it is readily apparent
that at any one moment of time there are many different interest rates.
Today, although you may receive an interest rate of about 4-5 per cent
on deposits at a savings account you will have to pay about 6 per cent to
borrow from the same bank. Interest rates on consumer credit of 12 per cent
and 16 per cent are common. A small firm pays a higher interest rate than
a giant company on funds it borrows from banks. Different government
bonds pay different rates of interest, depending upon the length of the
period for which the bond runs. Company bonds tend to pay interest at a
higher rate than government bonds, and there is much variation among
bonds of different companies. Since money is extremely mobile, why do
such great differences exist? Why doesn’t the flow of funds between different
uses diminish these differences ?
The answer is that the quoted interest rate is a payment of more than a
‘rental fee’ for the use of money. It must also covct die risk of default (i.e.,

the risk that the loan mi ght not he paid back^ aTlV.'^e costs involved m
processing apd collecting the loan, and the lenders loss of liquidity whichjs
a fun ction of both the period of time for which the loan is granted and I he
possi bilitv of persuading someon£-e]se_to take over the loan before its period
is up (by selling the evidence o Ld£bOPJhe_i>JIifirj3£rson).^Differen^loans
differ with-respect to each .o f the factors and th us they^Imve different
interest rates attached to them. I

PROFITS
Capital in any one use must receive a return sufficient to prevent it from
transferring to other uses. This necessary return on capital is a transfer
480 DISTRIBUTION

earning (see Chapter 33) and it is cillcd the interest on capital by the
*
economist
Profits as we ha\e used
the term are earnings abo\c the opportunity cost
of capital In a completely changeless world in which there w ere no barriers
to entry m any industry, profits would tend to zero so that in all industnes

the return to capital would be the same {making allowances for dilTcrcnces
in risks)
The existence of profits m the economists sense of the word is associated
with the dynamic factors of change and uncertainty Change comes from
both the side of demand and of supply \Vc hat c seen tn our study of the
theory of the firm how shifts m demand give rise to profits and to losses A
rise m demand will make an industry profitable, the profits will attract
capital into the industry and the expansion will continue until the profits
are competed away A fall in demand will make an industry unprofitable
and the losses will cause the industry to contract until opportunity costs can
again be covered fully Cliangcs in costs gi\c nse to profits and losses from
the supply side If a firm introduces new cost snving methods of production
or if it produces a new product that is prcfeiTed by consumers to existing
products Its profits will rise Uc have called such changes innova>
tions Successful innovations are undoubtedly a cause of profits and, as
emphasised by the writings of the late Professor Schumpeter, the possi
bility of earning such profits is a powerful incentive, for the firm to
innovate
^Ve have seen throughout the theory of the firm how variations m profits
signal that shifts in demand or in costs have taken place and so act as an
incentive or disincentiv e to investment in a particular industry Profits thus
fulfil a function in the free-market system of resource allocation Relative
profitabilities signal the relative attractiveness of investment in vanous lines

Highly profitable investment opportunities give rise to the demand for

investment funds The interest rale then, insofar as it does reflect demand
and supply for funds, serves to choke off the demand for funds at the avail
able supply, leaving those projects that yield a ‘profit’ (m the businessman’s
sense) higher than the rate of interest to be undertaken, and those that
yield a ‘profit’ less than this rale to go undone
VI. VV/CKA toA vaccssA vw w/d-atw/j tout. cS w.'.t'J.TOCTS.
no allocative function The classical opposition to monopoly and
that serve
to monopoly profits rested in pail upon the fact that the existence of
monopoly profits could not induce the resource flows that would eliminate
them

\ We saw in Chaplei 17 that ibe business man also calls ihis an mierest payment if be has
borrowed ihe capital but he calls il profit if it is a return on his own capital
 INTEREST AND THE RETURN ON CAPITAL 481

The Share of Profits in the National Income

We conclude by raising the interesting question of the share of profits in the
national income. We have no satisfactory theory of the share of national
income going as profits, and we can do little to explain past behaviour of
this share, nor do we have a body of predictions about the effect on this

share of occurrences such as the rise of unions, wage freezes, profits taxes,
price controls, etc. There is a great deal of basic research that needs to be
done by students of this subject. We do have a considerable body of syste-
matic observ'ations of how the share of profits in the national income varies
from one economy to another, and within economies from one time period
to another. We do not, however, have a theory capable of explaining these
phenomena, that has stood up to any serious amount of testing.

16
CHAPTER 36

CRITICISMS AND TESTS

OF THE THEORY OF
DISTRIBUTION

In previous chapters we have developed the traditional theory of distribu-

number of different contexts It is probably valuable to repeat vfhat
tion in a
18 basically the same analysis m a number of different guises Repetition
helps in de\ eloping the‘feel’ for the workings of a price system that is

thought to be so important to the economist Such repetition has the dis-

advantage, however, that it makes the theory appear to have much more
content than it actually has In fact, the whole of distribution theory is based
on only two or three basic hypotheses about behaviour In this chapter we
attempt first to lay out the basic structure of the theory and we then go on
to consider various criticisms and tests that have been put forward from
time to time

THE THEORY RESTATED

The theory of distribution asserts that factor pricing can be explained by

demand and supply
Factor supply is determined, according to the hypothesis of net advantage
When, taking pecuniary and nonpecuniary rewards into account, differen-
tials exist m the wages offered to factors, factors will move among uses,

among industries, and among places in such a way as to equalise the net
advantages to the owners of factors Because there are impediments to the
mobility of factors, there may be lags m
the response of factors to changes in
relative prices Thus the elasticity of supply will depend upon what factor
IS being discussed and what time horizon is being considered
The demand for a factor is a derived demand, depending on the marginal
 CRITICISMS AND TESTS OF THE THEORY OF DISTRIBUTION 483

physical productivity of a factor and the demand conditions for the com-
modity made by the factor. The addition to revenue caused by employing
another unit of some variable factor depends on the addition to physical
output contributed by that factor and the change in total revenue when the
rate of sales is increased by this amount. The curve that displays this infor-
mation is calle4 the marginal revenue product curve of the factor. As long
as the. factor is purchased under competitive conditions, and as long as firms
are maximising short-run profits, the marginal revenue product cuiY^e of a
factor is the firm’s demand curve for that factor. This is true whether the
product is sold under conditions of perfect competition, imperfect compe-

tition, or monopoly. In equilibrium, each type of factor will be paid a wage

equal to marginal revenue product.
its

A is not equating the marginal revenue products of each of its

firm that
factors with that factor’s price is not maximising its profits. On the other
hand, if the firm is maximising its profits, then it is necessarily equating
each factor’s price to the corresponding marginal revenue product. The
theory thus stands or falls with the theory of profit maximisation. It is
merely an implication of profit maximisation, and the only reason for
spelling it out in detail is that this may help us to develop interesting and
useful hypotheses about the effects of various changes in the economy on
the markets for factors of production.
When one thinks of all the heated arguments over the theory, of all the
passionate denunciations and defences that it has occasioned, it is surprising
to observe how few predictions it makes and how uncontroversial most of
them are. The theory predicts that demand for factors depends on, and
varies with, the demand for the products made by the factor. This was un-
doubtedly a great discovery when it was first put foiAvard no%v, however, it
;

is almost a platitude. The theory also asserts that the technical conditions
of production will influence the demand for a factor. The theory predicts
that, assuming the supply curve of the factor has not shifted, changes in the
factor price must reflect changes in the demand for the commodities made
by the factor. On the supply side, the theorj' predicts that movement of
resources between firms and between industries wll occur in response to
changes in factor prices. It is very hard to quarrel with any of these pre-
dictions; in fact, they seem so obvious as to be trite. They are, nevertheless,
important and often arise in practical issues of policy.
We have seen that marginal productivity theory explains the demand for
factors of production; it constitutes half of the traditional theory of dis-
tribution. The other half is the theory of supply, which asserts, as we have
seen, that factors will move between occupations in search of the highest
net advantage. It the marginal-productivity half of the theory that has
is

been subject to most criticism and about which there exist so many
484 PISTRIBUTION

misconcepiions, and we shall here consider a representative selection of

‘
these
1

SIX COMMON MISCONCEPTIONS ABOUT MARGINAL

PRODUCTIVITY THEORY
The theory assumes perfect (ompetthon tn all markets This is simply not correct
The relationship between the marginal physical product and the marginal
revenue product vnll be altered if the degree of competition alters, but the
marginal revenue product curve is the demand curve for the factor in per-
fectcompetition, imperfect competition, and monopoly Competition is
assumed in the factor markets, not in the markets svherc commodities are
bought and sold
2 The theory assumes full employment Marginal productivity by itself is a
theory of the demand for factors of production, it says nothing about the
supply of factors Unemployment ts a relation between the demand for, and
the supply of, factors of production Clearly, therefore, the marginal
productivity theory predicts nothing, one way or the other, about un-
employment
3 The theory assumes that the amount end value of the marginal product of a

factor « known enUepreneur The theory assumes no such thing* Critics

to the

argue that the 6rm will not pay any factor the value of its marginal product,
because the hrm will generally have no idea what that marginal product is
and would be unable to calculate this magnitude even if it tried This
cnticism IS irrelevant It has already been pointed out that payment

according to marginal revenue product occurs automatically whenever the

firm IS maximising its profits It does not matter how the firm succeeds m

maximising profits - by guess, hunch, luck, clairvoyance, skill, good judg-

ment, or calculating marginal quantities As long as profits are maximised,
factors will be getting the value of their marginal products The theory does
not purport to describe how businessmen calculate, it merely predicts how
they will react to various situations on the assumption that they are maxi-
mising profits
4 Since the theory predicts that all factors will receive a wage equal to their

marginal revenue product, the theory denies the possibility of exploitation offactors hy
their employers Thea theory about competitive factor markets all
theory is

employers are assumed to be wage-takers The> can choose how much of a

factor to employ, but no one of them can influence the wage If labour were
free, it would pay an employer to take on labour until its marginal revenue

product had fallen to zero If labour costs lOr per hour, it will pay the
J TTie first five are taken from F C Benham £ro«o»rttcj A Gineral [atrodaclton (6th ed ,

London Pitman 1960)

 CRITICISMS AND TESTS OF THE THEORY OF DISTRIBUTION 485

employer to add labourers until the marginal revenue product is I Or per

hour. If the employer stops short of this point - say, by using labour in a
quantity such that the marginal revenue product is 12r Grf - then the em-
ployer is paying workers less than the value of their marginal product.
Whether or not this constitutes ‘exploitation’ of the workers is debatable;
there no doubt, however, that it constitutes exploitation of the employer
is

by the employer! As long as the wage is less than the value of the marginal
product, the employer can increase his own profits by hiring more labour.
The theory assumes profit maximisation; it then deduces that the wage will
be equal to the value of the marginal product. Labour can be paid less than
its marginal revenue product in the case of monopsony. Indeed, this is a

prediction of the theory of monopsony ; it is not a refutation of the theory of

competitive factor markets.
5 The theory assumes that the price per unit of a factor will be the same in every
industry in which it is employed. This is incorrect. The theory does not assume
it. As we have seen, the theory predicts that factors willmove between in-
dustries until net advantage is There is nothing in the theory' that
equalised.
prevents it from accommodating unequal payments in various industries
due to such factors as a lack of mobility of labour between industries, non-
pecuniary considerations, or the existence of disequilibrium in a market in
which the adjustment mechanism works only slowly.
6 The theory is inhuman because it treats human labour in the same way as it
treats a ton of coal or a wagonload offertilizer. One must be careful to distinguish

one’s emotional reaction to a procedure that treats human and nonhuman

factors alike from one’s evaluation of it in terms of positive economics. Any-
one who accepts this criticism must explain carefully why separate theories
of the pricing of human and nonhuman factors are needed. He must also
show that his ‘human’ theory makes predictions that differ from those
made by the marginal productivity theory. The marginal productivity
theory only a theory of the demand for a factor. It predicts only what em-
is

ployers would like to buy. It predicts that employers’ desired purchases of

labour (and all other factors) depend on the wage rate, the technical con-
ditions of production, and the demand for the product made by labour.
Supply conditions may differ between human and nonhuman factors, but
these differences are accommodated within the theory. No evidence has
yet been gathered to indicate that it is necessary to have separate theories
of the demand for human and nonhuman factors of production.

Marginal Productivity and Justice in Factor Rewards

We have seen that it is a condition of equilibrium in the marginal produc-

tivity theory applied to a world of perfectly competitive factor markets that
 ,

486 DISTRIBUTION

all factors receive a payment equal to the values of their marginal products
Some supporters of the theory of marginal productivity have held that not
only was the theory correct, but that it satisfied the canons of justice,’ i c
that It gave rise to a just distribution of the national product, because
facton were rewarded according to the value of their contnbutions to the
national product Many cntics of the low levels of wages that prevailed m
the nineteenth century reacted with passion against a theory that was
claimed to justify these rates of pay
It IS beyond the scope of a book on positive economics to enter into
normative questions of what constitutes a just distnbution of income It is,

however, worth getting the facts straight According to the marginal pro
ductivity theory, each labourer (or each unit of any other factor) does not
receive the value of what he personally contributes to production He
receives, instead, the valueof what one more labourer would add to produc
were held constant If 1 million similar labourers arc
tion if all other factors
employed, then each of the I million receives as income an amount equal
to the extra product that would have been contributed by the millionth
labourer if he had been hired while capital and all other factors had re-
mained unchanged Whether or not such a distribution of the national
product IS regarded as just, one cannot say that each unit of a factor receives
as income the value of its mtm contribution to production Indeed, where
many factors cooperate m production, it is generally impossible to divide
total production into the amounts contributed by each factor of production

DO MARKET CONDITIONS DETERMINE FACTOR

EARNINGS?
Factors other than Labour

Most nonhuman factors are sold on competitive markets The theory pre-
dicts thatchanges in the earnings of these factors will be associated with
changes in market conditions The overwhelming preponderance of evi-
dence supports this prediction of the theory Consider some examples

1 Raw MATERIALS A dramatic example was provided during the Korean

War when a rapid increase in the demand for many strategic matenals sent
their prices soaring to the extent that the incomes earned by their owners
soared as well The pnees of copper, tin, rubber, and hundreds of other
matenals fluctuate daily in response to changes m
the demand and supply
of these products

I One of the most famous exponents of ihu v>ew the American economist John Bates
Clark
 CRITICISMS AND TESTS OF THE THEORY OF DISTRIBUTION 487

2 Land values: Land in the heart of growing cities is clearly fixed in

supply, and values rise steadily in response to increasing demand for it. The
value of the land itself often makes it worthwhile to destroy durable buildings
to convert land to more productive uses. The New York skyscraper and the
new London skyline are monuments to the high value of urban land. In
many smaller cities, the change in tastes from shopping in town to shopping
in outlying shopping centres has lessened the demand for land downtown
and influenced relative land prices. The increase in the price of land on the
periphery of every growing city is a visible example of the workings of the
market.
Agricultural land appears at first glance to provide counterevidence. The

classical economists predicted 150 years ago that, as population and the
demand for agricultural products grew, the price of the fixed supply of land
would rise enormously. The price of agricultural land, however, has not sky-
rocketed. Although the demand for agricultural produce did expand in the
predicted fashion because of the rise in population, the productivity of
agricultural land has increased in quite unexpected ways due to the inven-
tion of the vast range of machines and techniques that now characterise
modern agriculture. The prediction was falsified, not because the price of
agricultural land is not determined by market forces, but rather because
some of the market forces were incorrectly foreseen.

3 Taxicabs : The system of regulating taxicabs in London is ver)’ different

from the one in use in New York. The theory successfully predicts the
consequences of these different regulatory rules in each city.

The supply of New York taxicabs is rigidly controlled by a licensing

system, and number of cabs is kept well below what it would be in
the a
free-market situation.' The medallion, which confers the right to operate a
cab, acquires a scarcity value (presently about $26,000) ;
its market price is

predicted by the theory of rent, one aspect of which is the pricing of factors
As the demand for services of taxicabs rises due
in perfectly inelastic supply.
to increases in populationand average incomes, the price of medallions rises
correspondingly, so that new entrants earn only normal profits. If fares are
increased and the demand proves inelastic, so that gross income from operat-
ing a cab the price of the medallion rises correspondingly. The fare
rises,

increase thus amounts to a free gift to the current holders of medallions; it

does nothing to raise the net incomes of cab operators newly entering the
industry.
In London, fares are rigidly regulated but entr)' is quite free. Profits are
reduced to zero by entry which continues until each cab is only carrying
sufficient fares during the day to earn normal profits. Periodically the fares
1 This makes sense if cabs have social costs which are not borne by their private operators.
488 DISTRIBUTION

are raised in an efTorl to raise incomes If demand is inelastic incomes nse

in the short run but this attracts more cabs into the industry and excess
capacity increases until profits again fall to zero This case is m fact
analytically identical to that of the barber discussed on pages 368-70

Labour
When we try to apply our theory to labour we encounter two important sets
of complications first, labour markets arc a mixture of competitive and

noncompetitive elements, the proportions of the mixture diflering from

market to market, and, second, labour being the human factor of produc-
tion, nonmonetary considerations loom large m
its incentive patterns These

complications make labour economics one of the most diiTicult fields of all
economics They also make the question posed in the heading of this section
hard to answer Monopolistic elements and nonmonetary rewards, both
of which arc difficult to measure, require careful specification if the theory
tint labour earnings respond to market prices is to be made testable
Nevertheless, wc do hai e a mass of evidence to go on We do has c cases in
which a strong union one able to bargain effectnely and to restrict entry
of labour into the field - has caused wages to nse well above the competitive
level The West Coast longshoremen in the US could never have hoped to

obtain their present privileged position were it not for the extremely ciTec*

tive operations of Harry Bridges, president of the union When Bridges was
mobbed in San Francisco m Spnng, 1964, by unemployed labouren wanting
jobs as longshoremen, we were given impressive evidence that, if entry
could not be restneted, the high earnings of longshoremen could not be long
maintained Many other similar cases have been documented Unions can
and do succeed in raising wages and incomes when they operate in small
sections of thewhole economy, the high earnings do attract others to enter
the occupation or industry, and the pnvileged position can be maintained
only if entry can be cfTcctivcIy rcsincled
Earnings do then respond, at least to some extent, to monopoly power
Do they respond to normal fluctuations of demand and supply’ Here the
evidence ismixed The competitive theory predicts that a decline m the
demand for some product will cause a decline m the derived demand for the
factors that make the product, a decline in their income, and the exit of
factors to other uses Cases come easily to mind With the advent of the
motor car, many skilled carnage makers found the demand for their services
declining rapidly Earnings fell, and many in the older age brackets found
that they had been earning substantial rents for their scarce, but highly
specific, skill These men were forced to suffer large income cuts when they
moved to other industnes Many silent-screen stars who found their voices
unsuitable for the talkies suflfered disastrous cuts m mcotne and fell into
 CRITICISMS AND TESTS OF THE THEORY OF DISTRIBUTION 489

oblivion when the demand for silent films disappeared. A similar but less
dramatic fate hit many radio personalities who were unable to make the
transition to television and had compete in the greatly reduced market for
to
radio talent. Much earlier, the same fate met those music-hall stars whose
talents did not project on to. the flat, flickering screen of the early silent
movies. How soon will television entertainers, who have
enormous incomes
due to the high demand for their services, go the same way when a yet
newer entertainment medium sweeps away the present one? When in a
competitive, changing society you hear the bell toll for some once wealthy
and powerful group you should always remember that someday the bell
could easily be tolling for you
University lecturers in economics are an example of a group currently
enjoying gains as a result of changes in labour markets. The relative earn-
ings of an academic economist today are much higher than they were twenty
years ago. This is particularly so at the starting end of the scale where in-
tense competition for the scarce supply of good students who have just
obtained the relevant degrees forces up their price to levels that would have
seemed princely only fifteen years ago. Furthermore, the relative earnings
of different kinds of professors reflect the relative strengths of the demands
for their services innonacademic fields. Engineers, chemists, and economists
are all in heavy demand by nonacademic organisations, and their incomes
relative to those of their colleagues in less favoured fields reflect their com-
paratively high transfer earnings. Nor is this an isolated example. Why, if
you have the talent, can you make a lot of money writing copy for a London
advertising agency, whereas, if you have the talent, you wll not make a lot
of money writing books of poetry? Not because any economic dictator or
group of philosophers has decided that advertising is more valuable than
poetry, but because in the British economy there is a large demand for

advertising and only a tiny demand for poetry. A full citing of all such
evidence would cover many pages, and it would all point to the conclusion
that earnings of factors do very often respond to changes in market forces.
On the other hand, not only can monopoly elements raise incomes above
their competitive levels, but they can also prevent incomes from falling and
reflecting decreases in demand. Of course, if the demand disappears more
or less overnight (as itdid in the case of silent-movie stars and carriage
makers), there is nothing any union can do to maintain incomes. But the
story be different if, as is more usually the case, demand shrinks
may
steadily over a few decades. In this case powerful unions can often hold
wages up in the face of declining demand. Cases in which this has occurred
are found in railways and coal mining in the US.^
1 I cannot suppress the conjecture that the ‘restrictive behaviour’ of unions in these cases
has led to a more orderly, humane and civilised phasing out than would have occurred had

16 *
490 DISTRIBUTIOK

By way of summary, seems clear that the competitive theory docs help
ii

to explain the relative earnings of different groups of labour, clearly, how

ever, a strong dose of monopoly theory must be added if we hope to explain
much of what we see
There is a debate as to how much of the behaviour of labour markets can
be explained by economics Some argue that one must use one’s knowledge
of political science, sociology, or even psychiatry to understand the labour
leaders This debate cannot easily be resolved To do so would require
clearly specified predictions of alternative theories and careful testing of the
differences between these predictions The present state of the evidence
seems to be that economic explanations and wonomic forces do determine,
within limits the behaviour of factors, so that there is no need to abandon
such explanations and seek totally different ones Within these limits, there
IS room for much additional explanation, and economists should welcome

theories from wherever they can gel them Debates about conflicting
theories, each of which is logically consistent, can be resolved only by a
proper attention to the facts The view expressed in tins paragraph rests

upon the present state of the evidence and is subject to refutation

DO FACTORS MOVE IN RESPONSE TO CHANGES IN

EARNINGS?
In the previous section we saw that earnings do tend to change in response
to demand and supply conditions Changes in earnings are signals whose
purpose is to attract resources into lines of production in which they arc
more needed and out of lines in which they are less needed
In the case of nonhuman strong evidence that the theory
factors, there is

is able to predict the actual course of events with reasonable accuracy

Land is transferred from one crop to another response to changes the m m

relative profitabilities of the crops Land on the edge of town is transferred
from rural to urban uses as soon as it can earn substantially more as a build-
ing site than as a corn field Materials and capital goods move from use to
use in response to changes in earnings in these uses Little more needs to

be said here, the most casual observation will show the allocative system
working pretty much as described by the theory
Again, the complications come with labour Countless studies of labour
mobility have been made, but they do not point to a simple answer to the
question of whether the factor labour moves m response to monetary incen
the adjustment been left to a free market in which case those who remained tn the industry
and who were needed by >C would all have suRcred depressed conditions in order that the
disincentive could operate on those who did leave and on those who might otherwise have
entered
 :

CRITICISMS AND TESTS OF THE THEORY OF DISTRIBUTION 491

tives. On the one hand, it is clear that the great migration of Americans to
theWest Coast during the Second World War was induced by expanding
employment opportunities and soaring wages in the shipyards and aircraft
factories of California. On the other hand, why were the depressed areas of
Wales not depopulated ten years ago when the coal mines began to shut
down ?
At the risk of grossly simplifying a complex situation, we hazard the
statement that the existing evidence is consistent with the following
hypotheses

1 There exists a fairly mobile component in any group. This mobile

component tends to consist of the youngest, the most adaptable, and often
the most intelligent members of the group.
2 This mobile group can be attracted from one area, occupation, or
industry to another by relatively small changes in economic incentives.
3 Providing the pattern of demand for resources does not shift too fast,

most of the necessary reallocation can be accomplished by movements of

*
this mobile group.

As we go beyond these very mobile persons, we get into ranges of lower

and lower mobility until, at the very bottom, we find persons who are
virtually completely immobile. The most immobile are the very old, those
with capital sunk in nonmarketable assets, the timid, the weak, and those
who receive high nonmonetary rewards in their present occupation or loca-
tions. For them to shift is difficult; in extreme cases, only the threat of
starvation will motivate them.^ Thus, it may be relatively easy to create a
substantial inflow of workers into an expanding industry, or some outflow
of workers from a depressed industry, occupation, or area by a relatively
small shift in earnings. Such outflows from depressed areas such as
Appalachia and parts of New England in the US, the Maritime Provinces
of Canada, Sicily and southern Italy, the Highlands of Scotland, declining
areas of northeast England, and rural parts of central France have been
observed over long periods of time.
Although it is relatively easy to get some outmigration, it is difficult to get

large transfers in a short period of time. When demand falls rapidly,

pockets of poverty tend to develop. In each of the geographic areas men-
tioned above, labour has been leaving, but poverty has increased. The
reason is that the rate of exit has been slower than the rate of decline of the
economic opportunities in the area. Indeed, the exit itself causes further

decline, for, when a family migrates, both the supply of labour and the

1 Of course, the same individual need not move over and over again. The group is con-

stantly replaced by new entrants into the labour force.

2 Even this may not be enough. Some people believe, rightly or wrongly, that they will

starve even if they move.

492 PISTRIBUTION

demand for labour decline This is because all the locally provided goods
and services that the family consumed before they migrated now suffer a
reduction in demand
One of the main functions of inequalities in earnings is to signal labour to
reallocate It is sometimes asserted that if this signal is remoted, govern
ment compulsion would be the only allocative device available This
assertion is not correct
An alternative signal, still tai^ely within the free-market mechanism, is

to keep the earnings of the employed from varying, but to let unemploy
ment rates vary This is roughly what has happened the coal mining m
industry The mechanism we have been studying for reallocating labour
from industry A to industry 8 m the face of a shift in demand is a nsc in
earnings in B and a fall m A But what if wage rates m A and B are fixed
by powerful unions or by government decree^ Unemployment will develop
m A and severe shortages will devrelop in B Even though there is no dif
ference in the earnings of labour in A and B, the chance of obtaining a job
is mj^Ahighcr m B than m A This may well induce new entrants into the
labou?*torce to train for B rather than A, and it may induce some un
successful applicants for jobs in A to transfer to fi

Thus It possible to imagine allocative signals other than earnings

1$

differentials and government compulsion The question is now an empirical

one to what extent does labour respond to earnings differentials and to
what extent does it respond to other signals^ One recent study suggests that
the regional movement of labour responds more to relative regional un
employment rates than it docs to relative earnings *

On hand the unemployment in Appalachia, North cast

the other
England and South Italy seems to be fully as inefTeCtive as wage diffcren
tials in inducing sufficient movement, and there is some evidence that farm

unemployment docs not drive people off the farms as rapidly as the decline
*
in the demand for farm labour requires
Recently, nonmarket orientated policies have increased in populanty
with the central authorities of several countries, apparently in the belief that
no set of changes m market signals will suffice to secure the necessary move-
ment of labour withm an acceptable time period The US has given re
training a prominent role in its distressed area programmes The Area
Redevelopment Act of 1961, the Manpower Development and Training
Act (1962), and the War on Poverty Program of 1964 all include retraining

1 See a forthcoming article by B A Ci>ny in Emumua

2 A song popular soon after the end of the Fint World War asked the question How re
you going to keep them down on the farm after they ve seen Paree * This suggests that m
formation about and direct experience of alternatives may be more influential than the mere
existence of these alternatives
 CRITICISMS AND TESTS OF THE THEORY OF DISTRIBUTION 493

schemes.' The first report of the National Economic Development Council

in Britain and the subsequent ‘National Plan’ both laid substantial emphasis
on policies to aid the relocation and retraining of labour.

MARGINAL PRODUCTIVITY THEORY AND THE

FUNCTIONAL DISTRIBUTION OF INCOME
We have referred several times to the problem of the distribution of national
income into such broad aggregates as wages, rent, interest and profits. The
problem of distribution into these broad aggregates is sometimes called the
problem of macro-distribution. Figure 34.3 on page 468 shows the percent-
age of the national income going to wages and salaries over the last hundred
years in Britain. It has sometimes been argued that these data provided a
refutation of marginal productivity theory. The argument was seldom made
explicit but it ran somewhat as follows:

1 Over the last hundred years there has been an accumulation of

capital, theamount of capital per worker having increased greatly.^^^
2 According to the ‘law’ of diminishing returns the marginal product of
- capital must have declined relative to that of labour.
3 Therefore there should have been an increase in the earnings of
labour relative to capital and, hence, an increase in the share of national
income going to labour.
4 Such an increase has not occurred, therefore the marginal produc-
tivity theory is refuted.
Such an attempt to test our theories against empirical data is to be
lauded; it is a pity that the argument contains such an elementary theoreti-
cal mistake at step 3. Can you spot the mistake for yourself?
Step 3 confuses a change in the price of a factor with a change in total
earnings. It is well known, for example, that a rise in supply lowers price
but raises, lowers or leaves unchanged total expenditure according as the
elasticity of demand is elastic, inelastic or unitar>^ In 3 above the correct
inference is that the payment per unit of capital employed will fall but, since

there is more capital employed, total earnings of capital may rise or fall

depending on the elasticity of demand for capital. In fact all that the

marginal productivity theory predicts^ is that the payment per unit of

1 For an evaluation of the programmes, see G. Somers and E. Stromsdorfer, ‘A
Benefit-Cost

Analysis of Manpower Retraining', Ajutricau Ecotiottiic Review, May 1965. The authors conclude
that the benefits substantially exceed the costs.
2 Even this prediction can be extracted only by assuming no innovations. If there have been
changes in our knowledge about how to use capital then it no longer follotvs that the accumula-
tion of capital will be accompanied by a lowering of its marginal product.
Furthermore, in

order to test it we would have to know what we meant by a unit of capital which is, to say the

least, a very ambiguous concept.

491 DISTRIBUTION

capital wll fall relative to the payment per unit of labour Whether the total
earnings of capital will nse or fall as a percentage of the national income
depends on the elasticity of demand for capital
The marginal productivity theory is thus not refuted by any observed
time path m the distribution of income (given knowledge about relative
factor supplies) Or, to sa> the same tlung in a slightly different way, the
marginal productivity theory makes no prediction about the distribution
of income, unless we know the elasticities of demand for the various factors
of production IVith our present state of knowledge, the marginal produc-
tivity theory provides no predictions about the effect on macro-distribution

of such changes as shifts m total factor supplies, taxes on one factor, the nse
of unions etc We must, at the moment, admit defeat, we must admit that
we cannot deal at all with this important class of problems ‘
A few attempts to provide new theories of macro-distribution have been
made, the two most notable being those of Kalecki and N KaldorM
Kalecki’s theory attempts to explain distribution between wages and profits
m tefffis of the degree of monopoly The theory is open to many criticisms but

It none the less constitutes a significant and path-breaking attempt to

develop a theory specihcalty designed to explain and predict macro

distribution phenomena
Mr Kaldor has been developing this theory o\ er a number of years It js

probably fair to say ihat it is not yet fully worked out and that it has
certainly not been subjected to anything like a critical test ^ The theory
contains a number of novel elements and
it has been placed into the context

of a theory ofeconomic growth It is basically an attempt to explain macro

distribution between wages and profits in terms of a marro-thcory of national
income instead of in terms of a mtero theory of relative pnccs

1 leu a hazardous venture to try to guess the Tuturecoune of the development oDcnowledge
Personally however I would hazard the tentative guess that useful developments of macro

theories of distribuCion will notcome through the marginal productivity theory I say this
because it seenu to me that the informaiKm needed to make this theory predict anything at all

about macro-disiribution in a real world context is too detailed and hard lo come by, to make
development along these lines a hkcly bet I would be more inclined to believe that a break
through will come along totally diflercni hues
2 Kaldor s theory has a long inteUectual htstory traces of it being found m
the wntings of
Mrs Robinson Professor Kahn and Keynes among otheis For a summary of various theones
of distnbution together with some empirical tests of each (including a test of Kaldor s theory)
the potential economics specialist should consult M Reder Alternative Theories ofLabour s

Share m The Allocation of Economic Resources Essay in honour of B Haley Ed by

M Abramovitz and others (Stanford Univemty Press 1959}
 PART 6

THE ECONOMY
AS A WHOLE
CHAPTER 37

THE INTERACTION AMONG

MARKETS: GENERAL AND
PARTIAL ANALYSIS

The economy of any country consists of thousands upon thousands of inter-

related markets. There are markets for agricultural goods, for
manufactured
goods and for all types of consumers’ services; there are markets for semi-
manufactured goods such as steel and pig iron which are outputs of some
industries and inputs of others; there are markets for raw materials such as
iron ore, trees, bauxite and copper; there are markets for land, and for
thousands of different types of labour; there are markets for the lending of
new capital and for the transfer of existing loans. In the past 400 pages we
have studied many of these markets more or less in isolation (although in
the case of the markets for factors of production we have been concerned to
stress the relationbetween the price of consumer’s goods and the demands
for factors of production through the idea of the derived demand for a factor).
The economy should not, however, be viewed as a series of markets
functioning in isolation. It should be viewed instead as an interlocking
system in which anything happening in one market will greatly affect some
other markets, and could conceivably affect every other market in the
economy.
The markets of the economy are linked together and what goes on in each
is coordinated (more or less well) by the price system. We have already dis-

cussed the coordinating function of the price system in Chapters 5 and 13,
and the student should review both of these chapters at this time. The price
system does serve to coordinate what goes on in isolated markets. Changes
in surpluses and scarcities are reflected in price changes; these price changes
signal to decision takers what is happening in other markets and they alter
their behaviour in response to these changes in signals.
498 THE MARKET ECONOMY AS A WHOLE

AN EXAMPLE
The demand for cars has been nsing fairly rapidly in most Western coun
tries A particular nse in demand ivill be met fairly soon with a nsc in output
using existing plant and equipment but working it harder by means of over-
time and other expedients If the nse in demand is considerable, and judged
to be permanent, there will also be a planned increase in capacit) in the
car industry Employment will rise and an attempt may be made to attract
labour from elsewhere by offering higher earnings Thus, one of the first

impacts on other industries will be a loss of labour and possibly a need to

raise wages in order to compete with the car industry for labour This may
cause profits to fall m these other industries The increased employment m
the car industry may occasion some geographical movement of labour In
this case there would be a rise in the demand for housing in the car centres

and a corresponding fall in demand elsewhere New housing construction

m the car-producing areas would lead to a nse tn the demand for construc-
tion workers and materials Quarnes and brickworks will have to take on
additional labour and expand output There will also be a rise in the de-
mand for banking services, cinemas haircuts and the thousands of other
things that households moving to the car-producing areas will want to
consume Furthermore, there will be a rise in the demand for raw materials
used in car construction and the effects of this may be felt in such diverse
places as the glass making areas of the Midlands, the steel manufacturing
sections of Wales and the rubber plantations of Malaya If new investment
in plant and equipment takes place m the car industries, there will be a rue
in the demand for many capital goods, shortages and bottlenecks may

develop and other industries which use these materials may experience in-
creases in their costs and troubles with delivery dates There will also be a
change in consumers expenditure because some people’s earnings will be
increased and other people’s reduced Thus the effects of this one change
will spread out through the economy rather like the npples which spread out
over the smooth surface of a pond after a pebble has been dropped into it
The price system allows the adaptations to the initial shift to take place
without being consciously coordinated by some single central authority
When shortages develop, prices rise and profit seeking entrepreneurs are
led to produce more of the good that is in short supply ^Vhen surpluses
occur, prices fall and supply is voluntanly contracted The pnee system
produces a series of automatic signals so that a large number of different
decision taking units (firms) do, m feet, coordinate their efforts How well
they coordinate them depends on how well pnccs reflect current and future
scarcities and surpluses and on how fest and effectively firms respond to the
changing price signals
 THE INTERACTION AMONG MARKETS 499

FEEDBACK
One of the characteristics of the interrelated set of markets that forms the
economy is that a change in one market will affect many other markets and
the changes in these other markets may in turn affect (we say feed back
onto) the original market.
In the example given above, we showed that the decision to expand
capacity in the car industry would have many repercussions throughout the
economy. The changes in these other markets might well feed back onto the
car industry. To predict the precise effects of this feedback is a very difficult
thing to do. Certainly, the regional pattern of car sales wall be affected.
Sales will rise in the car-producing areas as workers migrate to these, and
sales will fall in areas which the migrating workers leave. It is also possible
that an overall national increase in car sales could occur if the increase in
employment in the car industry' brings more workers into the income range
at which they will buy a new car, rather than a second-hand one.
These feedbacks, where they are significant, make it difficult to use the
sort of economic theory that we have relied on completely up to this point.
This theory is best described as partial equilibrium theory. We have
already distinguished between equilibrium or static theory and disequi-
librium or dynamic theory* and we must now distinguish between partial
and general theory.

PARTIAL EQUILIBRIUM THEORY

The and general equilibrium are illus-
distinguishing features of partial
trated in Figure 37.1. We start by considering some sector of the economy
-
possibly the market for cabbages, for carpenters or for cars, and we call this
sector A. If there some change in sector A, this will cause changes in the
is

rest of the economy and these changes will in turn reflect back on sector A,
causing further changes in that sector. Let us assume, for example, that the
initial change in sector A
is a fall in the supply of cabbages. This
will cause

an increase in their price and a fall in the quantity bought. The rise in the

price of cabbages will cause other demands to change; in general, we would

expect to find an increase in the demand for goods that are close substitutes
for cabbages and a decrease in the demand for goods that are complemen-
tary with cabbages. As a result, the prices of all these other goods will

change. These are the induced changes in the rest of the economy.
The
our prediction,
original demand curve for cabbages, that we used to derive

was based on the assumption that all other prices were given. Now, how-
have risen and this will cause a shift
ever, prices of substitutes for cabbages
1 See Chapter 12.
500 THE MARKET ECONOMY AS A WHOLE

m our original demand curve Ibr cabbages This is the reflection back of
the induced changes in the rest of the economy on to the original sector II

IS a baste assumftion of parlial egaihbrtttm analysts (hat such effects art small tnoush

to be Ignored
All partial equilibrium analyses are based on the assumption of ceteris

panbus Strictly interpreted, the assumption is that all other things in the

0) Partial equilibrium analysis u possible

(ii) General eqmbbnum analysis is necessary
Fee 371

economy are unaffected by any changes in the sector under consideration

(sector A) This assumption ts always violated to some extent for anything
that happens in one sector must cause changes some other sectors What
m
matters is that the changes induced throughout the rest of the economy are
sufficiently small and diffuse so that the effects they in turn have on sector
A can safely be ignored There is no simple rule telling us when partial
 THE INTERACTION AMONG MARKETS 501

analysis can safely be employed. The final test is in whether or not the
predictions of partial theory are refuted by the facts. As a first approxima-
tion it is probably safe to say that the smaller is the sector under considera-
tion, the more likely is it that its behaviour can successfully be predicted by
partial analysis.

GENERAL EQUILIBRIUM ANALYSIS

General equilibrium analysis attempts to deal explicitly %vith interrelation-
ships between sectors. The micro-economic theories of Parts 1 to 5 are all
partial theories (interrelations are dealt with, but only in an impressionistic,
intuitive way). The macro-economic theories of Parts 7 to 11 are of the
general equilibrium type.
The father of general equilibrium analysis was Leon Walras. It was his
great insight to conceive of the economy as a system of simultaneous equa-
tions in which a shift in a relation (e.g., a demand or supply curve) in any
one equation affected the solution to every other equation (market). Today
we regard this view as excessively static for some purposes, but it represented
the final step in the realisation of the potential interrelation of all markets
in the economy and of the importance of the price system as the coordinator
of these markets. To get far in general equilibrium theory one needs mathe-
matical tools and so it cannot be developed in an introductory book.*

One word of warning regarding general equilibrium theory should be

given to the potential economics specialist. It is sometimes asserted that

‘everything depends on everything For example, it may be said that

else’.

the quantity demanded of any one commodity depends on the price of that
commodity and on the prices of all other commodities. Consider what such
a statement might mean. First, it might be an hypothesis about the world,
in which case it is open to testing. Second, it might mean that everything
could conceivably depend on everything else, in which case it is trite. Third,
it might mean that everything is known to depend on es'erything else,
in

which case it is wTong for, whether or not this is the case, it certainly has not

been established conclusively.

1 It is a which economists sometimes fall, to argue that everything might depend

pitfall, into
on everything and that a theory which does not allow for this is unrealistic or otherwise
else
unsatisfactory. These theorists will then take someone else’s theory’, which was
designed to

account for, and to predict, some real-world observations by assuming that each thing depends
on only a few other things, and they will generalise it by allowing for the possibility that every-
thing might (no one knows that it actually does) depend on everything else. They may
then
many formal properties. In
treat the theory by use of very elegant mathematics and display its

that all
contemplation of the generality and the beauty of the treatment it is often forgotten
exchange for this level of
possibility of usefulness is very likely to have been eliminated in
generality.
502 THE MARKET ECONOMY AS A WHOLE

Consider, for example, the demand for peanuts (ZJj) We may say,

assuming there to be n commodities in the economy, that the demand for

peanuts is a function of all n prices

Di = flPt,fi2, ,P„ >P.) (1)

If we say same thing about all commodities, we will have a demand

the
equation for each commodity, each equation depending on all pnccs In
practice such a theory will prove to be useless because we will not be able
(within the foreseeable future) to obtain sufficient cmpincal knowledge of
theway tn ithtck the demand for any good is influenced by the prices of most
other goods We might start conceptually from equations like (1), which
remind us that the economy is complex and that, conceivably, anything
might be influenced by anything else, but wc must then observe that, m
practice, most of the influences are likely to be negligible Consider, for
example, the effect on the demand for peanuts of a change in the price of
cotton shirts
The construction of a theory which will have cmpincal content, and thus
be useful in the real world, starts by removing most of the vanablcs from
equation (1) Conceivably every pnee in the economy might influence
peanut demand but m fact, so the theory will run, only two or three (or
possibly SIX or eight) prices really matter, the effects of the others ate
negligible and can be ignored Now the theonst will go on to build a theory
based on these simplifying assumptions and to test it against reahworld
observations On the other hand, a theory which says that everything in-
fluences everything else will be so general that all it is likely to predict (m
the absence of a truly vast amount of empirical knowledge) is that ‘any-
thing can happen’ A theory which picks out a few possible relations as the
critical ones will not cover all conceivable cases, but it may work in dealing
with the cases which actually occur

EMPIRICAL KNOWLEDGE ABOUT GENERAL EQUILIBRIUM

In recent decades a number of ambitious approaches have been made to the
problem of breathing the life of empirical applicability into general equi-
librium theory A typical problem would be to determine what would
happen to the economy if to produce an extra 1 million tons of
wc wished
steel for export An increase in the production of steel would require an
increase m the output of many other industries such as the iron ore, the
coal and the glass-making induslncs These industries will themselves require
steel and each other’s products as their own inputs Thus the steel industry
must expand its output not only because more steel is required for export
but because more coal is required to make steel and because the output of
 THE INTERACTION AMONG MARKETS 503

coal requires steel as an input. Furthermore, steel requires more iron ore
and this requires more coal to feed the engines to transport the ore
and this
in turn requires yet more steel. The problem is to discover when all actions
and reactions have been allowed for what ivill be the necessary changes in
the outputs of all the industries, including steel, in order that a net increase
in steel production (for export) of 1 million tons may be produced ?
The most important attempt to deal with these general-equilibrium
problems of the interrelations between markets is that of Professor Leontief.^
It has resulted in empirical measurements of the interrelations in the
economy. These measurements allow us to predict the effects throughout
the economy of changes in any one sector. The theoretical structure of
Leontiefs input-output models (as they are called) is based on many simpli-
fications. The basic unit is a broadly defined industry (or sector) that is

assumed to use both labour and the products of other industries. It is

assumed that industries use their inputs in fixed proportions. This assuming
away of the possibility of factor substitutability (see Chapter 1 7) is done in
order to focus on the adjustments among sectors and the nature of equi-
librium solutions. Actual interrelations are measured and then used to
predict the outcome of various changes in the economy. Performance of
input-output models in predicting the behaviour of market economies has
been moderately satisfactory over short periods of time. Over longer periods,
the absence of factor substitution causes considerable difficulties.
Input-output analysis, and its generalisations, which are called mathe-
matical programming, has become an important part of empirical eco-
nomics,. particularly in answering such questions as, ‘If we attempt to
expand output of military and space products at such and such a rate, where
will bottlenecks develop?’ and, ‘What is the maximum output, given
present techniques and patterns of use, that we can achieve with existing
known quantities of factors ?’.

Input-output measures, when applied to some of the problems of both

underdeveloped and planned economies, prove to be e.xtremely valuable.
In planning, it is desirable to be able to foresee all the major effects of cer-
tain measures. An input— output table of the actual interrelations thus becomes
an important piece of apparatus for the central planner.
Another relatively new approach to measuring the behaviour of complex
general equilibrium systems is to use simulation models?' In essence, the
approach builds an artificial and simplified representation or model of an

University
1 See Wassily W.Leontief, The Structure of the American Economy (2nd ed.; Oxford
Press, 1951).
2 The interested student will findan excellent and nontechnical survey of this field in the
Economic Remew,
paper by Guy H.Orcutt, ‘Simulation of Economic Systems’, American
December 1960.
504 THE MARKET ECONOMY AS A WltOLE

economy m much the same way as aeronautical engineers build wind

tunnels to simulate the atmosphere Expenments can then be performed on
the model In economics, simulation models have used water in pV^s and
electnc currents in wires to simulate the flows of goods and services, as well
as the stock of goods With the developmenl of large scale computers, very
elaborate models can be represented electronically Hypotheses about how
individual units behave can be placed into the memory of a computer, and
the economist can observe how hrge numbers of simulated households and
firms interact in the simulated markets (The behaviour of markets is also
governed by rules programmed into the computer ) The results of such
interactions over time become outputs of the machine Experiments can be
performed on the system for example, the economist can increase ‘taxes
and trace out the effects of the increase month b> month over many ‘years’
Since months of actual time can be simulated in a few seconds on a com
puter, the economist will be able to run many expenments (for example, he
can try out difftrmt lax policies) and compare the results of each This
approach is tn m infancy, and it is too early to evaluate its success, but it

certainly is one of the most exemng of the recent developments on the

frontiers of economic theory

A FINAL WORD
There no doubt that the millions of households and firms in an economy
is

and that their decisions individually small, impinge upon

are interrelated
one another through the market mechintsm Neither the theory of general
equilibrium nor the empincal measurements of the nature or process of
adjustments is anywhere near satisfactory But it is clear that the economy
docs adjust, sometimes imperfectly often slowly, to changing demands upon
It An enormous part of the adjustment process is handled by the pnee
system which no one planned and which is (by and large) unregulated To
say we do not undentand the pnee system nearly well enough does not
mean that we do not have a good idea of the way it works the speed with
which It works, the places where it appean to be least satisfactory, and the

places where it appears to do very wdl Some of our dissatisfactions with the

workings of the price system have led poIicymakcR, ev en m

the most market
oriented of economies, to adopt policies that interfere with the free working
of markets Some of the great efficiencies of the pnee system have led policy
makers m
economies where everyone is ideologically committed to planning
to introduce pricing schemes to guide a large part of the dcasion making
process
CHAPTER 38

MICRO-ECONOMIC POLICY

If you visitany Western country and talk to politicians, academics, business-

men and workers, you will discover two caricatures of each country you
visit. The details will vary', but by and large one caricature \sall present the
country as being involved in a mad, brutal chase for money with the pursuit
of private profit leading people to cast aside individuals, ideas and institu-
tions which do not meet the test of being ‘money-making’. In the second
caricature, the businessmen, workers and farmers of the country' will be
pictured as slowly strangling in a web of initiative-destroying regulations
spun by a spider called government control. If you have any doubt that
both such caricatures are abroad in your country', a careful attention to the
speeches of the members of various political parties at the next election
should dispel this doubt.
Neither caricature is realistic. It is true that many aspects of economic
hfe are determined by the operation of a free-market sy'stem. Private
preferences, expressed through private markets and impinging upon private
profit-seeking enterprises, determine what is produced, how it is produced
and how the product is shared. But even casual observation of, and assuredly
a close look at, any economy will make it clear that public policies and
public decisions play' a large role too. Not only are there laws that restrict
what people and firms may do, but there are taxes and subsidies that affect
choices, and large amounts of public expenditures that are not determined
solely by market demands and that lead to the production of a very
different ‘bill of goods’ and to a different distribution of national product
from what would exist in a system that relied entirely upon private markets.
In this chapter, we are concerned with the nature, form, purpose and
effect of various activities of the central authorities that impinge upon the
unrestricted workings of the ffee-market system. We
are particularly con-
cerned at this time with policies that affect the distribution of resources
among uses, and the distribution of income among people, rather than those
 MICRO-ECONOMIC POLICY 507

decide who he is going to satisfy. He might work on the first-come-first-

served principle, or he might serve people whose colour, manner, dress or
sex he liked and refuse to serve people whose colour, manner, dress or sex
he disliked.
Furthermore, the incentive for bribery and corruption is clearly present
in such situations. If, for example, state housing is available at prices well
below the market price, there will be a queue to get into the flats that do
become available. The official who controls the allocations from the waiting
list will be in a position of power and could take bribes for pushing someone
up in the list. If the state charged the market price for its housing, this
opportunity for corruption would be removed because, temporary and un-
foreseen fluctuations aside, everyone who was prepared to pay the price
could obtain accommodation.
There are other arguments in favour of the price system but these can be
left fora more advanced course, on the assumption that in most Western
countries students have a predisposition to accept the value of the free
market and most time needs to be spent on reasons why the free market
cannot be left to do everything, instead of persuading people that it can do
something. The major advantage of the free market, which is the one we
have studied throughout this book, is that it provides a method for decen-
tralising decision-taking while assuring that such decisions are at least
moderately well coordinated.

THE CASE FOR INTERVENTION

If motivation can be judged by behaA'iour, it is clear that a majority of the

Western countries believe that an unrestricted, unmodified,
citizens of all
free-market system is not the best of all possible worlds. What led free and
reasonably rational men to the conclusion that their own individual liberty

to do exactly what they wanted in every aspect of their behaviour should

be, tosome extent, curtailed by the central authorities? There are several
very important general reasons for interfering with the exercise of free
choice operating through a free market. We must now consider the most
important of these.

1 Market Imperfections

The mechanism of a free-market economy is the signals that are pro-

basic
vided by prices, costs and profits. These signal relative scarcities and sur-
pluses and induce businessmen and the owners of factors of production to
move from areas in which demand is low relative to supply to areas in which
508 THE MARKET ECONOMY AS A WHOLE

supply low relative to demand There are many reasons why this auto
IS

matic signal and reaction system may not work to our complete satisfaction
and why we might wish to intervene to improve its functioning

Factor movements If factors are relatively immobile, the supply of

factors will tend to be inelastic in the short run, and even large increases m
the price offered for scarce factors may induce only small movements of
such factors As we saw jn Chapters 33 and 34, if theoretical phj'sicists are

in short supply their pay will nsc, and an increase in the pay of existing
theoretical physicists may ultimately lead to an increase in the total supply

of physicists (for example, by persuading students who had planned to be-

come economists or engineers to go into physics instead) But such increases
in pay may have only a small effect on supply for several years, in view of

the training required to become a physicist Meanwhile, existing theoretical

and will command a larger share of national
physicists will earn high salaries,
income than they otherwise would have done It may well be that public
policy can bring about an increase m the supply of theoretical physicists
more quickly and at lower cost than can the unhampered market The
central authorities might, for example, provide very generous scholarships
and studentships to encourage students to take degrees m physics ^ In
Britain the Department of Scientific and Industrial Research and more
recently the Social Science Research Council have done precisely this In
America the same thing has been done by the Atomic Energy Commission
and the National Science Foundation
Factors may be immobile for reasons other than time lags required for
education For example, unemployed workers may be uninformed about
job opportunities or inadequately trained to take jobs that arc available
Agenaes such as Employment Exchanges can increase the speed with which
such workers can find jobs by collecting data on what kind of jobs ate
available m
what places, and informing unemployed workers about them
Special programmes of retraining are also provided by some governmental
units Factors may often be reluctant to move from one location to another
because of the heavy moving expenses, or the difficulties m finding housing,
or racial prejudice in hiring Public poUaes may lessen these imperfections
by giving subsidies for expenses incurred in moving from one job to another,
by building public housing (or subsidising pnvatc housing) in areas that
have labour shortages, and by legally prohibiting various forms of discnmi
nation with respect to the hiring of imnonty groups
Some attempts at encouraging mobility in these respects have been made
in Britain and m the United States The country that has gone furthest in
1 This provides incentives for new entrants to the labour force to become physicists without
providing extra earnings to ail these who have already decided to enter that profession
 MICRO-ECONOMIC POLICY 509

this respect, however, is Sweden. Here movement and retraining allowances

are extremely generous and have the purpose of enlisting the cooperation
of the worker in accepting the inevitable dislocation caused by economic
change.
Government policy can, on the other hand, discourage this type of
mobility. In Britain, for example, local councils provide housing at prices
often well below the market rate. As predicted by our theory, the quantity
demanded exceeds the quantity supplied at these prices. The rationing is

normally by queues often with a residence-in-the-local-area requirement in

order even to be accepted in the queue. This policy clearly acts as a disin-
centive to regional mobility, for the inhabitant of a council flat who con-
sidersmoving elsewhere must accept a long period of non-council dwelling
(at high rents) before he secures a council flat in his new community. This
sort of central-authority-induced impediment to regional mobility is the
opposite of what is required if we wish to encourage factors to move in
response to the changing signals of the price system.*
Another kind of market imperfection is ignorance. Signals cannot work
if they cannot be read. Providing market information, or enforcing a re-
quirement that others provide it, is a major activity of many government
agencies whose function is to make markets work better. These activities

range from requiring food manufacturers to identify the contents and the
net weight of their product, through the government inspection and grading
of meat, grain and other agricultural products to the provision of market
information on where jobs exist and at what level of earnings.**

Firm behaviour; Profit maximisation in a world of perfect competition

makes firms into cyphers, following market signals without exerting any
personal pressure on the outcome. When we depart from perfect competi-
tion and/or profit-maximising behaviour, firms achieve the power to inter-
fere with the behaviour of the market. Profit-maximising monopoly presents
one pure case. Monopoly power prevents resources from moving in response
to market signals. A rise in demand for a monopolist’s product will cause a
rise in the monopolist’s profits^; in a perfectly competitive industry this
would lead to an increase in resources allocated to the production of this

1 Of course this discussion is not sufficient to lead to a condemnation of public authority

housing. Public housing has many goals which it does achieve and many side effects, some
desirable and some undesirable. What we are discussing above is one undesirable side effect.
2 Before you dismiss this imperfection as trivial, ask yourself if you have any clear
idea of the

expected lifetime income of a person following the subject of study you have chosen and of
persons following the subjects your friends have chosen. Also can you compare lifetime
earnings

of the few people who are unusually successful with the earnings of those who are only averagely
successful ?
3 See Chapter 23.
510 THE MARKET ECONOMY AS A WHOLE

good until profits fell to normal* but if the monopolist can restnet entry
into his industry, no such resource re-allocation need occur
Even more important is the point that if the firm is not concerned with
profit maximisation, the whole market mechanism can break down com
pkttly A m demand wdl not even nectssardy be met by a nsc m out
put of the existing monopolist if he is not a profit maximiser
Oligopoly can also cause market reactions that are not m accordance
with the perfectly compctiuve result Since firms are conditioned by the real
and imagined reactions of their very few nvals, almost any conceivable
result can follow from such an unambiguous change in market signals as a
rise in the price of some commodity or the fall in the price of some factor

of production Thus m oligopolistic situations we arc unsure about how

satisfactorily firms will respond to changes m market signals

2 Differences Between Private and Social Costs and Differences Be

tween Private and Social Revenues

In a free-market economy, firms make decisions m relation to pnvate costs

and private revenues In Chapter 19 we noted that private and social costs
might diverge considerably and it is extremely important that the student
should re*read this critical section (sec pages 253-4) before reading on
Not only can social and pnvate costs differ, so also can social and pnvate
revenues differ
First, let us define these two revenue concepts

Private revenue The revenue that the firm obtains by selling its

product

Social revenue The money value of the gains that everyone obtains
(i e what they would be prepared to pay to get the gam) from the pioduc
,

tion and consumption of the good in question

Just as social and pnvate cost can diverge, so also can social and private
revenue diverge If a nun considers that u is jusi worth the £50 it costs him
Tt> rent; VriTTfttW tS an iTdcctiDos dnsease, then Viis pnvaTt hA®,

pnvate revenue Clearly, however, social revenue exceeds this figure be-

cause other people will not now be

exposed to the infectious disease,
whereas they would have been if the suRcrer had not elected to take the
cure Clearly, where individual A elects to do something that favourably
affects other individuals in the economy, total social gam of A’s decision
exceeds the pnvate gam to A Clearly, also, a case is imaginable in which

1 See CViapitr
 MICRO-ECONOMIC POLICY 511

A will just consider the action not worth taking, where if all the other
affected individuals could have their say would be worth taking.
it

The divergence between social and private costs and revenues is one of
the most important reasons for interference with free markets by central
authorities. Private cars, for example, excrete quantides of carbon mon-
oxide fumes into the atmosphere, thus contributing in a significant way to
atmospheric pollution. In the United States, the State of California has
recently required that anti-smog devices be attached to the exhausts of all
automobiles, presumably because the people of California (through their
elected representatives) attach a higher cost to a polluted atmosphere than
the costs individual drivers or automobile manufacturers are willing to pay
to avoid pollution.
There are many examples of policy problems caused by such divergences.
One important example concerns road versus rail transport. Many people
believe that the social costs of transporting freight by road greatly exceeds
the private costs. They point to the facts that trucks use the roads free, that
the cost of roads would be greatly reduced if they did not have to be built
strong enough to sustain the pounding of heavy lorries, and that road con-
gestion, which is aggravated by lorry traffic, confers heavy costs on other
users, particularlywhere a road goes through a town. A classic case of the
failure of theunhindered price system would occur if the cost situation for
freight transport were as follows:

ROAD HAULAGE RAIL HAULAGE

private cost < private revenue private cost > private revenue
social cost > social revenue social cost < social revenue

If this were the case the free market would throw freight traffic onto the

roads, while from society’s point of view the margin of advantage lies with
rail traffic.

Another example concerns zoning laws. If A owns a building lot in a

residential area (on which he intends to build a house in five years) and
decides to get some money out of it in the meantime by leasing it to a farmer
to pasture his goats or to a fairground operator, the neighbours may well
object. In a society that values individual freedom, we have a predisposition
not to interfere with bargains freely negotiated between adults. But if, as in
the above example, these bargains confer substantial costs on third persons
not party to the bargain, there is a case for intervention.
In cases where such divergences are clearly identified and measured it is
possible to adjust for them by taxes and subsidies. Where social costs exceed
private costs a tax on the production of the good may bring private costs
more in line with social costs and where the opposite relation occurs a sub-
sidy can be used. Road and petrol taxes provide an example, although
512 THE MARKET ECONOMY AS A WHOLE

many feel that they still leave pnvate costs well below soaal ones Free park-

mg m the centre of urban areas creates a situation of zero pnvate cosu in

the face of substantial social costs, and the recent introduction of parking
meters represents more than just another way of raising money, it is also
an attempt bring private and social costs into line and make the person
to
who incurs the cost on society bearit himself Taxes on cigarettes may to

some extent reflect the fact that the social costs of cigarette consumption -
m terms of expensive medical and hospital services required for smokers in
later life ~ exceeds the private cost of the cigarette to the consumer On the
other hand, subsidies on housing and medical services reflect the belief that
the social gams from removing slums and producing a high standard of
health exceed the private gains
Problems arise where divergences cannot be identified or where they can-
not be quantitatively measured even when they are located In such cases

the price system Will produce a use of resources other than the one that
would be adopted on social cost-benefit considerations

3 Collective Consumption Goods

Certain goods or services, if they provide benefits to anyone, nccessanly
provide them to a large group of people Such goods are called collective
CONSUMPTION GOODS National defence is a prime example of a collective
consumption good If we have an adequate defence establishment, it pH>*

tects us allIt protects you, even if you do not care to ‘buy’ any of it The
quantity of national defence to be provided must be decided cofltttwtly,
and there no market where you can buy more of it and your ncigbboui'
is

less Once you have agreed with him (and the rest of us) on a compromise
quantity, there is no market where you can be made to pay for your
share

of the good The government must acquire the funds to pay for it by the
compulsory scheme of taxation
There are many other examples The beautification of a city provides a
service to all residents and visitors A barrage that protects a city from a
flood IS also a collective consumption good So also is a humcane-waruing
system, very desirable in areas subject to hurricanes, but a public warning
IS by everyone Another important example is police
necessarily ‘consumed’
protection If a police force reduces thenumber of crimes, everyone gams
Even if m a market system you did not pay to have the police watch your
house you would gam from the fact that your neighbour did
One final exaropic parallels many real and very important situations
which have arisen m the United States where the production of electricity
by water power is extremely important Suppose the pnvately-owned Skunk,
Power and Light Company decides to build a dam across the Muskrat
 MICRO-ECONOMIC POLICY 513

River in order to produce hydro-electric power to sell on the market. Its

engineers and economists determine the best height of the dam to be 80
feet.* When these tentative plans are announced, someone realises that
if
the SP&L Company builds a 100-foot dam instead of an 80-foot dam, it
wall create a beautiful lake in back of the dam, provide an emergency water
supply for nearby communities, and provide flood control for 200 miles
downstream. Suppose the beneficiaries of these extras are prepared to pay
for them once they exist. The SP&L Company realises, however, that it has
no effective means of charging for them.^ It decides, therefore, to build the
smaller dam. If benefits are not marketable, it is unlikely that private pro-
ducers will be motivated to provide them. The very large role of the
American central authorities in multi-purpose, w'ater-resource projects is

explained to a great extent by the fact that they alone have mechanisms for
financing such projects.
In general, market systems cannot compel payment for a collective good,
since there is no way to prevent a person from receiving the services of the

good if he refuses to pay for it.^ Only governments, through their power to
tax, can compel payments by all. Indeed, it is the existence of collective
consumption goods that necessitates putting some of our production into
government hands and that prevents the gov^n^ment from selling ever)'-
thing it produces on the free market just as .irm does (and as the .

government does wdth postal services but Joes not and cannot do with
military and police services).

4 The Costs of Collecting Revenue

The power of tax enables governments to compel citizens to pay for collec-
tive consumption goods. Private concerns have no such power, and for this
reason no private concern finds it worthwhile to provide such goods. There
are similar problems \vith respect to some goods and services that are not
strictly collectively consumed. For example, suppose motorists in a big city

1 At this height, the company will maximise its (private) profits.

2. Consider the lake. The higher dam will flood certain lands but create valuable shoreline

property. The owners of this property will gain if the lake front is on their property. Suppose
the SP&L Company said to these people: ‘After we create this lake you should pay us
year because of the increase in the value of your property.’ And suppose the typical owner
answered; ‘Look, you’re flooding my south forty acres and you want me to pay you? 1 11 sue.
Governments, on the other hand, have means whereby they can acquire land, and where, if
land values rise, recover their costs through land taxes.
3 The demand for a collective consumption good the vertical sum of the demand curt es
is

of all individuals for the good. Compare this resultwith that discussed on page 88, and
especially with footnote on that page. The
I reason for the difference is that you and we can
consume the same unit of a collective consumption good. If you are willing to pay ^10 for it,

and I am willing to pay £1, together we are willing to pay £\1.

17
514 THE MARKET ECONOMY AS A WHOLE

arc Willing, even eager, to pay to have a high-speed urban motorway system
leading into and out of town Suppose that there are enough people willing
to pay 3</ a mile to cover the costs of building such a road system, but that
different groups of them want to use different sections of the system A
private company would find it profitable to build and operate the road if

It 3(f a mile from everybody willing to pay that much But, if

could collect
It must build a toll house at every entrance and exit to the road in order to

collect this money, the costs of the system would be vastly increased and the
venture might seem an unprofitable one Intra-urban motorways with
many access points and many short-joutney travellers are often unsuitable
for private ventures, because the cost of collecting tolls is too high It is no
accident that in countries where privately built toll roads arc common
virtually all toll roads are tnter-urban roads which require relatively few
access points and where the average journey is a long one
To take one more example, suppose every citizen of a given city is willing
to payof I per cent of his annual income in return for a concert hall and
-i

auditorium, but that any flat admission pnee to the auditorium would not
cover the costsA scheme of pnee discrimination according to income,
however, repaysall costs There is no mechanism by which the private

company can discnminaie among buyers of tickets, because it cannot pre

vent resale A government, however, with its power to tax incomes, could
readily finance the project

5 The Desire to Protect Some Individuals from Decisions Taken on

their Behalf by Others

In economics we take the household as the basic decision taking unit We

must not forget that the typical household contains several persons and
that the choice problem within the household is a political one Whose
desires are to be favoured and by how much in making the purchasing
decisions for the household ^One example should suffice to show how impor-
tant this point is In an unhindered free market the adult members of the
household will make deasions about how much education to buy and there-
by exert a profound effect on the lives of their children A selfish parent may
buy no education while an cquahtanan one, may buy the same quantity
for all his children The central authorities may interfere in this choice both
to protect the child and to ensure that some of the scarce educational re-
sources are unequally distributed according to intelligence rather than
wealth We
force all households to purchase education for their children
and we provide strong enducements, through scholarships, etc for the ,

gifted to consume much more education than either they or their parents
might voluntary select if they had to pay the whole cost themselves The
 MICRO-ECONOMIC POLICY 515

latter done probably because of a divergence between social and private

is

costs,but the former is done in the belief that the head of the household
should not have perfect freedom to take decisions that affect the welfare
of other members of the household, particularly when they are minors.

6 The Belief that Some Social Obligations Transcend the Market

Criteria

In a market system, if you can pay for something you can have it. If you
have to clean your house and if you can persuade someone else to do the
job for you in return for 15 shilh'ngs, presumably both parties to the trans-
action are better off: you would prefer to part with 15j rather than clean
the house yourself and your cleaning woman prefers 15 j to not cleaning
your house. Normally we do not interfere with people’s ability to negotiate
such mutually advantageous contracts. We do not feel this way, however,
about some activities which we regard as social obligations. A prime example
is provided by military service. At times and places in which military ser-

vice is compulsory contracts similar to the one between the housewife and
her cleaning woman could also be struck. Some persons faced with the
obligation to do military service could no doubt pay enough to persuade
others to do their turn of service for them. By exactly the same argument as
we used above, we can presume that both parties will be better off if they
are allowed to negotiate such a trade. But we usually prohibit such contracts
by law. Why ? We do this because we feel that there are values other than
those that can be expressed in a market. We feel that, in times when it is
necessary, military service by all healthy males is an obligation independent
of their tastes, wealth, influence or social position. We feel that everyone
ought to do and we prohibit trades between willing
this service traders which
in respect to most services we would allow.

7 Compassion

A free-market system rewards certain groups and penalises others. The

workings of the market may be stem, even cruel; consequently, it may seem
humane Should unproductive farmers be starved off the farm ?
to intervene.
Should men be forced to bear the full burden of their misfortune, if, through
no fault of their own, they lose their jobs? Indeed, even if they lose their
jobs through their own fault, should they and their families have to bear
the whole burden, which may include starvation? Should the ill and aged
be thrown on the mercy of their families? What if they have no families?
Should small businessmen have to compete with the chain store and the
discount house? A great many government policies are concerned with
modifying the distribution of income that results from such things as where
516 THE MARKET ECONOMY AS A WHOLE

one starts, how able one is, how lucky one is and how one fares in the free-
market world
The range of governmental activities designed to redistribute income is
enormous The graduated income tax, in which the proportion of one’s
income paid as a tax rises steeply as one’s income rises, is, of course, a major
policy tool of this kind, and we shall discuss it m detail later m this chapter
Direct public expenditures, such as payments to the aged (Old-Age Pen-
and government-supported social-insurance schemes (Unemploy-
sions),
ment Relief and National Assistance), have as their main purpose the
alleviation of financial hardship among those groups most disadvantaged
by the market economy Mtmmum-wage legislation may succeed in raising
incomes of some of those at the lower end of the income scale, although, as
we have seen, it may also make unemployable some people who could have
found jobs at wages below the legal minimum Most of our basic agricultural
policy of price supports can best be understood as an attempt to raise the
witriTnts n? vsoiV. cm cht land esamg fesTcvtrt’
incomes also has important effects on the functioning of the economy, but
the mam motivation for these policies seems to come from a special concern
for the well-being of the small farmer ) Small business is given preferential
treatment in many ways and is subsidised to some extent in its struggle to

survive in an economy of giant corporations

In the discussion above, we noted some of the reasons why people acting
through their government may wish to achieve a different quantity, quality
and distribution of goods than the free » the case
market provides This
against complete lausez/atre But members of a may have other goals
society
as well One of these is a belief in the individual’s freedom to act on his own
and to make hts own decisions Multiple goals often involve conflicts, and
conflicts require choices When and where and to what extent to interfere

with the free-market system, usually involves decisions as to whether a

certain specific objective is worth the loss of individual freedom that is
required to achieve it

ECONOMIC ANALYSIS AND ECONOMIC POLICY

The Pervasiveness of Policy Decisions

All governments have economic policies Governments derive their policy

authority from their police power - indeed, the words policy and police
come from the identical Greek word, pohuta
Some governments may lean toward a policy of laissez faire, or noninter-
ference, others aim for a policy of strict control over every facet of the
economy Even the decision not to act, to let nature take its course, is a
policy decision In a democratic country, jralicies are determined by pro-
 MICRO-ECONOMIC POLICY 517

cedures and are decided under rules that have been adopted by a govern-
ment selected with the consent of the majority; in a dictatorship, policies
may largely express the will of a small group. In a free-market economy, for
example, many ‘decisions’ are left to the market. This type of decision is as
much a policy decision as the decision of the government to le\y a tax on
cigarettes. Not every facet of policy is debated anew every year; indeed,
many policy decisions now in force (such as giving unions the right to
organise and women the were made decades ago. At any one
right to vote)
time only a few major policy decisions are being debated that receive a
great deal of attention. In any one year, thousands of economic-policy
decisions are made by local and national governments that never become
the subject of serious public debate.

The Relation Between Ends and Means

Any policy has two aspects the ends that the decision-makers are attempt-
:

ing to achieve, and the means by which the desired ends are to be achieved.
Governments pursue simultaneously many broad policy goals, such as
justice, progress, national security and economic stability. To achieve the
goal of justice, governments may decide to improve the economic status of
disadvantaged groups - the needy, the aged, immigrants, the mentally de-
fective. To achieve the goal of stability, the government may decide that
unemployment or inflation must be controlled, or that foreign-exchange
imbalances must be corrected. Once such a decision is made, it is translated
into a number of more specific goals. The broad concern with justice may
eventually be expressed, for instance, in a national minimum-wage law.
The economist asked to evaluate some particular policy- a minimum-
wage law, say - must ask himself two questions: ‘What is the specific objec-
tive this law is meant to achieve?’ and ‘What is the general objective of the
policy-maker?’. The economist wll have no chance of understanding
minimum-wage legislation unless he can relate this policy both to the objec-
tive designed to serve and to the underlying goal that dictated that
it is

objective. To take a different example, a change in the tax laws providing

for an investment credit is designed to stimulate investment which is an
objective- pursued in order to promote economic growth. The economist
asked to evaluate this policy must ask whether it will stimulate investment,
whether this stimulation will contribute to economic growth, whether there
are feasible alternative policies, and what the costs are that are involved.
The person who criticises the policy of the investment allowances on the

ground that it helps the big companies but does nothing for the farmer is

not coming to grips with the real issue: if the investment allowance in-
creeises the amounts businessmen invest, either because it increases their
518 THE MARKET ECONOMY AS A WHOLE

incentives or because it provides them with available funds, it mil lead to

an increase m production and thus help to realise the goal of economic
growth No one particular policy measure will contribute to the 3ch!e\e
ment of all our policy goals It is die economist’s business to determine if it
contnbutes to some of our goals and at what cost
People sometimes speak as if economic theory jusfijfcif certain poUcy objtc
tives For example, it is not uncommon to hear someone say that the policy

of rent control is tconomtcally unsound, or that the policy of the agncultural

pnce supports is tconomic nonsense, or that the only laltd economic policy is
to charge a pnce that co\ers costs Such viewpoints often represent an
effort to dismiss a policy mthout thinking through its consequences Every

time a student encounters the catch phrase ‘economic nonsense', his every
cntical faculty should be aroused ‘What does this really mean ^ he should,

ask, and Why is it dismissed as nonsense’ Sometime this form of words

represents an attempt to argue that economics proves what ought to be done
In this case the student will recognise it as an impossible attempt We have
argued in Chapter 1 that it is impossible to deduce a statement about what
ought to be from statements about what is Posmve economics concerns
whai IS and cannot, therefore, produce statements about what ought to be

The Rote of the Economist

What then can we expect from economics when we come to consider prob
lems of economic policy ’ Let us take a specific issue of policy to see in more
detail what we can and cannot expect from economics Consider the posi
non of an economist asked to examine the case for and against rent control
(i e government fixing of the rent of private dwellings) How should the
,

economist go about evaluating rent control ^ First, he should ask what goals
rent control is meant to achieve He might find that rent control is intended
to redistribute income from nch to poor, to ensure a minimum standard
of housing for everyone, and to maintain general housing standards Next
he must ask if rent control does in fact help to realise these policy goals If
the answer is ‘No then that is the end of the story The case against rent
control IS then clear it docs not achieve the objects for which it is intended

If the answer is ‘Yes then further study is necessary Let us consider, as an

,

example, the goal of income redistribution Rent control, if effective, means

that the tenant pays less rent than he otherwise would, and the landlord
receives less income than he otherwise would , thus rent control redistnbutcs
income from landlord to tenant But do landlords tend to be nchcr than
their tenants’ If a survey shows that most tenants are in fact nchcr than
their landlords, the economist can conclude that there is a case against rent
control, not because rent control is unjust or unethical, but because it does
 MICRO-ECONOMIC POLICY 519

not achieve the policy goal for which it was being used. If, however, the

surv'ey indicates that most tenants have lower incomes than their land-
lords, then the economist mil conclude that rent control is a means of
obtaining the desired goal of income redistribution.
The economist must next ask if rent control has effects that conflict with
other policy objectives. It may be, for example, that, although rent control
provides low-cost housing, it simultaneously causes the appearance of more
slum areas. When one measure helps to achieve one goal but hinders the
attainment of another, it is necessary' to decide which of these goals is pre-
ferred. The economist then must consider the alternative policies to see if
there are any other measures that will achieve the goal at a lower sacrifice
in terms of setbacks to other policy' objectives. It may be, for example, that
the progressive income tax redistributes income from rich to poor with more
certainty and precision and with fewer undesired side effects than does rent
control.
The economist must now' determine ifit is feasible to adopt the alternative
means. In other w'ords, when faced with evaluating a policy measure for
achieving some goal, he must ask whether other measures that are feasible at

the time and place w'ould better achieve the desired goal.
If, hating done all this, the economist concludes that rent control does
achieve the desired policy, that the undesirable effects in other directions
are judged (by the policy-makers) to be less important than the desirable
effects in achieving the stated policy goal, and that there are no other
practicable measures that w'ould better achieve the goals, he will then con-
clude that there is a strong case in its favour.
There are many pitfalls in attempting to apply this procedure. First, the
economist himself will usually have strong views on the particular measure
he is attempting to assess. If he does not like the measure, he is hkely to be
relentless in searching out possible unwanted effects and somew'hat less than
thorough in discovering effects that help to achieve the desired goals. It is

difficult to guard against an unconscious bias of this sort, and the best
method of doing it is to ensure that many economists of different political
persuasions ail examine the issue.

Second, it is easy to forget that the statements that emerge from standard
theory are predictions, the validity of which, in particular instances, can only
be estabhshed by' empirical testing. The economist might, for example,
‘demonstrate’ on the basis of a current economic theory that rent control
would cause landlords to spend less on repair and maintenance than they
would in its then might conclude that one of the valid argu-
absence. He
ments against rent control is that it will lead to inadequate repairs to hous-
ing, but he should say that standard theory' predicts that rent control w'ill
have these consequences (generally held to be undesirable) in regard to
520 THE MARKET ECONOMY AS A WHOLE

repairs to housing The confidence that we place in this prediction must

depend on the extent to which the general theory from which the prediction
has been extracted has already stood up to severe tesung If the general
theory has rarely or never been tested, then we should not have much con-
fidence in this particular prediction until it has been tested itself If, on the
other hand, the theory has been extensively tested m other contexts, rc
should examine the particular case m hand to sec if it differs from other
tested cases in circumstances that our theory suggest to be relevant Thus,
if as IS so often the case, we are dealing with theories that are substantially
untested, at least m a relevant context, we must conclude that the theorist

can only predict for us certain consequences of rent control, these predic-
tions tell us what to look for when wc turn to a study of the world It is a
common pitfall in considering policy measures to think that the job is done
once the economist has predicted the likely effects, at this state the job has
justbegun The predictions need to be tested
It vs the role of the economist not only to analyse the consequences of a
proposed policy (or to compare two or more policies), but also to suggest
policies Given a statement of the objectives, economic analysis can be used
to invent or publicise proposed policies that have not previously been under
consideration He might, for example, point out that some radical new
measure such as running all public transport in urban areas free of charges
might be a hiiherto-unthought-of means of achieving a number ofgcnerallj
accepted goals Free public transport
‘
m
all major urban areas might

vastly relieve congestion by dissuading people from bringing private cars

into these areas, it might also relieve urban overcrowding by persuading
people to live further out of town and there might be a large social saving
,

in time and effort brought about on both these counts, finally there would
be a direct saving in resources because conductors, ticket salesmen, ticket
collectors, inspectors and many traffic policemen could be dispensed with
This labour would be freed for use elsewhere to produce needed goods and
services The case for and against this particular policy would need much
more than can be given here but the point is made
careful consideration
that the economist has a function m
suggesting quite new ways of achieving
old goals In this case the goal of relieving transport congestion might be
achieved together with additional savings resulting from the freeing of
labour for other uses
One of the greatest advantages of the economist in dealing with policy
matters such as those mentioned here is that he is trained to look for

1 Tl bj no means novel pobey suf^eition is chosen for illustramc purpose* only A

IS IS

carefully conducied juantii^liDe study would be necessary before one could come to any con
elusion on the net advantage of this scheme
2 The discussion on the costs of rantng revenue (see page 513) is relevant at this point
 .

AlICRO-ECONOMIC POLICY 521

consistency in people’s ideas about the world. Where someone else may look
at ideas and policies one at a time, the economist has a predisposition to look
at themall at once to see if they make a pattern, and to ask if they are con-

sistentone with the other. Indeed, the training of the economic theorist often
leads him to value consistency above all other things. This is, of course, a
great strength for it is seldom useful to hold inconsistent ideas; to act, for
example, at one time as if all relevant demand curves were inelastic and
then immediately afterwards to introduce a policy which would make sense
only if they were all elastic. This predisposition is, however, not without
major disadvantages as well as advantages. It means that, if there is a
controversy between two groups, the side with a fully consistent theory
which, if it were tested, would be conclusively refuted, is put at a tremen-
dous advantage over the side which has looked at the world and has many
correct empirical observations in mind, but which has not yet worked these
into a fully consistent model and which holds some inconsistent views. The
history of economics provides examples where an incorrect but fully
elaborated model triumphed over a view that was substantially correct but
not yet rtade fully consistent. This is a very humbling obser\'ation and the
situation is made worse by the fact that we can say with finality whether a
given set of assumptions is consistent with one another, whereas we can
never hope to prove or to refute the empirical validity of our theories with
finality (although we can assess the probability of their being right or
wrong)
What then can we do ? First, we must beware of going to the other ex-
treme and arguing that logical consistency is unimportant while only
empirical testing matters. Clearly a theory that assumes at one stage that
all demand curves slope downwards and at another stage that all demand

curves slope upwards is not a desirable theory for it will always produce
self-contradictory statements. The demand for consistency - that we do not
contradict ourselves - is a necessary one for satisfactory explanation in

economics as in other fields. However, this is not enough. A perfectly con-

sistent theory may be utterly wrong and this can only be found out by
testing our theories against observations, A vast advance in the use of
economics in policy matters (as well as in economics in general) would be
achieved people accepted the statements of economists only after they
if

had asked ‘What is the evidence?’ and ‘Has this theory been subjected to
a test where it had a real chance of meeting with conflicting evidenceV. Let us men-
tion one further example. Some people believe that inflation actually
reduces output and thus reduces average living standards because it re-
duces the effectiveness of competition. Let us say that we challenge the
holders of this view for evidence and in response are given a model that
predicts this will happen (‘these are the theoretical “reasons” ’, we will be

17
522 TUr MARKET FCOMOMY AS A \SHOLr

told) The correct rcpl> to th« w I now sec i!iC tlieory from which >ou
derived this prediction but I where ts the evidence’ Untested prr
dictions of thcor> do not constitute evidence

Conflicts of Pol cy

Central authonties have a multiplicity of policy goals It is quite impossible

for them to fulfill all of these goals simultaneously Although one measure
may bring us closer to some of our objectucs it may take us further away
from others For this reason iffbr no other policy decisions arc subject to
disagreement and debate Any policy action limits someone s freedom but
if It IS a means to a preferred choice of goals or a preferred distribution of

income policy makers may well choose the policy that wall reap the greater
benefits Unemployment compensation for example may hinder the
quickness with which labour moves from labour surplus to labour scarce
occupations thereby mcreasmR the total unemployment in the country
but at the same time it may protect unemployed families from dehil tating
hardship Which is more important to protect some families from hard
ship or to risk raising the over all level of unemployment’ Economics cannot
answer such value questions hui someone must provide tlie answers for
decisions have to be made h ts not sufficieut for central authonties to
decide which objectives arc worth piinuing they must also decide on some
lait oj subiUlutm between them They must decide how much of one it is

worth sacrificing in order to get more of the otlicr

The significance of this point is frequently overlooked It is never enough
to show that a proposed policy advanccsoneof society sobjectivcs Virtually
any proposed action meets such a test and usually the exactly opposite
actvoft also meets vt ^For example owe might argue that meieastcig the
income tax on company profits and spending the increase on needy children
would increase the equity of the distribution of income But cutting the
tax on company profits and reducing the amount spent on needy children
might stimulate corporate expenditure for research and development and
thus contribute to growth ) What must be shown of some proposed policy is
tlial it advances certain of our objectives sufficiently to overcome the cost

in terms of the amount that it moves us away from other objectives

MAJOR POLICY TOOLS

The central authorities have three mam tools to affect the nature of what
goods arc produced and how they are distnbuted rule making taxation
and public expenditures
 micro-economic policy 523
1 Rule-making: Rules, in a multitude of forms, require, or prohibit,
certain activities. Rules require people to send their children to school and
to have them inoculated against diphtheria; rules prohibit people from
selling or using certain narcotics and from colluding in the pricing of com-
modities. Rules affect other areas as well, especially transportation, utilities
and communications.

2 Taxation; The nature of taxes and subsidies can affect both what is
produced and the distribution of income. Changes in the way taxes are
raised can redistribute income and can change the relative prices and profits
in different industries.

3 Public expenditures: Public expenditures can be used to produce

goods that the market leaves unproduced, to increase the income of some
groups in the population, and to change the signals to which households
and firms respond.
These different kinds of tools provide aliernattve mtans for achieving given
policy objectives.

Rule-making as a Tool of Policy

Rules pervade over society. Shop hours and ivorking conditions are regu-
lated. Children cannot be seiv'ed alcoholic drinks. Prostitution is prohibited
in many societies even ivhen it is between a \villing buyer and a willing
seller. Many business practices are illegal. In many countries you are forced

to purchase insurance for the damage you might do wth your private
motor car even if you don’t want to carry insurance. In many countries a
person who offers goods for sale cannot refuse to sell them just because he
does not like the customer’s colour or dress. There are rules against fraudu-
lent advertising and the sale of substandard, adulterated or poisonous foods.
In some coimtries, such as the United States, anyone who wants to can
purchase a ivdde variety of weapons ranging from pistols to machine guns.
In other countries, such as the United Kingdom, it is extremely difficult
for a private citizen to obtain a gun of any sort unless he has very good
reasons for doing so. Rule-making is an important method of regulating
economic and social behaviour.

Tax Structure as a Tool of Policy

In 1963-4, taxes raised in Britain amounted to ,^6,890 million which is

approximately 27 per cent of the value of all incomes earned in the country
in that year. As we saw in Chapters 1 1 and 27, taxes have some effect on
524 THE MARKET ECONOMY AS A WHOLE
pnccs, for reason s%c expect the pattern of taxes to have a substantial
this

effect on the allocation of resources among uses There is much evidence

about the effect of particular taxes on the allocation of resources to particu
lar markets We do not, however, have sufiicient knowledge to assess the
total effect of the whole tax system on the economy, and we arc thus unabfe

to ansvver such questions as *How different would the allocation of re

sources and the distribution of income be if all taxes were removed’

Taxes and the dcstribution or income In order to discuss the effect

of taxes on distnbution we need to define three types of taxes
A tax IS called proportional if u tikes amounts of money from people
in direct proportion to their income A tax is called regressive if it taka
a larger percentage of income from people the lower their income A tax
IS called progressive if a tikes a larger percentage of income from people
the larger their income
The words proportional, progressive and regressive arc treated as positive
terms Nothing m the definition implies that one is better than another It

IS usually not too dilTicuIc to classify individuil taxes as proportional, pro

grcssive or regrcssise
\Vhen dealing with rates of tix,wc need to distinguish between vvhat is
called the average rate op tax and the marginal rate of tax
Referring to the income tax the average rale of iix paid by some individual
IS his total income tax divided bv his total income fits marginal rate of tax
IS the rale he vvould pay on another unit of income To take an example,
m income of ;fl 000 per year would pay about
a single person in Bntiin wiih
4 I 86 miking in avenge nic of 186 per cent Ifhis income rises to
in tix

£l 001 he would pay 7s 9d in fix on the extn one pound of income making
a marginal rate of 38 8 per <cnt
Consider first the progresstvity of some individual taxes The income tax
in theUnited Kingdom is steeply progressive as it is in most countries On
the other hand, the heavy uixes on cigarettesand beer tend to be regressive
since the proportion of income spent on cigarettes and beer, and hence the
proportion paid in tobacco and beer tax, tends to decline as income nsa
A modem government will have to use many individual taxes to make up
Itstax system This is because diflcrent taxes have different strengths and
weaknesses and catch different income levels or types of incomes with
varying degrees of efficiency Popular discussion often concentrates on
thcprogrcssmtyorrcgressivityofaparticulartax certain taxes such as the
tobacco tax, arc often condemned because they are regressive, also the
image of the United Kingdom as a steeply progresswe-tax country 1 $ based
1 In other words these commod lies have income claiticilies of demand which are less ll an
un ty
 MICRO-ECONOMIC POLICY 525

solelyon the progressivity of the income tax, without any correction for the
influence of other taxes which are either less progressive or even regressive.
Clearly, if we wish to judge the effect of the tax system on income distribu-
tion we wish to look at the progressivity of the whole tax system. We wish to
ask of a typical household what happens to the proportion of its income
that is absorbed by all taxes as income rises.

Table 38.1 shows the progressivity of the income tax in the United States
and the United Kingdom while Table 38.2 gives the results of a study into
the progressivity of the whole tax system. The contrast between the two
tables is striking.
The tax system distinguishes among households not only according to the
size of household income, but also according to a host of other characteris-
tics, size of family, age, occupation and .source of income. The tax
such as
system undoubtedly does redistribute income among households when they
are classified by some of these other characteristics. Consider two examples.

Table 38.1

AMOUNT OF INCOME TAX TO BE PAID AT DIFFERENT

INCOME LEVELS BY A CHILDLESS COUPLE, 1965

(A) US Federal Tax

!
1

Net income before

exemptions {but Average lax rate, Marginal tax rate

after all Personal percent {

= tax on extra
i

deductions] ' income tax [=(2)-(l)]xl00 !

dollar)

(1) (2) (4) percent

i
;
1 1

Less than $1,200 0 0 0

2,000 t
112 5-6 1
14

3,000 260 8-7 15

;

4,000 !
418 10-5 1
16
'

586 11-7 17
5,000
10,000 1,556 15-6 22

20,000
t

4^044 20-2 28
50,000 16,460 32-9 50
100,000 44,460 1 44-5 60

200,000 109,972
'

55 69

400,000 250,140 62-5 70

670,140 67-0 70
1,000,000
6,970,140 69-7 70
10,000,000
—
526 THE MARKET ECONOMY AS A WHOLE

(B) UK Income Tax

1 Amount of tnrome tax Average lax Marginal tax

Earned income ^
and surtax rate '

rate

(£> (£) (%) (%)

m
coo
\

27
0 0
45
!

1
0
30
1,000 133 13 3 38 8
1,250 208 167 38 8
1,750 359 20 5 1
38 8
3 000 735 24 5 38 8
1
1

7,000 2,294 1

32 8 1

38 8
15,000 7,787 51 9 76 3
j
'

30,000 20,972 ! 69 9 88 8
50,000 38,722 77 4 68 8
100,000 83,097 83 I 88 8
1

By giving deductions for wives and children the tax system tends to redistn-
buie income away from households with only one person and towards
households with many children In the United Kingdom, income arising
from the sale of the services of labour (called earned income] is taxed more
lightly than income arising from the sale of the services of capital (called
unearned income) This tends to be favourable io people whose incomt is
from work and to be relatively less favourable to people whose income is
from investments

The structure of taxes and the ooods the economy produces

Resources move, m a market economy, in response to changes in the ulatiie
prices of different goods and factors, and in response to the relative profit
ability of different industries A neutral taX is defined as one that does
not
change any of the which resources are supposed to respond It is
signals to
of any Western country is not neutral m
perfectly clear that the tax system
this sense It would be surprising if it were The central authorities may

wish to allocate resources m a way different from the way the market alio
cates them, and taxes can be used toward that end
One way m which tax structure affects resource allocation is through the
deductions allowed on the personal income tax An allowable deduction is
 micro-economic policy 527

Table 38.2
EFFECTIV E TAX BURDENS BY INCOME LEVELS^
US

Tax burden as
percentage
Family income level of income
^

Under $2,000 28
2,000-3,999 1 26
4,000-5,999 26
6,000-7,999 26
8,000-9,999 24
10,000-14,999 24
Over 15,000 36

Average for all 27

1 Estimates are for 1958.

Source; Allocation of the Tax Burden by Income
Class (New York; The Tax Founda-
tion, I960).

an expenditure that may be deducted from income before computing the

amount of taxes payable. Consider a wealthy man in the 70 per cent mar-
ginal tax bracket (see Table 38.1). Every ;fl00 he spends on an item that is
tax deductible costs him ^^30 in after-tax income. Every 100 he spends on
items that are not tax deductible costs him 100 in after-tsix income. In
many countries contributions to charitable and educational institutions are
tax deductible, as are interest payments on a mortgage, while contributions
to political parties are not, nor is rent paid on a home. The effect of these
laws is to restructure incentives by making the relative cost of giving to educa-
tional institutions less than the cost of giving to political parties, and the
‘price’ of owning a home relatively less than the price of renting one. If our
theory of the effect of changes in relative prices on demand is correct, people
should be induced to contribute more to charities and less to political

parties than they would in a free-market situation; similarly, they should

be motivated to invest more money in owner-occupied housing and less in
rented housing, and so on. Interest paid on bank loans is an allowable
deduction from United Kingdom income taxes. This means that a rich
man in the 70 per cent tax bracket who borrorvs money from the bank can
528 THE MARKET ECONOMY AS A WHOLE

deduct this from his taxable income and thus save 70 per cent of the interest
payments which would otherwise have gone m
taxes If the bank charges
him 7 per cent for a loan he will m
fact only have to pay 2 1 per cent for it,
the other 4 9 per cent wiU be paid by the exchequer m
terms of foregone
taxes The increased incentive to borrow is obvious If the intention of the
policy IS to shift incentives in this way, our theory predicts that it will be
successful
Corporate tax non-neutralities work out in a similar fashion The ‘invest-
ment credit’ found m
various forms in both the UK
and the US provides
special tax relief for firms that arc engaging m new investment In the
United States, income received from the production of minerals is given
favourable tax treatment, through the so-called depiction provisions which
allow a mineral producer to deduct from his income amounts that often
exceed the costs actually involved Because of this, explorers and producers
of minerals earn greater profits after taxes, and empirical studies leave no
doubt that the depletion provisions have greatly increased investment in

discovering and developing mineral supplies

Excise and sales taxes raise the relative prices of taxed commodities
Commodities such and petrol are taxed
as cigarettes, alcoholic beverages
particularly heavily As we saw in the discussion of tax incidence in
Chapter 11, excise and sales taxes will have some effect on prices Depend-
ing upon the elasticity of demand, this effect will, in turn, create varying
effects on the quantities consumed It is not clear whether taxes of these
kinds are intended to curtail consumption or merely to raise revenue The
more inelastic the demand for a product {tetens panbus), the less the curtail-
ment of consumption and the greater the revenue yield
The quantitative extent to which tax policies that affect the relative
prices of goods in the private sector actually change the pattern of goods
and services produced is not really known It is clear that man> pnees are
higher than they would otherwise be, and that some arc lower On the basis
of our theory, we are led to predict that this leads to less production and
consumption of the former commodities and more of the latter, but the
question of how different is tbc biU of goods, is an empirical question that

requires, but has not received extended study Nor is it clear whether such
effects as have occurred arc intended consequences of a policy designed to
achieve them, or arc incidental side effects of policies designed to achieve
other ends

Public Expenditure as a Tool of Economic Policy

Much government expenditure goes on what is called transfer pay-
ments These are payments made not in return for any goods or services
 MICRO-ECONOMIC POLICY 529

They transfer money from one person or group to another ivuthout adding to
total production. Public welfare, old age,
unemployment and social security
payments are all examples of transfer pa^unents.
Much government expenditure goes on the provision of goods and ser-
many of ivhich are provided free to the households that consume them.
vices
Some of these goods are collective consumption goods such as defence,
police and goods for which the cost of collecting payment is too high such
as urban roads. For these commodities there is no problem of choice be-
tween the public and the private sector; as they are provided by the state
or not at all. Other goods and sendees such as education, hospital and
medical care, rural roads and motonvays could be provided by private
firms, and indeed were in the past and are now in some countries, but are
nonetheless provided free by the state in other countries. Other goods such
as postal and telephone services and the products of nationalised industries
are provided by the state but are sold to consumers on a commercial basis
just as if a private firm were producing them.
A full consideration of the pros and cons of state versus private production
of these goods and services belongs to a more advanced course and we can
only mention one or two of the most important points here.

Collective goods; There is no problem here since private firms cannot

adequately provide them.

Free goods: The considerations involved here are many and complex.
The case for subsidising goods such as medical care and education rests
partly on a divergence between social and private costs and revenues, partly
on compassion and partly on more subtle welfare arguments. There is
nonetheless a very strong argument against the prorision oTfree goods (as
opposed to subsidised goods) and we must now consider it. This case is
based on the theory' of household demand and before reading on the student
should re-read pages 178-86 of Chapter 15.
If a good is provided free and all demand is met, then households will
go on consuming it until the last unit consumed has a zero marginal utility.
As long as extra units consumed have a positive marginal utility (no matter
how small) and a zero marginal cost, households can raise their total utility
by raising their consumption of the commodity. Thus, resources \rill have
to be used up producing units of the commodity which have zero marginal
utility to each and every' household. Since resources are scarce, these re-

sources must be taken from the production of other goods that have positive
marginal utilities for all households (i.e., households would like to have more
of them). To use scarce resources to produce goods with zero or even very’
low marginal utilities when the same resources could produce goods with
530 THE MARKET ECONOMY AS A WHOLE

(i) Providing the good free greatly increases the amount of resources allocated to
producing the good
(ii) Providing the good free increases only slightly the number of resources
allocated to producing the good
Fig 38 I
 MICRO-ECONOMIC POLICY 531

high marginal utilities ensures that all households Mdll have

lower total
than they could have. If a small price were charged for the com-
utilities

modity, its consumption would decline and resources would be freed to

move to uses where their product had a higher marginal utility. ‘
The quantitative extent of this problem depends on the shape of the
marginal utility schedule for the good. Good I in Figure 38.1(i) has a ver\^
flat marginal utility schedule, the difference between the consumption {Oa}
when the market price is charged and consumption (Ob) when the good is
free is very great. Good II has a steep schedule and the consumption at
zero price is not much higher than at the market price. The case against
providing good I free is that this w'ill absorb a great deal of the nation’s
scarce resources in providing units of the commodity which have a very
low marginal The case against providing good II free is the same,
utility.

but quantitatively it is much less serious.

Why then does the state provide so many goods free? We have already
discussed the case of water (see pages 179-81).^ This is a prime example
of a very weak case. Water undoubtedly has a flat marginal utility schedule
of the sort illustrated in Figure 38.1 (i), thus the no-price policy means that
a great deal of the economy’s scarce resources must be committed to pro-
ducing units of water which have a ver)' low marginal utility to the users
(for examples of low-marginal utility' uses, see page 180). Furthermore,
water does not provide a case where there are social gains from encouraging
water consumption beyond what the individual would voluntarily choose.
Indeed, if a commercially profitable market price were charged there is
little reason to believe there would be serious divergencies between social

and private costs and revenues. Here is a case in which it is hard to see a
rationale beltind existing policy.
The and schools is somewhat different. First, there is
case of hospitals
some doubt that many people would waste free hospital care in the way
they are observed to waste free water. Studies that have been made suggest
a low incidence of unnecessary hospitalisation in a free-hospital system.
In the case of education up to the statutory age, consumption is compulsory
in any case.^ Secondly, there is compassion. Necessary medical and

1 This case against free goods can be maintained without having to accept that the proposi-
tion that perfect competition provides an optional allocation of resources has any relevance to
practical policy problems.
2 In many countries a flat charge is levied on domestic water users. The flat charge is

irrelevant from our point of viesv. It is absence of any charge on additional units consumed
that is important.
3 Indeed, there is that the consumption of ‘free education paid for by all
some evidence
taxpayers W'hether or not they are parents of school-age children leads to a lower consumption
of education than would result from education paid for solely by the households that
were

consuming it, and obviously benefiting from it.

532 THE MARKET ECONOMY AS A WHOLE
hospital care can sometimes be very expensive, the annual cost can easily

be in excess of a household's annual income This is dilTercnt from water

where it is not a great burden for a household to pay a commercial rate for
allthewaterthatisneccsstrytoamodcrately civilised life Thus, to charge a
price that covered costs of production would deny medical and educational
services to many * In both these commodities social and private costs and
revenues are thought to diverge substantially Thus, if I choose not to or
am unable to, afford to buy a cure from an infectious disease, the effects arc
not felt by me alone If all children are better educated, not only they and
their parents but everyone gains from any nsc in output that results from
an increase in their labour productivity Thus, there arc good reasons
accepted by everyone for reducing the private cost of these services below
the market rate by means of a subsidy Tlie only argument is whether the
cost should be merely lowered or should actually be reduced to rero Many
people feel that the costs of making agonising choices of a bit more medical
care for one memberof the household or a bit more education for another
arc degrading and far greater than the costs of making the jcrvicc free {the
cost being measured in terms of a few more resources devoted perhaps
unnecessarily to the provision of these services) Others disagree, and the

provision o{ free welfare services is still hotly debated and the full set of

arguments cannot be set out here Enough should have been said to show
that the case for and against providing a commodity free vanes greatly with
the nature of the commodity being considered

Goods sold on the open market The third class of public expenditure
for purposes of purchasing factor services to produce goods and services
concerns nationalised mdustnes Typically these industries produce goods
and sell them to the public just as any other firm docs They do not how
ever, have to seek to maximise profits, and they often seek just to cover
costs Various reasons for nationalising industries have been put forward
and wc can only give very brief mention to these
1 To confiscate the promts for the general public's aelfare instead of for the
capitalists' In so far as nationalised industries are profitable ones and in so

far as they are not any less efficient under nationalisation than m private

hands this » a rational objective Quantitatively, however, it is insignificant

besides such redistributive devices as the progrcssiv c income tax In so far

as unprofitable and declining mdustnes arc nationalised (e g coal rail

,

ways, airlines in Britain) this motive cannot be a dominant one

1 Of course K is possible lo insure against medical expenses in countries wiiho it naiionil sed

medicine But as everyone who isacquainted with aged people on average incomes in the Ln ted
States knows the insurance policies lend to let the client down injusi those disasters m which
they arc most needed When ihn happens the life savings of even the mosi frugal tahoviring
household can be eaten up bv medical and bos|Mt3l bills in an alarmingK '1 ort period < 1 n e 1
 MICRO-ECONOMIC POLICY 533

2 To get more coordination where private costs and benefits do not refiect social
costs and might be the case that a nationalisation of all forms of
benefits. It

transport could produce a single transport policy wherein decisions made

by railways took into account the costs imposed on the road haulage indus-
try and vice versa. Whatever the facts of such interrelationship there can
be little doubt that the nationalised industries in Britain have up to the
present time made little attempt to look beyond their own parochial
boundaries when taking decisions.*

3 To obtain a radically different pattern of production than would obtain under

private enterprise. If this were a goal, the nationalised industries would require
a clear directive as to how their output and pricing policies were to be
determined. In fact they have usually been given the major task of trying to
avoid which requires that they make average revenue equal to
losses,
average In industries presently making losses, the attempt to cut
costs.
losses is profit maximising behawour. Thus behaviour in these industries is
indistinguishable from what the private firm’s behaviour would be. In the
case of a profitable industry, on the other hand, the directive ‘cover costs’
{AR=ATC) will lead to a higher output and a lower price than the direc-
tive ‘maximise profits’ {MR=MC) and the student should not read on until
he has drawn a graph to demonstrate this.

4 To control effectively a natural monopoly. Industries such as the postal,

telegraph, telephone, gas and electricity ineritably are monopolistic. The
alternatives here are public regulation of a privately owned industry' or

public ownership. One of the main arguments for nationalisation is to

secure effective control over such national monopolies. All countries
seem to accept this argument in the case of the post office and many coun-
tries also accept it for the great public utility industries such as gas and
electricity.

5 To and a more dynamic growth policy than under private

get greater efficiency
industry. The relative efficiencies of private versus public production
is still

the subject of heated debate. Although universal agreement is not forth-

coming, I would hazard the conclusion that the experience of nationalised
industries in the United Kingdom suggests that they have
neither been

vastly better nor vastly worse in running their day-to-day affairs than
ivere

their private predecessors.

Coal mining - an example; This view that public control was needed to
private
save an industry from the dead hand of third rate, unenterprising,
of costs and revenues
1 The Beeching report for example seldom goes beyond calculations
in the railway industry- in making its recommendations.
 '

334 THE MARKET ECONOMY AS A WHOLE

owners was very commonly held about the British coal industry m the inter
war period and was undoubtedly a factor leading to its nationalisation m
1946 This view was clearly held by the Commission which reported in
1926 on the state of the coal industry when they said

It would be possible to say without exaggeration of the miners’ leaders that they
were the stupidest men m England, if we had not had frequent occasion lo meet the
owners

On the other hand, Sir Roy Harrod has recently argued that the run
down state of the coal industry m
South Wales and Yorkshire and the
advanced Northamptonshire and Derbyshire represented
state of the pits in
the correct response of the owners to the signals of the market He writes

The mines of Derbyshire and Notnnghamshiie were nch and it was worth
sinking capital in them If similar amounts of capital were not sunk in other parts
of the country, this may not have been because the managements were inefficient
but simply because it was known that they were not worth these expenditures
Economic efficiency docs not consist m
always introducing the most up-to date
equipment that an engineer can think of but rather m the correct adaptation of the
amount of new capital sunk to the earning capacity of the old asset In not intro
ducing new equipment the managements may have been wise, not only from the
point of view of their own interest, but from that of national interest, which requires
the most profitable application of available capital it is right that as much

should be extracted from the inferior mines as can be done by old ^shioned methods
(i e with equipment already installed) and that they should gradually go out of
*
action

Declining industries always present a sorry sight to the observer Revenues

have fallen below total costs, and, as a result, new equipment is not brought
in to replace old equipment as itwean out The average age of equipment
in use thus rises steadily The superhcia! observer seeing the mdustnes
very real plight blame
is likely to it on the antiquated equipment m use,

whereas the antiquated equipment is the effect, not the cause, of the indus
try’s decline To
modernise at high capital costs merely makes the plight
worse since output and costs will nsc in the face of declining demand and
prices The correct response to a steadily declining demand is indeed not to
replace old equipment but to continue to operate what exists as long as «
can cover tts variable costs of production
It would take a major, and carefully planned, econometric study to

determine who is right about the British coal industry, but, whatever the
ments of thu particular case, the general point is an extremely important
1 Quoted David Thompson SaglonJ pi tie TwenMk Crnlurf Pelican Books 1965 p llO
in
But see also L S Amery i reply that the Commission had ignored the very strong claim of the
government to be so considered This is quoted in context of the discussion of government
policy on p 782 of the present book
2 Roy Harrod Thi Bnlish Ecanomy McGraw Hill 1963 54 p
 MICRO-ECONOMIC POLICY 535

one; a declining industry will always display an old age-structure of capital

and thus ‘antiquated’ methods; this is the consequence and not necessarily
the cause of decline/ To instal new plant and equipment in a genuinely
declining industry is to use the nation’s scarce resources
where they will not
lead to large increases in the value of output. Capital resources are scarce: if
investment occurs in mines, there is less for engineering, schools, roads,
computer research and a host of other things. To re-equip a declining
industry which cannot cover its capital costs is to use scarce resources where
by the criterion of the market their product is much less valuable than what
it would be in other industries.

This of course is a very brief discussion of the reasons for having a sector
of nationalised industries. But enough has been said to show that there are
many possible motives for nationalisation and judging from the behaviour
of the industries, some of the most publicised motives cannot have been the
real ones since the action necessary to give effect to them has not occurred.

MICRO-ECONOMIC POLICY: A FINAL WORD

We have in this its length, merely touched on some public
chapter, despite
policies that affectmicro-economic behaviour. The most difficult problem
for the student of the economic system is to maintain his perspective about
the scope of government activity. There are literally tens of thousands of
laws, regulations and policies that affect firms and households as they seek
their own welfare in a complex world. But seek it they do, and with an
enormous amount of discretion about what they do and how they do it. One
become so impressed (or obsessed) with the number of ways in
pitfall is to

which governmental on the individual that one fails to

activity impinges
changes in the signals and the constraints under
see these as relatively small
which the decision-maker operates within a system that basically leaves
him free within that vast area of the economy called the private sector. In
the private sector, the individual chooses his occupation, earns his living,
spends his income and lives his life. In this sector, the firm, too, is formed,
chooses its products, andgrows and sometimes dies. The opposite
lives,

pitfall is to fail to see that the highly significant sums paid by the private

sector to the government as taxes also buy services and goods that add to
the welfare of the individual and his society. To a great extent the public
sector is complementary to the private sector, doing things the private
sector would leave undone, or would do very differently. The public and
private sectors compete, of course, in the sense that both make claims on the
real resources of the economy.
1 Of course an industry' could decline because its orvners refused stupidly to instal new'

equipment even though it would pay to do so.

536 THE MARKET ECONOMY AS A WHOLE

The particular collection of pobcies m

operation at any time is not the
result of a single master plan that specihes precisely where and how the
public shall seek to complement, help along, or interfere with the workings
of the market mechanism Rather, as individual problems arise, the central
authorities attempt to meet them by passing ameliorative legislation These
laws stay on the books, and some of them become obsolete and unenforce
able This is true of systems of law m
general As a result, it is always
possible and often easy to find outrageous examples of inconsistencies and
absurdities m any system Many anomalies exist in our economic policies,
for example, laws designed to support the incomes of small farmers have
produced some vastly wealthy farmers
The free-market system is retained today as a conscious choice Since
advocate doing away with it and they are not
political parties exist that
voted into power, wc can assume that this is because the majority docs not
want to abolish the system By and large it is valued because of its ability
to do much of the job of allocating resources with a reasonable degree of
efficiency, because it provides scope for the exercise of individual initiative

and because works with a degree of impersonality which obviates giving

it

a host of direct orders on production and consumption decisions On the

other hand, if we judge again from actual behaviour the majority of house
holds are not mesmerised by the free market Some of its limitations are

recognised and the public sanctions many interventions by the central

authorities who seek to change things in order to provide a better world in

which to live
CHAPTER 39

FROM MICRO- TO
MACRO-ECONOMICS

The economy of any Western country is a fantastically complex mechanism.

To use the analog)’ of a machine, it is a mechanism tnth over 200 million

moving parts. The parts - the individuals, households, firms, unions, etc.
independence of action. Yet they fit together, and they display reasonably
coordinated behaviour for at least most of the time.
In the preceding sections of this book, we have been looking at these
parts, how the)' behave, and how they interact. The central concern has
been with the forces that make the economy function. In the remaining
sections, we have the identical concern, but our approach is somewhat
different. Let us pause and see where we have been and -ivhere we are going
and why we are changing approach at this time.

THE DISTINCTION BETWEEN MICRO-ECONOMICS AND

MACRO-ECONOMICS
Micro-economics is concerned with a detailed study of the working of indi-
vidual markets and of the relation between these markets. The central prob-
lem is that of the allocation of scarce resources between alternative uses,
which in turn is related to the problem of the determination of prices and
quantities in all the markets of the economy. We now move our attention
away from a study of the interrelations between the various bits of the
economy, and toward a study of the behaviour of a number of broad aggre-
gates and averages, such as the level of total employment and unemploy-
ment, the total quantity of output (possibly divided into the total quantity
of consumption goods and the total quantity of capital goods), and the
 !

FROM MICRO- TO MACRO-ECONOMICS 539

fed, clothed, entertained, healthy and secure. They have, in var^dng

amounts, resources - income, assets, time and energ\' - ivith which to
attempt to satisfy these wants. But, their resources are insufficient to permit
them to satisfy all their needs and desires. They are forced, therefore, to
make choices, and this they do through markets where they are offered
many ways to spend their money, their energy, and their time. The signals
to which the members of households respond are market prices; for each
given set of prices, they mtU make a given set of choices. In so doing, they
also, in the aggregate, affect those prices. These prices serve as signals to
firms of what goods they may profitably pro\ade. Given technology'
and the
cost of factors, firms must choose among the products they might produce,
among the ways of producing them, and among the various quantities (and
qualities) they can supply. In so doing, they affect prices. Firms demand
factors of production, the quantities demanded depending on the output
decisions, which, in turn, depend upon consumers' demands. These derived
demands for factors will in turn affect the prices of labour, managerial skill,
raw materials, buildings, machinery, use of capital, land and all other fac-
tors. The owners of factors (or the possessors of the skills that can provide

the factor services) respond to factor prices and make their choices about
where to offer their services. These choices determine factor supplies. Pay-
ments by firms to factor o%vners provide the owners of the factors wth
incomes. The recipients of these incomes are people who have needs and
desires for goods and sendees. We have now come full circle
. . .

This circular flow of expenditure from firms to households and back to

firms again is a critical feature of the economy.
In the chapters to this point we have had many insights into how the flow
behaves. But we have also neglected certain key issues. We have said noth-
ing about the size of the flow, or of what forces tend to increase it or de-
crease it. We have not examined the conditions under which the flo^v Mali
be constant from year to year or the conditions that %vill make it fluctuate.
The reasons why the flow increases or decreases, remains stable or fluctuates,
constitutesome of the most compelling issues of economic theory, economic

research and economic policy. These are the central issues of macro-
economics.

The Basis of Macro-economics

In answering general questions about economic s>'stems, we are
many
house-
interested perforce in the aggregated behaviour of large groups of
holds and firms. A
group is made up of its components. If we had perfect

knowledge, we could aggregate or disaggregate at will to find out how any

subgroup or supergroup would behave. Because \s'e do not have any such
540 THE MARKET ECONOMY AS A WHOLE
perfect knowledge, tve attempt to side step such aggregation (and dn
aggregation) problems by h)'potbcstsing stable behaviour patterns at the

aggregate level ' \\ e assume, for example, th^t indivadual households vnll
beha%e in such wavs that stable relations will be found to exist between the
total expenditure of all households and the total incomes of aU households
The hypothesis that stable relations exist between macro variables is the

basis of macro economic theory If such stable relations are found to exist,

there will be a large payoff for economic policy The payoff will take the

fonn of being able to influence the size of certain macro v anabics in v.hich

we are interested by changing the size of other macro vanables that can be

influenced by pohev Ue can hope, for example, to influence the over all

level of unemployment in the economy by regulating the over-all size of

the government’s budget deficit or surplus Where such stable macro rela

tions are not found to exist, we shall have to foil back on a more detailed

study of smaller groups of households, firms and markets before we can

'iwde’tt.v.awd eswiv.g,Vv ibvsviv vhe vs iWa to ku/aw wbva.*.
measures will achicv e our goals
A detailed study of all the interrelationships all the markets in the
economy would be, at best, extremely laborious, at worst, it would be
impossible in our present state of knowledge Macro economics is a search
for a short cut Policy-makers find it possible to affect certain aggregates,
bv culling income taxes, for example, they csin increase the total amount
of income left in the hands of households But the behaviour of households

in theaggregate depends upon what each one of millions of households does

It may be that it is necessary to predict the behaviour of individual house
holds before we can predict their aggregate behiviour accurately * It nav
be, on the other hand, that we can predict the aggregate effect with suf
ficicnt accuracy even though we arc unable to make highly detailed
predictions about the difFcrcnces in behaviour between individual units

The Value of the Micro — Macro Oislmction

As IS of any procedure is judged by whether or
usually the case, the value
not works If individuals do behave
It such a way that macro vanables
m
show stable relations with each other, then a theory based on such an
tiypothesis is likely to prove workable and useful If individuals do not so
1 Thus It u to some estent true to »y that iliedi\’ision b^l" tea mjcro- and maero'ftonooiics
arises from our Ignorance and is to be accepted only as a worbng rule that
enables us to nia e
progrew tn understand ng cenain proUeins
of ii
2 This IS groupoTablerconomuts For a very forceful statement
the hypothesis of one
ami
view see GH Orcutt VficroanaKuc Models of the United Slates Economy Need
D-yelopment Amtruait Ennamic Rerw Mav 1962 RtgHt or wrong it ts at ibis stage a
this stage
distinctly minority Mew The discuss on jn Chapter i pages 10-13 is relevant at
 FROM MICRO- TO MACRO -E C ON O M ICS 541

behave, then the theory will soon be refuted when we tr>' to use it to predict
the outcome of various changes in the economy. The test is thus a prag-
matic one: ‘Does the theory work?’ In the longer run, however, we are
seeking a fuller understanding of the behaviour of the economy. This means
that we must try to move constantly between levels of aggregation. We
want to know what a theory that is successful in predicting at one level
of
aggregation implies about behaviour at other levels. In particular, if a
theory contains only macro variables, we want to know what it implies
about the behaviour of individual households and firms. Once we have
done this, we shall know whether or not the theory is consistent with the
observationswe already have about the behaviour of individual units. This
knowledge may lead us to formulate further tests of the theory in cases in
which there are definite implications about behaviour at other levels of
aggregation but no existing observations.

THE PROBLEMS OF MACRO-ECONOMICS

Why has the level of unemployment in the United States since 1947 oscil-
lated between a low of 2-9 per cent of the labour and a high of 6-8 per
force
cent, while the level in Britain has oscillated over the same period between
IT per cent and 2‘6 per cent? What are the effects on unemployment of
changes in taxes and government spending? Is it really possible that a
country can spend itself out of trouble? Why did British prices double
between 1939 and 1949 and why have they continued to increase, but at a
much slower rate, since that time ? Why do some South American countries
experience a more or less permanent inflation, with prices doubling every
few years ? To what extent do the actions of bankers affect the lives of
ordinary citizens ? How and why does the Bank of England seek to control
the actions of British bankers and households? Is poverty and heavy unem-
ployment in the north-east of England, the maritime provinces of Canada,
and the Appalachian area of the United States the inevitable price of
economic progress? Will the economies of Germany and Japan continue to
enjoy rates of economic growth weU in excess of that ruhng in Great Britain
and the United were the conditions of the fifteen years 1950-65
States, or
exceptional ones, not to be repeated? If the experience of the post-war
period does continue into the future, how long will it be before the United
States loses position as the country with the highest average standard of
its

living in the world ? If Germany and Japan do overtake the United States
in their standards of living, will Russia follow as well ? If present growth
rates continue, how long will it be before England looks to a visitor from
France or Germany as backward and depressed as Southern Spain and
Sicily now look to the visitor from Britain ?
542 THE MARKET ECONOMY AS A WHOLE
These questions all come within the
of macro-economics Perhaps
field

the best way of indicating the scope of this branch of economics is to list the
most important sets of problems with which we shall be concerned in the
remainder of this book, and to contrast these where possible with the related
problems dealt with in micro-economics
1 Problms relating to fiatuattons in tfu Iml of rtscourct U3t, particularly
fiucluahons in the leiel of employment of labour In micro economics we take the
total volume of employment as given and consider how it is allocated among
various sectors of the economy
2 Problems relating to fluctuations tn the outrage level of prices ~ problems, that

IS, of inflation and deflation In micro-economics, we take the absolute pnce

level as given and account for the structure of relative prices
3 Problems relating to fluctuations tn the general level of money wages In micro
economics, we are concerned with the relation between wages in different
areas, occupations and industries
4 Problems relating to the alloeation of resources between the production of con

goods on the one hand and Ike productions of capital goods on the other This
turners'

isan allocation problem similar to the one encountered m micro economic

theory The level of aggregauon is, however, different here wc are dealing
with the allocation of resources between two sectors that together account
for the whole economy, whereas in micro economic theory we split up the

economy into many small sectors

5 Problems relating of growth ofproductive capacity These problems
to the rate

have already been considered m

Chapter 21 but they reoccur in the con
text of macro economics
6 Problems concerning the relation between international trade and the levels of

employment, prices, and growth in the economy These are some of the most impor
tant policy problems facing British and European governments today
7 Problems concerning the control by the central mthonties of the level of activity
in the economy, the tools of control beingflscal and monetary policy In Chapter 38
v.e considered government economic policy in relation to the problems
studied in micro economics In subsequent chapters, we consider govern
ment economic policy in relation to the problems studied in macro
economics
 PART 7

THE CIRCULAR FLOW

OF INCOME
CHAPTER 40

THE MODEL OF THE CIRCULAR

FLOW OF INCOME

In our economy, both commodities and factors of production are constantly

being exchanged for money. If one household buys a currently-produced
commodity from another, it hands over money in return. The selling
household has now earned
this money: it is its income. The household
may money, passing it on to another household in ex-
in turn spend the
change for some other currently-produced commodity. The second house-
hold has now earned the money, and it may in turn spend it. And so it goes
on, the money passing from household to household in exchange for com-
modities: each time money changes hands in return for currently-produced
goods and services, it is income for the recipient. The money flows from
hand to hand in much the same way as water flows through a pipe or
electricity through a circuit.

GOODS AND MONEY FLOWS BETWEEN INDIVIDUAL

HOUSEHOLDS
Imagine a simple economy that consists of four households. A, B, C and D,
each of which makes a different commodity. Assume that there is only one
pound note in the economy. Household A buys goods valued at from
B, giving the pound note in return. This pound is income for household B,
and it spends this income by buying a pound’s worth of goods produced by
household C. Household C has now earned a pound that it uses to buy goods
produced by D, and D in turn buys a pound’s tvorth of goods produced by
A. The and money in this economy are illustrated in Figure
flows of goods
40.1. Goods flows are shown with broken lines and money flows with solid
lines. A single pound note has passed round the circle and has been used
four separate transactions. Each household has earned ,^1 as its income

18
546 THE CIRCULAR FLOW OF INCOME

each has spent £\ If these are all the transactions over, say, a week, then
the total income earned by all individuals in our economy over the week is
£‘^ Note that the total income exceeds the quantity of money (jTl) because
the same pound note was used for more than one transaction This illustrates
an important fact total income is not necessarily equal to the total quantity
of money, because a single unit of money can create income each time it
changes hands Thus, if the typical unit of money changes hands more than
once during the period under consideration, income earned will be greater
than the amount of money m
existence When wc refer to the number of
times that money changes hands over some time period, wc speak of the
VELOCITY OF CIRCULATION of moncy *

Fig 40 J Flows of goods and monr

between four households

GOODS AND MONEY FLOWS BETWEEN HOUSEHOLDS AND

FIRMS
In a real economic system, production is organised by firms, each separate
household does not earn its income by making its own goods for sale
Instead, firms make the commodities that households consume Firms pur
chase the services of factors of production from the households that own

V The quanirty of money is a stock while income is a tlow {see Appendix lo Ghipter 2)

A close analogy would be a circular pipe filled with a certain quantity of water which
*Demg forced round fne circuit 'oy a pump Ttoe quantity dS water in 'fne circuit is a VaaV
might for example be 100 gallons But theamountorwaterpassing any one point dunngsome
period of time is a flow it might for example be 400 gallons per hour If the slock of
water

remains unchanged but the punqi is speeded up to that the water moves faster round the
circuit then the flow will be increased to say 600 gallons per hour In this example the stock
of water is analogous to the quantity of money while the flow is analogous to its velocity hence
to income
Since Income is a flow it must have a time dimension For simplicity wc iviH measure ever)

thing m terms of years Thus income means amsai income which is the amount of income
which accrues over one year
 THE MODEL OF THE CIRCULAR FLOW OF INCOME 547

them, paying wages, rent, interest and profits in return and then use these
make commodities that are sold to households. Households earn
factors to
income by selling factor sendees to firms; they spend this income by buying
goods and services from firms. ^
These transactions give rise to a flow of money in one direction, and a
flow of goods and services in the other. Money flows from firms to house-
holds in return for the factor services purchased from households by firms.
Money also flows from households to firms in return for commodities pur-
chased from firms. This flow is illustrated in Figure 40.2.
Money paid for factor services becomes income for households, and the
money paid for commodities becomes the income of firms. Money flows

S
V
/ \
/ \
/ \
/ \
Wages Paymems
Factor Rent for
Services
Goods
Interest goods
Profit purchased /
\
\
\
\
N
N
N
%

Fig 40.2 Real and money flows between firms and households.

around the from firm to household and back again; thus

circuit, passing

we speak of the circular flow of income. Macro-economic theory is concerned

to explain this flow of income between firms and households.
In this chapter, we shall construct a theoretical model of the circular
flow of income in the economy. Like all models, this one will be an abstrac-
tion from, and an extreme simplification of, the actual flow's in the world.
Economists believe, however, that this model retains the most important
elements and abstracts only from the less important ones. The proof of
whether or not it is a useful abstraction depends, of course, not on whether
1 This is usually but not always the case. People in domestic service, for example, earn their
income by selling their services directly to other households.
548 THE CIRCULAR FLOW OF INCOME

or not It appeals to us but on how well it is able to predict the outcome of

those changes in the economy m which we arc interested If the abstraction
ISa poor one, the predictions of the model will tend to be refuted when the)
are confronted with real-world observations If the abstraction is a success
ful one, the predictions of the model will tend to conform with what we
actually observe
In constructing the basic model, we identify, and allow for, the most
important complications In subsequent chapters, we examine the condi
tions for equilibrium of the circular-flow model and consider m some detail
hypotheses about the behaviour of the determinants of the circular flow
Once we have built up our model, we can see how it behaves as various
changes, such as an increased desire to spend on the part of households,
occur This will allow us to predict how corresponding real-world magni
tudes will vary when the change occurs in the world

Stages in Production

We begin by considering a simple economy that consists of three firms and

several households This economy is illustrated m Figure 40 3 The first

Ftg 40 3 Real and money flows between households and firms involved
different stages of production

to
firm, R, uses factors of production that it purchases from the households
produce raw materials that it sells for ,(^100 to firm I Firm 1 uses factor
rvices to work these raw materials into a semi-manufactured state and
 THE MODEL OF THE CIRCULAR FLOW OF INCOME 549

then sells them to firm F for ;^130. Firm F employs factor services to turn
the semi-manufactured goods into a finished state, and then sells the goods
to the households for ;^180.
The total value of all sales made by the three firms comes to ;^410. The
total r’alue of final goods produced and sold to households is, however, only
/^180. How could we arrive at the figure? ^Ve might do so in two ways:
first, by looking at the sales made by firm F to households, and, second,
by
summing by each of the three firms, R, I and F. In our
the values added
example firm R from scratch and produces goods (raw materials)
starts
valued at £100; the firm’s value added is £100. Firm I starts rvnth raw
materials valued at £100 and produces semi-manufactured goods that it
sells for £130. Its value added is £30 because the value of the goods is

increased by £30 as a result of the firm’s actiHties. Firm F purchases the

semi-manufactured goods for £130 and works them into a finished state,
selling them for £180, Firm F’s value added is thus £50.
In general, the value added by the productive activities of a firm is de-
fined as the difference between the value of the goods that it produces and
the value of the materials that it purchases from other firms. When we
speak of the value of total production we shall be speaking of the value of
final goods produced, which is the sum of the values added by each indi-

vidual firm (£180 in the present example). We shall use the term ‘value of
transactions’ to refer to the total value of sales by all firms (£410 in the
present example).
In our model of the circular flow we aggregate transactions beUveen firms
that represent stages of production. The model displays only the output of
goods and ser\aces by firms in general.

FACTOR INCOMES
The money that flows from firms to households pays for the factor ser\’ices
sold by households to firms. Payments for the services of factors such as land
and labour are made soon after the factors are hired. In the case of labour,
for example, payment is customarily made weekly to wage earners and
monthly to salaried employees. The quantity of factor services hired by
firms varies as current production varies.' Thus the incomes of the onmers
of such factors will also vary with production, rising when production rises

1 In the case of fixed factors, these statements are true only if there is a very long time lag.
A great deal of research has been done recently on the reaction of labour inputs, and hence of
labour income, to changes Although it has supported the general statement made
in production.

in the text, the research suggests that theadjustment of labour inputs to a change in output is
spread over a considerable period of time. In the terminology of Chapter 12, we can say' that
it is true as a general rule that L=flQ}, but that it is also true
that L is related to by a long
distributed time lag (L is employment of factors and 12,1^ quantity produced).
548 THE CIRCULAR FLOW OF INCOME

or not It appeals to us hut on how

well it is able to predict the outcome of
those changes m the economywhich we are interested If the abstraction
in

IS a poor one, the predictions of the model will tend to be refuted when they
are confronted with real-world observations If the abstracuon is a success
ful one, the predictions of the model will tend to conform with what we

actually observe
In constructing the basic model, we and allow for, the most
identify,

important complications In subsequent chapters, we examine the condi-

tions for equilibrium of the circular-flow model and consider in some detail
hypotheses about the behaviour of the determinants of the circular flow
Once we have built up our model, we can sec how it behaves as various
changes, such as an increased desire to spend on the part of households,
occur This will allow us to predict how corresponding rcaJ-world magni-
tudes >«1J var) when the changcoccurs in the world

Stages m Production

We begin by considering a simple economy that consists of three firms and

several households This economy is illustrated m
Figure 40 3 The first

Fig 40 3 Real and money flows between households and firms involved in
different stages of production

1,
R, uses factors of production that it purchases from the households to
duce raw materials that it selk for ;(;i00 to firm I Firm I uses factor
services to work these raw materials into a semi-manufactured state and
 THE MODEL OF THE CIRCULAR FLOW OF INCOME 549
then sells them to firm F for ;^^130. Firm F employs factor senfices to turn
the semi-manufactured goods into a finished state, and then
sells the goods
to the households for ;^180.
The total value of all sales made by the three firms comes to ;^410. The
total \nlue of final goods produced and sold to households is, however, only
/;i80. How could we arrive at the figure? We might do so in two ways;
first, by looking at the sales made by firm F to households, and, second, by-
summing by each of the three firms, R, I and F. In our
the values added
example firm R from scratch and produces goods (raw materials)
starts
valued at ;^100; the firm’s value added is ;^100. Firm I starts with raw
materials valued at j^lOO and produces semi-manufactured goods that it
sells for ;0130. Its value added is ;{)30 because the value of the goods is

increased by as a result of the firm’s activities. Firm F purchases the

semi-manufactured goods for £130 and works them into a finished state,
selling them for £180. Firm F’s value added is thus £50.
In general, the value added by the productive activities of a firm is de-
fined as the difference between the value of the goods that it produces and
the value of the materials that it purchases from other firms. When we
speak of the value of total production we shall be speaking of the value of
final goods produced, which is the sum of the values added by each indi-
vidual firm (£180 in the present example). \Ve shall use the term ‘value of
transactions’ to refer to the total value of sales by all firms (£410 in the
present example)
In our model of the circular flow we aggregate transactions bettveen firms
that represent stages of production. The model displays only the output of
goods and services by firms in general.

FACTOR INCOMES
The money that flows from firms to households pays for the factor services
sold by households to firms. Payments for the sertdces of factors such as land

and labour are made soon after the factors are hired. In the case of labour,
for example, payment is customarily made weekly to wage earners and
monthly to salaried employees. The quantity of factor services hired by-
firms varies as current production varies.' Thus the incomes of the o%vners
of such factors will also vary tvith production, rising when production rises
time lag.
I In the case of fi.ved factors, these statements are true only if there is a very long
great deal of research has been done recently on the reaction of labour inputs, and hence
of

labour income, to changes in production. Although it has supported the general statement made
in the text, the research suggests that the adjustment of labour inputs to a change in output is
spread over a considerable period of time. In the terminology of Chapter 12, we
can say that

it is true as a general rule that L — UQ), but that it is also true that L is related to ffby a long
distributed time lag {L is employment of factors and ffis quantity
produced).
550 THE CIRCULAR FLOW OP INCOME

and falling when production falls The households that hire factor services
out to firms receive payments from firms in return for providing the factor
services that help to produce current output
Now, take the goods currently produced by some firm, value them at
current market prices, and subtract the cost ©fall materials and hired factors
that the firm uses The amount left over is the firm’s remuneration for the
use of Its own factors, plus any pure might hate earned The
profits that tt

firm regards this whole sum as profits, but the economist regards much of
It as the cost to the firm of using its own lactors ‘ These ‘profits’ do not, of
course, actually accrue to the firm until the goods are sold Once the profits
do accrue, they are income /or the households that own the firm (i e ,
those
households that hold shares in the firm) Business profits thus differ from
other factor incomes with respect to the time at which they arc earned
Factors that are used in the course of production gam their incomes while
they are producing goods, a nse m production will lead to a nsc in the
incomes of these factors Profits are obtained when goods arc sold a nsc , in
production will not lead immediately to a nse in profits and thus will not
lead to an increase m theincome of households that own firms
When a firm hires factor services, it pays income to the households supply-
ing the servicesA rise in output produces, after only a short lapse of time,
a flow of income actually paid out to households (except in rare cases, such
as bankruptc>) In the case of households that have a claim to the firm’s
profits, the link between changes m output and payments to households is

not so precise Often, firms retain profits that belong to the ownen and use
them for various purposes These unJufrtduUd profits belong to the households
that own the firms, but they are never received by households as income

SOME DEFINITIONS
Beforewe build up our theory of the circular flow we need to be clear on
what we mean by a few terms that we shall use throughout the analysis
We shall confine ourselves to the minimum number that is necessary at this
stage
One of the mam vanables that concerns us m macro-economics is the
amount of unemployment {(/) We define the labour force as the number of
persons willing to work (n), which number is made up of those actually
working (/) and those willing to work but unable to find jobs, i e unem- ,

ployed, (u) Thus n — (+u We define the percentage of the labour force
unemployed as

0 )

^
1 This IS discussed in detail m Chaplm 19 and 35
 THE MODEL OF THE CIRCULAR FLOW OF INCOME 551

and the percentage of the labour force employed as

^

xl00 = 100-Z7 /2)‘

l+u '

The volume
employment depends on the volume of output,^ and we
of
next need to define a number of output concepts. The gross national
product (GNP) is defined as the total value of all goods and services pro-
duced in the economy over some period of time. (We are taking a year as
our unit of measurement). The net national product (NNP) is defined as
the GNP minus an allowance for depreciation. This allowance is an estimate
of the value of capital goods (plant, machiner)', etc.) used up during the
process of production, and it is thus an estimate of how much of total
production is needed just
keep the capital stock intact.
to
Which of these concepts we
use depends on the problem in which we are
interested. Employment, for example, depends on GNP not NNP employ- :

ment is created whenever factors are set to work to make something; it does
not matter what the product is used for, whether it is consumed, added to
the capital stock or used to replace old capital that has been scrapped. It is

also worth noticing that GNP tends to be easier to estimate statistically than
is NNP since in practice
it is very hard to esdmate depreciation reliably.^

Next we need to define inventories. All producers keep in hand stocks of

materials and finished goods. Raw materials are held so that production can
continue smoothly in spite of irregularities in delivery. Stocks of finished
goods are held so that unexpected orders can be met without waidng for
changes in output; and also so that unexpected variations in output wlJ not
immediately cause variations in the fulfilment of orders. Such stocks are
called INVENTORIES. There is a simple relation between production and
sales and the change in inventories: if the rate of production exceeds the
rate of sales, then inventories of unsold goods will be piling up, and we say
that inventories are accumulating', if the rate of production is less than the rate
of sales, then inventories will be running down, and we say that inventories
are decumulating.
Although we are mainly defining terms at this stage, we need to intro-
duce one behavioural hypothesis that allows us to avoid a number of very
difficult problems that belong only in an advanced treatment of the subject

1 Notice that we are here using lower case letters to refer to absolute numbers and upper
case letters to refer to percentages.
2 With the mentioned in note 1 on page 549.
qualifications
3 This is main
because the records we have to go on are the books of firms which are pre-
accountant’s idea of
pared by accountants. As we saw in the Appendbe to Chapter 19, the
depreciation often differs from the economist’s and even this is influenced by what the tax
authorities udll permit the accountant to charge.
 552 THF riRGULAR FI 0\\ OF INtOMF

we assume that the value of goods sold dunng tht >ear is equal to the valm
of goods produced This means that there is no change in business inven-

tories Later we shall relax this assumption

We now take the value of all output (Ghjp) and break it up m the manner
shown in the accompanying Chart in order to relate our output concepts to
some income concepts If we take the GNp and deduct payments made to

(n (111 (III)

Gross minus Cost of factors et/uals Gross

national product hired by the firm business profits

Households that self Distributed profits Undistributed profits

factors services to
firms

To buy capital goods To buy capital goods

to replace old ones that are net additions
that are scrapped to the firms stock
of capital goods
JO J The relation between various concept of national income and product
(Arrow* drnoic flow* of payment*

households m return for factor services we get the gross profits of business
If we count as income all income paid to households for their factor services
(II) and all income earned by firms (III), whether paid out to the house-
holds that Own the firms (V) or retained by the firms (VI) then we have
defined income to be the same as output Thus gross national income (GNI)
 : +

THE MODEL OF THE CIRCULAR FLOW OF INCOME 553

is exactly the same as GNP. In
what follows the term national income is
meant to refer to the GNP and GNI and for brevity it is given the symbol
P.
Net national product (NNP) is shown in the Chart by I minus VII
and
this is the same as net national income which can
be looked at as IV+V +
VIII. (Do not read on until you are sure that I — VII is the same
as IV
V+VIII.)
We are also interested in what is disposable income. This is the
called
amount of income that households actually receive and thus ha\'e available
for spending. In the Chart this is IV plus V minus any income taxes on these
earnings. This determines the amount actually available to households.^

A SIMP LI FI ED MODEL OF THE CIRCULAR FLOW OF INCOME

First consider a simple economy described by the following assumptions:
(1) households spend incomes on the purchase of goods and services
all their

produced by firms; (2) firms keep production exactly equal to sales so that
there is no change in inventories (3) firms pay out to households in wages,
;

interest and profits all the money they receive from the sales of goods and
services. In this economy, the payment to factors is equal to the value of
current production. Total income is equal to disposable income. Since
households spend all their incomes on goods, firms’ incomes will be the
same as households’ incomes. AU the money that is paid out to households
comes back to firms when households spend their incomes. Since production
is equal to current sales, all the money that comes into firms is paid to
households by way of profits, interest, rent and wages. Clearly, this circular

flow of income, once begun, can continue at the same level forever; there is
no reason why it should change. We say that such an economy is in neutral
equilibrium whatever set of flows we establish will persist forever since there
is nothing to cause them to change.
The abovean extremely important result and it is worth illustrating
is

with a numerical example. Suppose current production is ,(^1,000 per week,

of which is a payment for factor sendees hired by firms and ;^350 is

business profits, which includes a compensation to the owners of their firm

for the use of their capital. Households’ incomes will be £1,000 per week.
Households’ expenditure will also be ;(;i,000 per week. This expenditure
becomes the incomes of firms, and this income must therefore be ,((1^000

1 In statistical work the value of income is defined to be identical with the value of output.
This means that income and output both refer to the same thing. If we think of output as
producing a flow of income payments, these two values (output and income) need not be the
same because of leads and lags and because of changes in inventories. Xo problem is caused in
GNI = GNP as long as one is not misled into thinking that this tells us anything about
defining

the world.

18
 554 THE CIRCULAR FLOW OF INCOME

per week Since production is assumed to be at the same rate as sales, ;^1, 000
will continue to accrue to households as their income return for supply- m
ing the necessary factor services
The circularity of this process should be readily apparent It should also
be apparent that nothing m the argument depended on the levels of the
flows example are multiplied by 10 (or by 100, or
If all the figures in this
by any other number), the same argument can be re-applicd The new
flows will continue indefinitely
W'e now take the next step towards developing a theory of what deter-
mines the size of the flow of income First, we define carefully what we
mean by the circular flow the ctTCular flow of income is ofpayments from
the jiow
domestic households to domesticfrms and back again * As long as households spend
all the money on buying goods and services from domestic
that they receive
firms and pay out to domestic households all the money
as long as firms
that they receive, the circular flow can go on unchanged forever All money
received by one group is passed on to the other in return for goods received
or services rendered iVothmg is ever taken away fiom the flow and nothing
IS ever added to it We say that there are neither withdrawals from nor

INJECTIONS into the flow Although this was true of the example considered
above it is not true of the world There are both withdrawals from and
injections into the real circular flow of income and we must now define these
concepts carefully and introduce the most important withdrawals and
injections
A wUhdraual ts any income that is not passed on in the circular jlow Thus, if

households cam income and do not spend it on

domestically produced goods
and income is withdrawn from the circular flow Similarly, if
services, this
firms receive income from the sale of goods, and neither spend this money
on purchasing factor services nor distribute it as profits, they have with-
drawn It from the flow
An injection :j an addition to the income of households that does not arise from the
spending of domestic firms, or an addition to the income offirms that does not arise
from the spending of households An
injection occurs, for example, if firms
borrow money from banks to pay households for the use of factor services
Household income is increased not because households are buying more
goods from firms, but because firms are borrowing money from outside the
circular flow to hire additional factor services
Figure 40 4 shows the circular flow model amended to include injections
and withdrawals Injections are shown as additions to the circular flow of
income and are labelled J ^Vithdrawak arc shown as subtractions from the
circular flow of income and arc labelled W
Withdrawals represent income received by households and not passed
• Domestic is used in distinction to foragn
 the model of the circular flow of income 555
back to domestic firms (upper right-hand arrow) and income received by
firms and not passed back to domestic households (lower left-hand arrow).
If households elect to put some of their incomes aside rather than spend it
all, they withdraw
it from the flow. If firms do not pay out profits
to their
owners but hold the money back for use at some future time, they \vithdraw
this money from the flow.

Injections create income for households that does not arise from
the
spending of domestic firms (upper left-hand arrow) and create income for
firms that does not arise from the spending of domestic households (lower
right-hand arrow).An example of the first sort of injection %vould occur if
some households gained income by selling their services to foreign firms,

Fig 40.4 The circular flow of income

with withdrawals and injections.

for such income from the spending of domestic firms.

clearly does not arise
An example if some firms borrow
of the second sort of injection arises
money from banks and use the money to purchase goods produced by other
firms. In this case, the income of firms goes up without there hawng been
any increase in the spending of households.
In the next two chapters we shall discuss in detail the causes of changes in
the circular flow. In the meantime, however, it should be intuitively
obvious that withdrawals tend to reduce the flow and that injecrions tend
to increase it.^ Thus it would seem reasonable to suppose (what we shall
prove formally in the next chapter) that the flow of income increases or
decreases according as the volume of injections exceeds or falls short of the
volume of withdrawals. If more is being withdrawn than is being injected,
the flow of income around the circuit will be diminishing, whereas, if more
is being injected than is being withdrawn, the flow
around the circuit will

1 To see this in its simplest form go back to the model of page 553, that was in neutral

equilibrium and make one change in the assumptions: households save (keeping the money

under their mattresses) 10% of all income they receive. Now what u-ill happen to the circular
flow?
556 the circular flow op income

be increasing OnI> if the two flows of injections and withdrawals are equal,

will the flow of income around the circuit be unchanged In formal language,
the equilibrium condition for the circular flow of income to be constant
o\ er time is that the volume of injections should equal the volume of with-
drawals
We shall now consider the most important withdrawals from and injec-
tions into the circular flow

Savings and Investment

It IS usual to speak about saving and investment m the same breath since
much investment expenditure is financed by borrowing money that has
been saved by households From our point of view it is very important to
separate the decision to save from the decision to invest

SAVING If households save some of their income, they must be spending

lesson goods and services than they receive as income Savings are a with
dravval from the circular flow Of course, some of the savings may even
tually find their way back into the circular flow as it would for instance if
households lend to firms the money they have saved and firms use it to build
new factories On the other hand, some of the money saved may not be
returned to the circular flow, for example, households could hoard their
savings m tin boxes kept under the bed
Firms as well as households can save money If firms do not distribute all
their profits, the undistributed part represents business savings This
money, too, may find us way back into the circular flow, as it would for
example, if the firms use it to build new factories On the other hand, it

may not be returned to the circular flow, for example, firms may hoard it,

keeping it aside for some anticipated ‘rainy day’

Whatever subsequently happens to the money, the act

of saving withdraws funds from the circular flow of
income. In our model, we thus treat all savings as a
withdrawal.

Investment Investment is the expenditure on goods not for current con-

sumption The principal forms of investment are increases in the inventones
*

and expenditures on captial goods, such as plant and equipment Investment

in capital goods is sometimes broken down mto replacement investment and
new investment (The distinction between net and gross national product rests
upon this division
)
Total investment, or gross investment, as it is often called,
1 This can be a slippery concept
 THE MODEL OF THE CIRCULAR FLOW OF INCOME 557

includes all of these kinds of im’estment. Net investment is gross investment

minus replacement investment.
Money for investment expenditure comes from many sources. Investment
expenditure may be made from borrowed money, from funds currendy
received but not distributed to households or from funds accumulated in
the past.
In all cases, the expenditure of money on investment adds to the revenues
of those firms selling the investment goods and thus in turn adds to the
incomes of households selling factor ser\'ices to these firms. Investment
expenditure represents an addition to the circular floiv of income. It creates
income for households, but, since the goods are sold to other firms and not
to households, it does not arise directly out of the expenditure of households.
If, for example, firms spend funds accumulated in the past in order to build

Fig 40.5 The circular flow of income

with savings and investment.

plant and equipment now, households’ incomes will rise even though there
is no increase in the spending of households.

In the model of income flows, gross investment is re-

garded as an injection, since it adds to the circular

flow of income without there having been any change
in the spending of domestic households on goods and
services.

These flows are illustrated in Figure 40.5. There are two sat-ings with-
drawals, since both firms and households save, but there is only one
invest-
involves the sale
ment injection, since virtually all investment expenditure
of capital goods from one firm to another.

Foreign Trade
produced by
Domestic households may spend part of their income on goods
output households
foreign firms, and domestic firms may sell part of their to
558 THE CIRCULAR FLOW OF INCOME

located in foreign countries The first set of expenditures results in imports

goods and services (to which we assign the symbol ^^) and the second results
m exports (to which we assign the symbol X)

Imports Imports constitute a withdrawal from the circular flow because,

when households spend their incomes on imported goods instead of domestic
make these goods instead
ones, they create incomes for the Foreign firms that
of for domestic firms If, to take an extreme example, households were sud«

denly to decide to spend all of their incomes on imported goods, no money

would be passed back to domestic firms, the income of these firms would
fall to zero, and income paid out to households would fall to zero To take

a less extreme example, assume that Bntish households decide to buy fewer
Austin cars and more Volkswagcns This means that a smaller proportion
of the income received by Bntish households will be passed back to the
Bntish Motor Company and a lai^er proportion will be withdrawn from
the circular flow in the Bntish economy BMC will now earn less mcome
and will hire fewer factors so that incomes ofUK households will as fall

well

Whatever subsequently happens to the money, the act

of importing withdraws funds from the circular flow.
In our model, we treat all expenditure on imports as
a withdrawal

Exports One country’s imports are another country’s exports Exports

constitute an injection into the domesuc arcular flow of income In the
previous example, there was an increase of imports into the UK, but there
was simultaneously an increasein exports of Volkswagens from Germany
This increase in exports means that the German car company gains an
increase in Us income and since more factors will be needed to produce
more Volkswagens, there will also be an increase in the income of German
households This increase in the circular flow in Germany did not ansc
because of any change in the expenditures of German households (it was
the expenditure patterns of UK households that changed), and so it con-
stitutes an injection into the German circular flow of income
Whatever subsequently happens to the money, the act
of exporting injects funds into the circular flow In
our model, we treat all expenditure on exports as an
injection

Figure 40 6 shows imports and exports added to the model of the circular
flow of income
 THE MODEL OF THE CIRCULAR FLOW OF INCOME 559
The Activities of the Central Authorities

The central authorities affect the circular flow both by

irithdraiving income
from it through taxes and by injecting income into it through their
spending.
We must now consider both of these effects. The term government will
be
used interchangeably with the term central authorities throughout
the dis-
cussion.

Taxes: Taxes withdraw money from the circular flow of income in just
the same way as do savings and imports. If the government taxes firms,
part of the money received by firms is not available to be passed on to
households. If the government taxes households, part of the income earned
by households is not available to be passed back to firms. Of course, some of

Fig 40.6 The circular flow of income

with savings and investment and
imports (M) and exports (X).

the tax revenue may find its way back to the circular flow if the government
subsequently spends on commodities purchased from firms or on factor
it

services hired from households. On the other hand, if the government does
not spend the money but merely lets it accumulate as a reser\'e against some
future expected expenditure, it will remain outside of the flow. We assign
the symbol T to the total value of taxes raised in a year.

Whatever subsequently happens to the money, the act

of taxing withdraws funds from the circular flow of
income. In our model, we thus treat all taxes as a
withdrawal.

Expenditure: We may divide the expenditure of the central authorities

into three main groups.
Expenditure to produce goodsto private households. In this aspect of
to be sold
1

be regarded as just another producer

behaviour the central authorities can
households.
buying factor services from households and selling goods to
560 THE CIRCULAR FLOW OF INCOME

This the case, for example, with the Post Office and the nationalised
IS

industries In the model of the circular flow, we shall include government

commercial acmities m which factor services are hired and the resulting
production is sold to households, in the firm sector Thus when we talk of
the activities of firms, we now mean the activities of all bodies public or

parate, that buy factor services from households and earn other income by
selling their output on the free market
2 Expenditure on gifts, grants, etc , made to private households and firms but not

in return services Payments in this category include subsidies

for any goods or
to unemployment payments, old age pensions, social assistance,
firms,
maternity grants and a host of other welfare payments The main charac
teristic from our point of view is that they are not one side of a two sided,

commercial transaction They are not made in return for any services pro
\ ided and they do not serve directly to increase total output Such pay-

ments are called transfer payments We shall follow standard practice

by deducting transfer payments from both our revenue and expenditure
figures for the central authorities Thus when we speak of taxes, revenues
or expenditures of central authonties we are excluding transfer payments
from our calculations
3 Payments for goods and services providedfree A large part of government
expenditure is for things such as roads, education, defence, national health,
monopolies commissions and the salaries of MP s and civil servants, that
are made m return for goods and services that add to current output, but
payments which are provided from tax revenue rather than by selling
for
the things produced on the market These expenditures create income for
firms and households that docs not arise as a direct consequence of the
expenditures of other firms and households and it is thus an injection into
the circular flow of income

In our model, all government expenditure on goods

and services is treated as an injection Such expendi-
ture IS assigned the symbol G

We must now incorporate these various government activities into our

simple model of the circular flow Figure 40 7 represents various govern
mental injections into and withdrawals from the circular flo\\, ignoring
both transfer payments and the activities of nationalised industries
In order to incorporate the role of the government into our general
circular-flow analysis, we shall combine all of these different withdrawals
and injections into aggregates, and merely show G entering as one injec-
tion in each branch of the circular flow, and T as one withdrawal in each
branch
 THE MODEL OF THE CIRCULAR FLOW OF INCOME
561

In Figure 40.8, we summarisethe various elements we have discussed in

this chapter. If you compare this Figure with the simple circular flow
model
in Figure 40.2 (page 547) you will see how far we have gone in this chapter.

A LINK BETWEEN WITHDRAWALS AND INJECTIONS?: ^Vithdrawals and

injections are not completely unrelated to each other. We mentioned above
that much
of the investment expenditure undertaken by firms is financed
by borrowing (and spending) the money saved (i.e., not spent) by house-
holds. Also much of government expenditure is financed by taxes raised
from domestic firms and households. Finally, as we shall see later, imports
and exports are not totally unrelated to each other.

Fig 40.7 The circular flow of income with taxes and government spending.

If carried these relations to an extreme in which investment was

we
always equal to savings, government expenditure was always equal to tax
revenues (i.e., the government budget was always exactly balanced) and
borrosring
imports were always equal to exports (i.e., there was no foreign
irithdrawals and injections
or lending), then, in spite of having introduced
position of neutral
into the circular flow, we would be back
into the
changed there
equilibrium described on page 553. Every time an injection
offsetting change in a withdrawal, and ever^^ time a
would be an exactly
withdrawal changed would be an exactly offsetting change in an
there
injections would have no effect on
injection. In this case withdrawals and
entirely closed and whatever
the circular flow; the whole circuit would be
was withdrawn would be exactly replaced.
5G2 TUr CIRCULAR Flow OF INCOMF

It IS and fundamental h>poihcsii of the tlicory of the circular flow

T basic
of income that ssitl drawals and injections are not closely linkctl at least
o\cr such short periods of time as a few months or yean Specifically

1 If households decide to save more the funds may he idle since there
ISno reason uhy firms siiould decide to make a corresponding increase in
msestment
2 If households should decide to save less there ii no reason sshy firms
should spend less on investment since alternative sources of funds arc
available (at least for some time even if not permanently)

Fg 40 8 full elemenur> motlcl « f ihr cirnibr flow of income with three

vtiildrawah imporii an 1 raves and three injections mvesimeni
savanj^s

expt ns and gf vemment spending

3 Governments can raise or lower one of taxes or expenditure vMthout

changing the other If lax revenue is made to exceed expenditure this
merely means that the government w accumulating funds vshilc if revenues
arc reduced below expenditure the dincccnce can be made up by govern
ment borrowing
4 Exports and imports can dtange independently of cacli other as long
as voluntary international borrowing and lending is occurring or as long
as countries have exchange reserves to accommodate the difierence between
these two magnitudes (This is a more difficult point which cannot be clan
fied until we have studied the theory offoreign trade )
The critical assumption of the theory of national income is that the
 THE MODEL OF THE CIRCULAR FLOW OF INCOME 563

volume of injections is independent of the volume of wnthdrawals. This

means that either one of withdrawals or injections can change without the
other automatically changing. It is this hypothesis that gives rise to the
possibility of changes in the circular flow of income being caused by changes
in either withdrawals or injections. We shall have more to say about the
validity of this hypotheses later in our study of macro-economics.
CHAPTER 41

EQUILIBRIUM IN THE
CIRCULAR FLOW OF INCOME

In Chapter 40 we laid out the basic model of the circular flow of commodi-
ties and income In the present chapter we study the determination of the

equihbnum level of the flow of national income In ihe next chapter we shall
develop some predictions of this simple theory, and then m subsequent
chapters we shall consider in detail theories about the behasiour of each of
the major components of the flow
The formulation of the model of the circular flow in the previous chapter
deals vnth continuous flows and this invites a dynamic analysis of how the
flow behaves from minute to minute, whether in or out of equilibrium
Unfortunately, we cannot have recourse to such a dynamic treatment be-
cause it requires mathematics beyond the range of most beginning students
(and indeed of many professional economists) For this reason we adopt
exactly the same approach that we have used m previous parts of this book
that of comparativt statics We determine the equilibrium level of the flow of
income chapter) and we then introduce certain changes in which we
(in this
are interestedand compare the new equilibrium after the change with the
old eqmlibnum before the change The shift m
the equilibrium values must
be due to the change introduced, and we then predict that similar results
will be observed in the world Thus if we find that a rise in tax rates will,
teUns pajxbns, lower the equilibnum level of income and employment wc
will predict that, in the real world, a nse in tax rates will tend to be
accompanied by a fall in income and employment This method of com-
parative statics was used extensively in our study of micro-economic theory
and the student who feels at all uneasy about it should now re-read pages
152-154
At the beginning of our study wc shall make some simplifying assump-
tions about the behaviour of output and prices Later we shall be able to
 EQUILIBRIUM IN THE CIRCULAR FLOW OF INCOME
565
work ivithout these assumptions, but in the meantime
they allow us to study
the behaviour of the circular flow of national income
within a simple and
yet empirically relevant framework.

SUPPLY'AND PRICES: SOME SIMPLIFYING ASSUMPTIONS

We shall assume, first, that there exist unemployed supplies of all factors
of production including labour and capital equipment. In chapter 51 we
remove this assumption in order to deal vath situations of full employment
and inflation. Second, we assume that the prices of all goods and ser\'ices
are constant and that changes in the demand for goods are met by changes
in output. These two assumptions mean that the level of output and
employment is determined by aggregate expenditure. Anything that in-
creases aggregate expenditure will increase output and employment, and
anything that lowers aggregate expenditure will lower output and employ-
ment. Because prices are constant any change in the money value of national
income must reflect a real change in output and employment. Third, we
assume that the size of the labour force is constant. A consequence of this
assumption is that (given our first two assumptions] a rise in the expenditure
for goods, that increases output, also increases employment and decreases
unemployment. (This assumption is removed in Chapter 56.)
We shall continue to maintain our assumption that inventories remain
constant. The consequence of this assumption is that the value of output

produced in a given year is the same as the value of output sold in the same
year. We discuss situations where this assumption is inapplicable in
Chapter 45.
Finally, we shall assume that the productive capacity of the economy
remains constant: the value of output attainable when all resources are
fully employed does not change. This assumption rules out the possibility of
economic growth in our model. It does not make the model useless because
—
as long as the rate of economic growth is low - only a few per cent per year
then the assumption wll be approximately true over any short time period,
for, say, a few months to a few years.
To summarise, we assume that we are dealing wdth situations of un-
employment of all resources; that prices do not change so that fluctuations
in the value of output represent fluctuations in quantities
of output, that
passi\ ely
output is demand-determined, by which we mean that it adjusts

to fluctuations in total expenditure in the economy and that employment

;

direcdy with output, whereas unemployment

fluctuates
of labour fluctuates
of the fluctuations in aggregate
inversely with it. Under these assumptions, a theory
level of national income and of
expenditure provides a theory of fluctuations in the
employment.
566 THE CIRCULAR FLOW OF INCOME

It IS to be stressed that in Parts 7 and 8 we are dealing with a theory of

demand We are asking how, with a fixed maximum production capacity,
the output and employment of an economy vanes as the demand for output
varies and also what determines these variations m demand In Part 10 we
drop this assumption of fixed productive capacity and ask what forces
determine the growth of the capacity to produce goods and services

WITHDRAWALS INJECTIONS AND AGGREGATE

EXPENDITURE
In the previous chapter we descnbcd three mam withdrawals and three
mam injections in the circular flow These arc summansed in Table 41 1,
first mwords and then by assigning a symbol to each
Aggregate expenditure is the total of all the expenditure in the economy In
the simplified model that
we arc considering, aggregate expenditure deter
mines the volume of income and employment Expenditure has two main

Table 41 1

WITHDRAWALS FROM AND INJECTIONS

INTO THE CIRCULAR FLOW
Withdrawals = Savings+imports+taxes

S+Af+T)
Injections =* Investmeni+exports+government expenditure

U=J+X+G)

components the expenditure that is injected into the system and the ex-
penditure that arises from withm it because of the spending of domestic
households Total expenditure m the economy is thus total injections plus
and services In Table 41 2
household expenditure on domestically produced goods
the components of aggregate expenditure are summansed
We shall study m detail the behaviour of the components of the circular
flow in subsequent chapters In the meantime we shall introduce some very
simple hypotheses and assumptions that allow us to study the essential ideas
about the determination of national mcome These hypotheses are as
follows eoerylking that originates within the areular Jlow is assumed to inaease as
income increases, while everything that
is injected from without ts assumed to be

unaffected by a change Thus, consumption expenditure and all with

in income
drawals vary directly with mcome, while ail injections remain unchanged
mcome changes Consider these flows one at a time
 EQUILIBRIUM IN THE CIRCULAR FLOW OF INCOME 567

1 Consumption is assumed to rise as income rises; the greater the in-

come of a household, the greater the amount of money it will spend on
purchasing goods produced by firms. Hence the greater is the level of
(national) income the greater is the level of aggregate con-
sumption expenditure.^
2 Household saving is assumed to rise with income: the greater
the
household’s income, the greater the amount of money it saves. Hence the
greater the level of aggregate income the greater the level of aggregate
savings.
Imports are assumed to rise with income: as incomes rise, households
3
spend more on all kinds of commodities, including those produced abroad.
4 Tax revenue is assumed to rise \vith income: with all tax rates held
constant, a rise in household income wall increase the yield of income taxes,
and a rise in spending on domestic consumption will increase the revenue
from excise and sales taxes.
Now consider injections. We hypothesise that exports depend on the
prices of our goods relative to the prices of foreign goods and on the incomes
of foreign households. For the moment, we hold both of these things
constant, so exports must be constant. Government expenditure depends on
decisions of governments, and investment expenditure depends on decisions
of firms. Both of these are important determinants of aggregate expenditure.
We shall have a great deal to say about their behaviour later, but in the
meantime we assume them to be constant. IVe do this not because we
believe that they actually are constants but because we wish to treat them
at the moment as exogenous variables; something, that is determined by
factors outside of, and thus unexplained by, our theory.^ These injections,

1* It is common intreatments to define consumption expenditure as all expenditure

many
on goods and services whether produced at home or imported. tVe have defined consumption
more
expenditure as the purchase of domestically produced goods and services because this fits
developed in
naturally into a behavioural model of the actual flows in the economy of the sort
Chapter 40. Of course the two alternative approaches give the same results when handled

appropriately. In our terms

E = C+I+X. (')

Now subtract imports from both sides of the expression

E-M = C+I+X-M 12)

and rearranging
E= {C-tA1)XI-^{X~

If we let C* be our C plus M and we can svrite

E = C*+I+{X-M),
which is the formulation often used.
2 See pp. 35-6.
568 the CIRGHLAR flow of irSCOME

which do not depend on domestic income and which for the moment are
assumed to be constant, are often referred to as autonomous expendi-
tures They affect the system but are unaffected by it ‘ These assumptions
and hypotheses are summarised m
Table 41 3

Table 41 2
AGGREGATE EXPENDITURE IN THE CIRCULAR FLOW
Aggregate expenditure = Total expenditure on goods and services pro-
duced by the economy
Aggregate expenditure = Expenditure ansmg from within the system
injections of expenditure into the system

{E = C+J)
Aggregate expenditure = Expenditure of households on domestically pro-
duced commodities + investment + exports -f-
government expenditure
(£ = C+/+X+G)

Table 41 3
ASSUMPTIONS AND HYPOTHESES ABOUT THE
BEHAVIOUR OF THE COMPONENTS OF CIRCULAR
FLOW
Behavtour as total

Components income rises

Withdrawals
Savings Rise
Imports Rise
Taxes Rise
Consumption Rises

Injections
Investment Assumed constant
l

(until Chapter 44)

Exports Assumed constant
(until Chapter 46)
Government expenditure Assumed constant
(until Chapter 47)

1 Autonomous has the connotations as exogenous See the Appendix to Chapter 2

’5-6
 equilibrium in the circular flow of income 569
Graphical Representation of Withdrawals, Injections
and Aggregate
Expenditure

Our various assumptions and hypotheses about beha\dour are

plotted on
Figure 41.1. National income is plotted on the horizontal axis and
expendi-
ture on the vertical. The same scale is used on both axes so that an inch along
the horizontal axis indicates the same flow per year as does an inch up the
vertical expenditure axis. Both of these are fiows and are measured as amounts
per year. Thus if we go out
a certain distance on the income axis we are
indicating a certain amount of national income produced each year. Any
point on the graph shows some combination of a flow of income and a flow
of expenditure. For example, the point x shows Ob national income per year
combined with bx of expenditure per year.

Fig 41.1 Injections, consumption

and aggregate expenditure.

The 45° line: If we locate all the points showing annual expenditure
exactly equal to annual income and join them up we shall trace out the line
labelled £=
F in the figure. This line makes an angle of 45° with the F-axis^
and often referred to as the 45° line. Points above or to the left of the
is

45° line show combinations for which expenditure exceeds income; points
below or to the right of the 45° line show combinations for which expendi-
ture than income. Note that we have not yet dated these two flows,
is less

we have not said whether E and F refer to the same or different periods of
time. This critical matter is taken up later in the chapter.

income
Injections: Injections are assumed to be independent of income as :

rises or falls there is no expectation that injections will change. This means
that the annual
that they are shown on a graph as a horizontal line so
volume of injections is the same whatever the level of income. In Figure
41.1

1 Why?
570 THE CIRCULAR FLOW OF INCOME

the annual flow of injections (C?+X+/) is equal to Oa pounds per year The
line relating injections to income is labelled J and it shows that injections
\mI 1 be Oa per year whatever is the level of annual income
The aggregate injection schedule is the sum of the three separate schedules
for investment exports and government expenditure This is shown in
Figure 41 2

Consumption expenditure ‘ This is shown in Figure 41 1 as an upward*

is national income, the larger is con
sloping line, indicating that the larger
sumption expenditure At the low levels of income below Ob, annual
consumption expenditure exceeds annual income This indicates that house-
holds must cither be borrowing money or using up their past savings At
incomes above Ob, annual consumption expenditure is less than annual
income Since the 43° line shows points where expenditure equals income,
and the 45° line shows the
the vertical gap between the consumption line
amount by which domestic consumption expenditure exceeds or falls short
*
of income

Aggregate expenditure In order to represent aggregate expenditure

we must add amount of consumption expenditure that would occur at a
the
given level of income to the amount of injections that would occur at the
same level of income If we do this for each level of income and plot the
points, we shall trace out an aggregate expenditure line showing how total
expenditure vanes as national income vanes Since injections are a constant,
the aggregate-expenditure Imc in Figure 41 1 has the same slope as the
consumption line This indicates that aggregate expenditure vanes with
income on!> insofar as consumption expenditure vanes The aggregate
expenditure line is, however, above the consumption line by the amount of
the injections the vertical distance between E and C is equal to Oa

Withdrawals Wc have defined a withdrawal as ‘any income not passed

on Thus withdrawals are by dejintUon the difference betueen Y which is total

income and C which u the proportion of total income that is passed on {through buying
domestically produced goods) In Figure 41 1 withdrawals at each level of in-
come are represented by the vertical distance between the 45° line and
At income Ob, withdrawals are zero all of income is passed on At

1 Remember that we have defined consumption as expenditures of domestic households on

domestically produced goods
W e have drawn the relation of consumpuon to income as an upward sloping straight line
with a slope of less than 45” The hypotheses that underlie this particular shape are discussed
in Chapter 43
3 Notice that the vertical distance between Cand the 45* line does not represent household
saving but rather the difference between total national income and consumption expenditure
Some of the difference is saved some goes in taxes and some is spent on imports
372 THE CIRCULAR FLOW OF INCOME

higher levels of income withdrawals are positive, and they get larger as
income rises At levels of income lower than Ob, consumption exceeds in
come, and therefore withdrawals arc negative What does this mean'^ It
means that the amount being spent by households on domestically-produced
consumption goods exceeds the whole of national income This means that
the amount being passed on to firms because of household expenditure
exceeds the value of income currently being produced
Because 41 1 is getting crowded, we plot the withdrawal schedule on
Figure 41 3 The scale is exactly the same as in Figure 41 1, and the points
0, a, b, etc ,
are precisely the same

Fig 41 3 The determination of the

equilibrium level of national income
from the withdrawal and injection
schedules

The aggregate withdrawal schedule is the sum of three separate schedules

for savings, imports and taxes Thisis emphasised on Figure 41 2

The withdrawal line rises with income, as we hypothesized it would An

increase in income causes consumption to increase but by less than the
increase m income Thus withdrawals, which are the difference between
income and consumpUon expenditure must rise as income rises The with
drawals schedule shows for any level of income the amount of that income
not passed on through spending on domestically produced goods The
consumption schedule shows the amount of income that is passed on through
spending on domestically produced goods Clearly, the amount that is not
passed on plus the amount that is passed on must equal the whole of income
Therefore, the addition of the W
line to the C line must yield the 45° line,
intficatmg that fhe totaf of the amounts spent pfus the amounts not spent
must account for all of national income When we do this we are back at the
45° line from which we began our discussion

SHIFTS OF CURVES AND MOVEMENTS ALONG CURVES

In Chapter 7 we were concerned to distinguish between movements along
a demand curve and shifts in the curve The former showed changes m the
 A

EaUlLlBRIUM IN THE CIRCULAR FLOW OF INCOME 573

quantity demanded in response to changes
in the prices, while the latter
shoived that there was a change in the
whole relation betiveen quantity
demanded and price (either more or less demanded than
previously at each
pnce). The same distinction is important
ssdth respect to wthdrawal
injection and expenditure curves.
Consider the rarves shown in Figure 41.4. At
income Oa, M-ithdrawals
are ad per year, injections are ac and aggregate
expenditure is <Z(f. This is a
particular magnitude of each of these three flows
associated ivith a particular
national income. If national income rises to Oe per year,
-withdrawals rise to
ef, injections stay constant at eg and aggregate
expenditure rises to eh. In
each case we have a particular rate of injections, isathdraivals and total

Fig 41 Shifts in functions and movements along given functions.

574 THE CIRCULAR FLOW OF INCOME

expenditure associated with a particular level of income We have moved

from one point on a given schedule to another point on the same schedule
Next we can ask to what extent each of these flows responds to a change in
national income On each graph the change in income is at and it is marked
on the graph as AY The change in withdrawals is AIV, in injection is zero
and in aggregate expenditure is AE The ratio of the change in the flow in
which we are interested to the change m
the flow of income is called a
MARGINAL PROPENSITY Algebraically we have the following three
expressions
marginal propensity to withdraw expenditure
from the flow of income =
^
-jtt
^

marginal propensity to inject expenditure

into the flow of income
AY

marginal propensity to spend

’ay
Graphically these arc given by the of the three lines W, J and E, in
slopes

Figure 41 4 The marginal propensity to inject expenditure is zero by

is assumed not to respond
assumption, indicating that injection expenditure
to changes inincome The other two propensities are positive but less than
unity indicating that as income rises both expenditures and withdrawals
rise

In the second column of Figure 4l 4 the schedules have been shifted The
change tn the IV schedule indicates that there is a lower volume of with
drawals associated with each level of national income than before the shift
took place The change in the injection schedule indicates that the \olume
of injections has increased (but that injections still do not vary as income
vanes) These two shifts are independent of each other, a shift m
either the
IV or the J schedule docs not cause a shift in the other schedule A shift m
either schedule does, however, cause a shift in the aggregate expenditure
schedule a downward shift m
the VV schedule causes an upward shift m
the E schedule, while an upward shift in the y schedule also causes an up-
ward shift m
the E schedule In Figure 41 4 the E line is shifted upwards
indicating a nse in the volume of expenditure associated with each level
of national income
It is extremely important to distinguish a movement along one of these
schedules, indicating a change m
the annual rate of withdrawals, injections
or expenditures in response to a change in actual income, from a shift in one
of the schedules indicating a change m
the volume of injections, with-
drawals or total expenditure associated with each level of income
 EQ^UILIBRIUM IN THE CIRCULAR FLOW OF INCOME 575

THE EQUILIBRIUM LEVEL OF INCOME

The may not find this section easy on first reading. It is, however,
student
critical toeverything that follows. He is strongly advised to read through
to the end of the chapter to get an idea of the main drift
of the argument
and then to return to this point and study the material with great care. An
extra hour spent on this chapter now will pay dividends later in
the book.
National income is said to be in equilibrium when it is unchanged
through time. In symbols we may express this as follows

^ t
— ^t + l — ^1 + 2'
'
(1)

where t refers to the first time period, possibly the first week, month or year,
/ + 1 to the next time period and so on.* If national income is not in equi-
librium it will be changing through time, either rising

7, < T, + i
< F,+2--- (2)

or falling:^

Tt > ^1 + 1 > T,+2''’ (3)

In words, the relations in ( 1 ) merely say that national income is the same
in successive time periods, the relations in (2) say that national income is

rising through successive time periods while those in (3) say that income is

fallingthrough successive time periods.

The reason why national income can be at a level other than its equi-
librium one is because adjustments do not occur instantaneously. A rise in
aggregate expenditure leads, after a lapse of time, to a rise in output. After
a further lapse of time, the rise in output places more income in the hands
of households and this after a while leads to a rise in households’ spending.
If all of these adjustments occurred instantaneously, we would always be at
tbe equilibrium level of national income. They do not occur instantaneously
and we describe this by saying that there are lags in the circular flow of
income. In a full theory of the circular flow we would have to allow for all of
these lags, many of which are quite complex. We shall concentrate here on
a very simple treatment designed to reveal the essential ideas of the theory
while concealing many of its most interesting complexities.
We begin by noticing two points about our use of words. First, what is
bought and what is two sides of the same transaction.
sold merely represents
Thus if we look at the value of households’ expenditure on the purchase of
goods and services over some time period t it is the same thing as the value

1 See page 37 on the use of this notation for dating variables.

2 It can also oscillate as we shall see in Chapter 45. In the present treatment we stick to

very simple time-paths of income.

576 THE CIRCULAR FLOW OF INCOME

of firms’ receipts on account of the sale of goods and services over the same
time penod Wc may express this symbolically as follows

E. = 5'. (4)

We have used the term national income todesenbe the total value of all
output m the economy Expressing this m symbols we have

y, = GEP, (5)

where Y is national income and GNP is the total value of all goods and
services produced
So much for our use of terms Next we must recall the assumption made
on page 565 that output remains equal to sales so that there is no change in
inventories We may express this as follows

S, = GNP, (6)

This says nothing more than the total value of sales during some period of
time IS equal to the total value of goods and services produced during that
penod of time
If wc substitute (4) and (5) into (6) wc obtain the following expression

y. = E, (
7)

This tells us that the volume of expenditure is equal to the total volume of
national income in the same period This does not follow from our use of
words but is an implication of our behavioural assumption about
inven tones
Next we need to express symbolically the assumptions wc have already
made about the behaviour of aggregate expenditure We have already
noted that expenditure can be divided into two parts an autonomous com
ponent that docs not vary with income (call this A) and a component that
does vary with income Assuming that expenditure in any period depends
on income earned in the previous period' we can write

E,= A+/{y,i) (
8)
which says, in general terms, that expenditure in one period has an
autonomous that w the same. GxMn.pfxuid in peruidaod a part that
depends upon (is a function of) national income in the previous period
Assuming the form in the graphs of Figure 41 2 we can write

E,=^A + bY,_i (9)

Equation (9) is merely more specific than is equation (8) about the way
1 This makes very good sense if the penods arc short since wages and salanes are almost
always paid either one week or one month in arrears The A includes the constant on the
consumption function
 EQ^UILIBRIUM IN THE CIRCULAR FLOW OF INCOME 577

that expenditure varies with income. Equation

(9) says that if income rises
by some amount, expenditure in the next period will rise by b times
that
amount. We
have already introduced the assumption that a rise in income
causes expenditure to rise but by a smaller amount than the rise in income,
which implies that b is positive but less than unity:
0 < i < 1. (10)
If, for example,were -8 then 80 per cent of any rise in income would be
b
reflected in a rise in spending on domestically produced commodities and
20 per cent by a rise in taxes, imports and savings (withdrawals).
The model that we have introduced can be described by three equations
that we can now draw together:

E, = A + bY,_,, (9)

II
(7)

1 II II
.M (1)

Equation (9) expresses our assumption about the behaviour of aggregate

expenditure. Equation (7) expresses our assumption that output adjusts
instantaneously to sales so that inventories never change. Equation (1)
states the condition for equilibrium that income should not be changing
from one period to the next, the equilibrium value being denoted by 7^.
In order to get the model into a form suitable for graphical presentarion
we can substitute (7) into (1) and reduce it to two relations
£, = A + br,.„ (9)

E, = Y,^,. (
11 )

Equation (9) describes the aggregate expenditure function while equation

(11) gives the condition for national income to be unchanged: that ex-
penditure this period should equal income last period. If this is not the
case, if expenditure falls short or exceeds the prewous level of income, then
that level of income cannot persist unchanged.

An alternative form. We
can also display the equilibrium in terms of with-
drawals and injections. Let us see how this can be done. First look at the
definition of income and expenditure and also at the equilibrium condition
used above:
= c,+w„ (12)

(13)

= r,-i. (11)

The first relation says that all income earned in period i-1 is either spent

in the next period (C,) or not spent [W^]. The second relation says that all

19
578 THE CIRCULAR FLOW OF INCOME

expenditure m period t must ansc from within the system (C,) or from
without (J,)
The third equation is the condition for national income to be
in equilibrium that we have derived above If we substitute (12) and (13)
into (II) we get

C, + IV, = C,+J„ (14)

and cancelling out the C’s we gel

in=y, (15)

Thus another way of stating the condition for national income to be in

equilibrium is to say that the volume of withdrawals from the circular flow
should be exactly equal to the volume of injections
For those who do not like symbolic presentations the above may have
seemed rather heavy going but it is really not mtcllcctually demanding
Trouble can be caused, however, because the form of such arguments is
unfamiliar Everyone should be prepared to work hard on the section and
when It IS re-read after a few more chapters have been studied jt should
seem quite commonplace

A ORAPHiCAL REPRESENTATION To display thc equilibrium of national

income we need only make a slight amendment to the previous graph This
IS done m Figure 41 5 First we date our two variables, plotting income in

one period on one axis, and expenditure in the next period on thc other
Any point such as x now says that if income is Or this penod expenditure
will be Os next period The second line on our graph is the 45® line whose
meaning is also suitably amended « joins up all points for which expendi-
ture in one period is equal to income of the previous penod Thus thc two
lines correspond to the two equations (9) and (11) that we used to describe
our theory in the previous section (sec page 577) The intersection of these
two lines shows the level of income and expenditure which will repeat them-
selves period after period at any other level, expenditure in one period will
,

not equal income m the previous period and income must be changing,
since, m this case, income this penod cannot equal income last period
^

The same thing can be shown on a withdrawals and injections diagram

The equilibrium level of income determined by the intersection of the IV
1* It 19 common m
many elementary and intermediate texts to define the 45° line as an
identity by defining E as identically equal to Y This is totally unsatisfactory The 45° line
must be the expression of an equilibrium candition and if points off it are to have any meaning,
then E cannot be identically equal to F £ and T must refer to difiercni things so that their
have two independent vatiaWes in one model
equality can be a condition of equilibnwn If we
we need two independent conditions todetcrminc their equilibrium values One of these cannot
be an identity
 equilibrium in the circular flow of income 579
and / schedules in Figure 41.6 is, of course, the same as that determined
in the previous diagram.

Behaviour out of equilibrium: We must now see how an economy

described by our simple assumptions would behave out of
equilibrium, and
how it is that it would move back towards the equilibrium level of income.

Fig 41.5 The determination of the equilibrium level of national income using the
aggregate expenditure function.

Consider an economy with an expenditure function illustrated in Figure

41.5 and with current income of Oz. According to our hypothesis about
aggregate expenditure we can see that expenditure in the following period
will be Za. We have assumed that production reacts to changes in expendi-
ture without a time lag so that income earned in period t on account of pro-
duction in that period is exactly equal to expenditure made in that period on
580 THE CIRCULAR FLOW OF INCOME

account of the sale of output ‘

To find the level of income m period t we go
over to the 45° line (by= Za) By the construction of this line, this income is
equal to the level of expenditure (income Oy is equal to expenditure ;'i>)

According to our behavioural hypothesis, expenditure the next period m

will be yc which produces an income in that period of Oai ( = u</=yf) In the
next period expenditure will again fall, this time to ue We can now see by

Fig 41 6 The determination of (he equilibrium level of national income using the
withdrawal and injection functions

inspection that on any point to the right of the equilibrium level of income,
expenditure will continue to fall short of income and that income will thus
be falhng period by period This general proposition is symbolised in
Figure 41 5 by the right-hand arrow indicating a dotvnward pressure on
income at incomes above the equilibrium level
The reader is left for himself the exercise of tracing out the adjustment
when income is below the equilibrium level In this
process case he will see
that the level of expenditure, of income, vnll be rising period by
and level
period This upward pressure of income is symbolised by the left-hand
arrow in Figure 41 5

1* Some behavioural assumptions about whether or not changes in expenditure are foreseen
and how businessmen adjust production to changes m sales are needed if we are to desenbe a
dynamic adjustment Each set of assumptions will lead to a different path towards equilibrium
Our assumption is made so we can see one path, and in the knowledge lhat having studied one
adjustment path we are then going to concentrate mainly on equilibnum conditions But in
spite of appearances to the contrary m many treatments there is no mechanical force ineviubly

pushing income to its equilibrium level We cannot say that income will be pushed to its
equihbnum level until we have made H>me behavioural assumptions about ibe reactions of
firms to disequilibrium situations
 ECiUILIBRIUM IN THE CIRCULAR FLOW OF INCOME 581

Common Sense
Common sense arguments are treacherous. The most intuitively reasonable
hy-potheses are often rejected by scientific observation. We should never let
the burden of any theoretical reasoning fall on common sense. However,
once a result has been obtained by formal methods it often helps to give a
common sense explanation of it.

1 Withdrawals eq^ual injections; If the volume of withdrawals from

the circular flow of income equals the
volume of injections, then income will
remain unchanged. In Figure 41.3, the level of income for which with-
drawals equal injections is Og} If more is being withdra-wn than is being
injected, however, the flow of national income will be declining. If less is
being withdrawn in terms of savings, imports and taxes than is being in-
jected in terms of autonomous expenditure, then income must be rising.
Since -withdrawals rise as income rises and injections are constant, it follows
that at all incomes in excess of Og withdrawals %vill exceed injections and
income must be falling. Thus, the arrow to the right of the point of inter-

section of the W
and J lines in Figure 41 .3 indicates a downward pressure
on national income. By the same argument, at incomes below Og, with-
drawals must fall short of injections and an upward pressure must exist on
national income. The equilibrium level of national income is given by the
intersection of the withdrawals schedule and the injections schedule; at any
other level of income there will be pressure pushing income in the direction
of the equilibrium level.

2 Aggregate expenditure equals income: The second way of look-

ing at equilibrium in the circular flow is to find the level of income for

which aggregate expenditure is equal to the current level of income. If we

think again of the circular flow of income, we kno\v that total expenditure
determines the amount of output, the amount of output determines the
amount of income earned, and the amount of income earned helps to
determine the amount of expenditure. If the amount being spent is equal
to current output, then there is no reason for output, income, or demand to
change. If the amount spent exceeds output and income, then output wll
rise as will incomes earned by households. If, on the other hand, total
expenditure is less than current output and income, then both output and
income earned will faU.
The level of income for which aggregate expenditure equals income is
Og in Figure 41.1. At higher levels of income, aggregate e.xpenditure is less
than current incomes so that a downward pressure is exerted on output and

1 The same result is also apparent in Figure 41 . 1 . At income Og, the distance between C and
the 45° line is exactly equal to the distance between C and E.
582 THE CIRCULAR FLOW OF INCOME

incomes At lower levels of income, aggregate expenditure exceeds current

incomes and an upward pressure is exerted on output and income

The dating of variables It is common practice in economics to lea\c

the \ariables E and Y undated thus implying that current £ is a function
of current Y If the lag in the adjustment o(E to I’ls not a long one rclatne
to the time period co\ered by E and I', this may not cause any serious
errors For instance, if expenditure lags only a week behind income (because
wages are paid one week m
arrears), and if wc are dealing with annual
income and expenditure, then wc will not do too much violence to the facts
if we say that this years’ expenditure is related to this years’ income Since ‘

the lag structures can sometimes be quite complex wc shall follow the exist
ing practice by leaving these magnitudes undated m subsequent chapters
This should be read to mean that expenditures oief an appropriate lime period
are related to current income and not necessarily that current expenditure is

related to current income Indeed it is very important to notice that if we

ate going to explain disequilibrium behaviour tlicrc must be some lags in
the circular flow If everything adjusted instantaneously to everything else
we could never be out of equilibrium and, quite obviously, a model that is
^
based on such assumption can not be used to analyse such behaviour
l If we are dealing with weekly obiervatiocu however thii will not do for if we lay Ihi*
week t expenditure depends on this week s income we will make senoua errori if in fact ic
depends on last weeks income
2* Assume E = J+bY
and Y,^ B,
so that expenditure adjusts instantly to income and income adjusts instantly to changes in
expenditure Then
Y ay+*r.
n-*).
Y.0-b) =7
y * 7/( 1 -*)
This holds for all I The only possible value for F is 7Kt — 4) No equilibrium condition need
be introduced No other value for Y is possible
Many of the errors in elementary introductions to the subject follow from not dating the
variables For example the common proof that savings must in fact be equal to investment
(even though planned S and / may diHer) runs as follows

0 = C+I (!)

r = C+5 (2)

0 = 7 (3)

/=5 (*)
where 0 is current output Two
errors are common Either (3) is introduced as an identity
defining O and Y as equal
which case we have learned nothing about 5 and I in the world
in
we only know O and 7 synonymously we are forced to use S and I synony
that since we use
mously or (3) is meant to be an empirical statement that is always true, this case dating m
reveals that (3) holds only in equilibrium nnee there are leads and lags between output and
various factor payments
CHAPTER 42

SOME PREDICTIONS OF THE

SIMPLE THEORY OF NATIONAL
INCOME

In the last two chapters we have developed the basic model of the circular
flow of income, and have determined the equilibrium of this flow. In the
present chapter we shall use comparative static analysis to derive some very
important predictions from the theory. We noted in the previous chapter
that Ave can express the equilibrium of national income in terms either of
the aggregate expenditure function and the 45° line or of the Avithdrawals
and injections funcdons. We shall use both of these approaches inter-
changeably. The student should always try to demonstrate for himself the
result in question using the approach not adopted in the text.
We are interested in predicting what a« 11 happen to the level of national
income if there is a change in the behaviour of households, firms or the
central authorities. We shall first consider those changes that shift the
injection function and then those that shift the Asithdrawals function.

A SHIFT IN THE INJECTION FUNCTION

What will happen to national income if there is a change in the investment
expenditure of private firms, in governmental expenditure or in exports ?
Fortunately, the same analysis applies to all three changes. In Figure 42.1
an increase in investment, exports or governmental expenditure is shown
by an upward shift in the injection schedule from Ji to Ji- At the original
level of income, Og, injections exceed rvithdraAvals so that income must rise.
As income rises, withdrawals, which are a function of income, also rise, as
shoAvn by the upward slope of the withdrawal function The rise in
income continues until withdraAvals are again equal to the (now higher level
of) injections. This is at income Od in the figure.
584 THE CIRCULAR FLOW OF INCOME

extremely important to remember that we are dealing with con-

It IS

tinuous flows measured as so much per penod of time At the original

equilibrium level, withdrawals and injections were both steady at Oa per
year, injections roseautonomously to Ob per year and as a result income
rose until withdrawals had risen to the rate, Ob per year as well
VVe can analyse a fall in investment, exports, or government spending by
assuming that income starts at Od and that autonomous expenditure then
from the level indicated by J2 {Ob per year) to that indicated by
falls

{Oa per year)

Ftg 42 I The effect on the

equttibnum level of national income
of a change m the level of mjecoons

We have now derived two sets of predictions from our theory

1 A rise in investment expenditure, exports, or

government expenditure, ceteris paribus, will raise
the level of national income.

2 A fall in investment expenditure, exports, or

government expenditure, ceteris paribus, will lower
the level of national income.

A SHIFT IN THE WITHDRAWALS FUNCTION

A downward shift m
the savings,* import or tax schedules will shift the
withdrawals function downwards, indicating that, at each level of income,
the flow of withdrawals will be lower than it was previously A fall in the
withdrawals schedule is shown in Figure 42 2 The original level of income
ISOg with the annual flows of withdrawals and injections are equal to Oa
The withdrawals schedule shifts downwards from IV, to IV2 so that at the
original level of income the new annual flow of withdrawals (gh) is hi less

1 'Remembenng that household! cither tpend ihar income (C) or save it (S), a fall in the
saving schedule means a rise in the consumptum schedule
 :

PREDICTIONS OF THE SIMPLE THEORY OF NATIONAL INCOME 585

than the annual flow of injections. The fact that less is being removed from
the flow than is being injected into it causes the flow to rise until with-
drawals are once again equal to injections. This occurs once income has
risen to Os per year.
An upward shift in the M-ithdrawals schedule is caused by a rise in the
savings/ imports or tax schedules. This is shown in Figure 42.2 by starting
from the schedule W
2 with income Os and letting the withdrawals schedule
rise toW^, so that at the original level of income, withdrawals exceed
Income must fall as a result. As income falls, the
injections (by kl per year).
volume of actual withdrawals falls as well. This continues until income has

Fig 42.2 The effect on the

equilibrium level of national income
of a change in the withdrawals
function.

fallen to Og and wdthdrawals are once again equal to injections at the rate
gi {=ks) per year.
We have now derived two further important sets of predictions

3 A the withdrawals schedule accomplished

fall in
by a fall in tax rates, or the desire to save, or the desire
to import causes a rise in the level of national income.

4 A rise in the withdrawals schedule accomplished

by a rise in tax rates, or the desire to save, or the desire
to import causes a fall in the level of national income.

THE PARADOX OF THRIFT

The above and counter-intuitive applica-
predictions have one interesting
tion. We normally assume, by analogy with an individual household, that

as a fall in the consumption schedule. Sec note i, page 584.

1 Which is the same thing

19 *
586 THE CIRCULAR FLOW OF INCOME

thnft IS is a bad thing, the former leading to

a good thing and prodigality
increased wealth and prosperity, and the latter to eventual bankruptcy
Now what happens if ctitiis panbus all households try
consider, however,
to increase the amount they save at each level of income This increase in
thnftiness shifts the withdrawals function upwards as shown in Figure 42 2
by the shift from Wj to Wi Indeed, as long as income remains at its ongmal
equilibrium level Os the volume of savings is increased (by kl) But now in-
come will begin to fall Indeed, income will fall until total withdrawals have
been reduced to their original level Oa (=:gi=isk) so that they arc again
equal to injections
Now money on purchasing
consider an increase in the desire to spend
goods and services at each level of income This means that less wih be
saved at each level of income and the withdrawals schedule will be shifted
downwards m
Figure 42 2 say from IK, to IV2 But this will mean that income
IS no longer in equilibrium Income will rise until withdrawals arc once
again equal to the unchanged level of injections In this case the downward
shift in the schedule from income from Og to Os
to IKj raises national
We have now derived the predictions of the so called paradox of thnft

5 Ocher thmgs being equal the more frugal and

thrifty are households, the lower will be the level of
national income and total employment. The more
prodigal and spendthrift are households the higher
will be the level of national income and employment.

This striking prediction brings out clearly the basic assumptions on which
our elementary theory is based First, wc are speaking of situations in which
there is some unemployment of all resources (see page 565} so that the level

of output, income and employment depends on the volume of spending in

the economy, anything that increases spending increases income and
employment If we were already at a situation of full employment of
resources then a nse in spending would only cause an inflation since output
cannot be increased, a rise m expenditure can only cause a rise in prices
The second even more critical assumption is that injections are genuinely
autonomous specifically the volume of injections must be independent of
,

the volume of withdrawals If households save more, so goes the theory,

there is no reason why anyone else should decide to increase their spending
This assumption probably would prove to be incorrect, for example, if
firms had been unable to fulfil their investment plans because they could
not get the necessary funds In these circumstances the extra savings made
by households would probably be lent to firms and this would lead directly
to a nse m investment expenditure The most cntical assumption of all the
PREDICTIONS OF THE SIMPLE THEORY OF NATIONAL INCOME 587

theory we have been studyingis that a change in the withdrau'als

schedule
will not cause a change in the injections schedule.
There has been much study of this assumption and currently (1966) there
is a major debate around the investigadons into it by Professor Friedman of
the University of Chicago. It is unlikely that an unequivocal answer will
ever be given that the assumption is always right or always wrong. Suffice
it to say that there are circumstances
- particularly those in which there is a
margin of unemployed resources - in which it is probably correct
substantial
to assume that a change in households savings will not cause a change in
investment expenditure by firms. In these circumstances the prediction
which we have called the paradox of thrift will be confirmed by the evi-
dence. This leads to the policy conclusion that substantial unemployment
is best combattedby encouraging governments, firms and households to be
more intuitively reasonable advice that,
as spendthrift as possible while the
in times of unemployment and depression, frugality and parsimony should
be encouraged will, if followed, only serve to make things worse.
The paradox of thrift was not generally understood during the great
depression of the 1930’s (indeed it is doubtful if it is well understood today
outside of the body of professional economists) and we find, at that time,

many mistaken policies such as the one suggested in the followng passage
from King George V’s message to the House of Commons on 8 September
1931. The message was delivered on the occasion of the formation of a new
national government after the collapse of the Labour administration.^

The present condition of the National finances, in the opinion of His Majesty’s
Ministers, calls for the imposition of additional taxation, and for the effecting of
economies in public expenditure.

At the time the unemployment rate stood at 21 per cent of the labour force
should not read on until you are absolutely sure you know what
You
national income theory predicts to be the outcome of the policies recom-
mended in this Royal message. The suffering and misery of the unhappy
decade of the 1930’s would have been greatly reduced if even a few of the
people in power knew as much economics as is contained in the present
chapter.

COMPENSATING SHIFTS IN INJECTIONS AND WITH-

DRAWALS
The level of national income is in equilibrium when the total volume of
withdrawals equals the total volume of injections. As far as equilibrium in

David Thompson, England in the Twentieth Century, Pelican, 1965, p. 136.

1 Quoted in
588 THE CIRCULAR FLOW OF INCOME

national income concerned, there *s no need for any particular Mth-

is

drawa) to equal any particular injection We may write the condition for
income to be in equilibnum as

it'=y (1)

which was derived on page 578 But specifying withdrawals and injections
in more detail we can rewrite this as

S-\-T-\-M = l+G+X (2)

This condition clearly says nothing about pairs of these terms Specifically
it does not imply that or that T=G or that M^X
This means that
it IS have compensating changes in various withdrawals and
possible to
injections Indeed, if any injection changes, an equal and opposite change
m another injection will leave total injections, and hence total income,
unchanged Also, an equal change m some ivithdrawal, but m the same
direction as the change in injections, will leave income unchanged
Consider, for example, an equilibrium situation in which both the
budget and international payments are m balance This means (?as7',
Af— A, and, since in equilibrium total withdrawals must equal total
injections, it follows that S^I Now, assume a decrease in the desire to
invest on the part of private firms, causing a downward shift in the injection
function, say from Ji to /, m Figure 42 1 {page 584) Our theory predicts
that, alms paubus, this will lead to a fall in income and employment (sec
Prediction 2, page 584) Butother things do not have to remain equal There
arc two possible offsetting policies First, the central authonties can in*
crease governmental spending by exactly the same amount as I has
decreased If in Figure 42 1, C is increased by ab, the amount by which /
fell, total injections will be unchanged, and national income will not depart

from Its full employment level In the new position, income will be un*
changed, but G>T and I<S, so that the fall in investment is exactly
compensated for by a government budget deficit In effect the government
will be spending the amount that private firms no longer wish to spend The

second possible policy is to cut taxes sufficiently to compensate for the fall in
investment expenditure In this case, there is a fall in withdrawals to com-
pensate for the initial fall in injections In Figure 42 1 it will be necessary to
shiftthe IV schedule downwards so that it intersects the new injection
schedule Ji at income Od Again, mcome will not change but the govern-
ment will end up with a budget deficit, this time because its tax revenues
fallwhereas its expenditures remain unchanged These compensating
changes in injections and withdrawals are extremely important, and they
PREDICTIONS OF THE SIMPLE THEORY OF NATIONAL
INCOME 589
form the basis of governmental full-employment policy. We
shall consider
them in detail in Chapter 47. In the meantime, the student should not read
on until he has used the equilibrium condition

S+M+T ^ I+X+G (2)

to show what the central authorities should do to G or to in order to T

offset a rise or a
fall in each of the four other withdrawals and injections.
In
each case, will the government’s offsetting action lead to a budget deficit
{G> T) or to a budget surplus {G<T)1

THE MULTIPLIER
In the previous sections we have considered the effect of shifts in the with-
draivals and injections schedules on national income, concentrating on the
direction of the changes. We now want to look more precisely at the magnitude

of the changes. If the annual flow of investment changes by some amount,

AI, by how much will income change? Such an analysis is important when
considering the effects of changes either in investment or government
expenditure. We might, for example, have a severe depression with the
government being urged to spend money in order
to create employment.
The an increase in government
analysis of the previous section suggests that
expenditure tvill in fact inorease income and employment, but we will wish
to know by how much we can expect employment to increase when govern-
ment expenditure increases. If there is a fall in private investment of some
stated amount, we will wish to know by how much this would reduce
national income and employment.
There are many actual examples of such policies. In 1965 the British
Government placed a six months’ moritorium of expenditure on new invest-
ment projects and it was clearly important to have some idea of how much
this would affect national income and employment. In 1964, in the face of
rates of unemployment around 6 per cent of the labour force, the American
government made a set of tax cuts calculated to reduce the annual flow of
tax revenues by just over S8,000 million. This was predicted to raise
national income and employment in America but it was important to
know by how much. In the event the quantitative prediction made by
American economists proved surprisingly accurate. How did they do this
and what is the basis for such successful prediction ?

A VERBAL statement: Let us put the question in the specific form of what
you would expect to happen to national income if there was an autonomous
590 THE CIRCULAR FLOW OF INCOME

nse m government expenditure on university construction of million

per year with no corresponding nse m
taxes Would national income nse by
only million’ Anyone who has mastered the theory developed so far
should not have too much trouble in replying ‘No, national income will
nse by more than ;^1 million ’ He could argue in one of two ways First, he
could concentrate on the expenditure function and say that a permanent
nse in autonomous expenditure of million per annum would cause
further induced rises in consumption expenditure so that expenditure and
income would in the end nse by much more than the £\ million auto
nomous nse The impact of the mibal nse will be felt by the construction m
dustry and all those industries that supply it Income and the employment
offactorswiUnseby£l million as a result But these newly employed factors
will spend much of their income buying food clothing, shelter, holidays,
cars, refrigerators and a host of other products This is the induced nse in
consumption expenditure, and when output expands to meet this extra
demand, employment will nse m all of these industries When the factors
that are newly employed spend their incomes, demand, output and em-
ployment will rise further, more income will then be created and more
expenditure induced Indeed, at this stage we might begin to wonder if the
increases m income will ever come to an end This question is more easily
answered if we look at the second way m
which an imaginary reader might
argue about the initial question The initial nse in government expenditure
represents a rise in injections This, he would go on to argue, will raise
income, but, as income nses, the volume of withdrawals will nse (tax
receipts, imports and savings will all rise as income rises) Income will con
tinue to nse until additional withdrawals of ;01 million have been generated
At this point, withdrawals will have nsen by as much as the original
(permanent) nse in injections and, assuming we began from a position of
equihbnum, we shall be back in equilibrium Indeed, for example, if
20 per cent of all income is withdrawn through taxes, savings and imports,
then the nse in income will come to a halt when income has risen by £b
million At this new level of income an extra million m withdrawals will
have been generated and since the nse in withdrawals equals the initial nse
m injections income will no longer be rising
Very good, we would say to our imaginary student, your intuition is very
fine but, as we have already mentioned, intuitive arguments do not prove
propositions Can you provide us with a formal proof of these intuitive
arguments ’ At this stage our imaginary dialogue ceases and the teacher
takes over again

An algebraic statement The multiplier is defined as the ratio of the

in income to the permanent change in the flow of expenditure that
 :

PREDICTIONS OF THE SIMPLE THEORY OF NATIONAL INCOME 591

brought it about. The multiplier is given the symbol K and is defined by the
expression

s;

1
( )

If K—2 then any permanent rise in the annual flow of injection expenditure
vvdll lead to a twofold increase in the annual flow of national income.
We must now see on what the value of K depends. In equilibrium we have
W = J, (2)

which says that the volume of withdrawals must equal the volume of in-
jections. Now, if J increases by AJ, then W
must increase by the same
amount in order to re-establish equality (2). The symbol A means ‘a change
in’, so A J refers to a change in injections. Thus we have

AW = AJ. (3)

This says nothing more than that, if withdrawals equal injections originally,
and if injections rise hy A J (say by £\ million per year), then withdrawals
must rise by the same amount in order to restore the equality of and J. W
According to our theory, withdrawals depend on income thus the change ;

in withdrawals depends on the change in income. Thus we may write

AW=wAY. (4)

If, for example, 20 per cent of extra income is not passed on through new

spending, then w=.2, and the change in withdrawals \vill be 20 per cent of
the change in income. Expression (4) merely says that, of any new amount
of income, some fraction will not be passed on in new spending but will

instead be withdrawn.
If we substitute (4) into (3), we obtain

wAY= A J,

or, by dividing both sides by w,

AY=^AJ. (5)
w
Thus, if we write

AY = KAJ, (
1 )

we have
1
K
w
PREDICTIONS OF THE SIMPLE THEORY OF NATIONAL INCOME 593

6 The larger the proportion of income passed on and

the smaller the proportion withdrawn (the smaller
is ip), the greater the fluctuations in income for any
given shift in aggregate expenditure.

7 The value of the multiplier is equal to the re-

ciprocal of the fraction of income not passed on (iv).

If we know, or can estimate, the proportion of any extra income that will
be passed on through spending, we can estimate the magnitude of the effect

Fig 42.4 A graphical demonstration

of the effects of different slopes of the
withdrawals schedule on the value
of the multiplier.

on income of a change in any injection or of a shift in consumption. It was

the estimation of just this magnitude that lay behind the successful pre-
diction of the final effect on US income and employment of the tax cuts
in 1964.
 HOUSEHOLD CONSUMPTION 595

Many factors influence aggregate consumption expenditure: the level of

household income, the household’s stock of wealth, advertising, the ease or
difficult)- of obtaining credit, the outlook for the future,
the weather, and
so on.
Followdng the procedure we have used so often before, we start by focus-
ing on the relation of consumption to one variable in which we are particu-
larly interested. The one we choose
is disposable income. Although

disposable income not the only determinant of consumption, it is shown

is

by empirical studies to be an important one. Changes in other variables can

also have an important influence on consumption. IVe can, however, con-
centrate on the relation betiveen consumption and disposable income and
regard changes in other variables as shifting the simplified function. Recall
in Chapter 7 how we examined a functional relationship beuveen demand
and price rvhile changes in income and other prices were seen to shift the
demand cuix'e.

Consumption as a Function of Disposable Income^

That consumption (C) is a function of disposable income (7^) is an hr-po-

thesis that ever)' study has confirmed. Let us start by stating three more
specific h)'potheses about the nature of the relationship.

Hypotheses.-

1 The greater disposable income, the greater aggregate consumption.

2 Any given rise in income will cause a less than proportionate rise in con-
sumption. To state this another way, any rise in income of X pounds will cause

consumption expenditure to rise by an amount greater than zero but less than X pounds.
3 For very low levels of income, consumption will exceed disposable income (i.e.,
households in aggregate will be going into debt or using up past savings), but beyond
some level income not all of the income received by households will be spent on
of
consumption.

Graphical Representation of the Consumption Function

Figure 43.1(i) and (ii)presents two cur\'es (labelled C), each of which con-

forms to the We
h)-potheses just stated. shall describe the straight line in

(ii), but everything w'c say about it

applies as well to the cur\'e shown in (i).

At low levels of income, consumption expenditure exceeds income. The

income level, Oa in our diagram, at which consumption equals income,

is

above
sometimes called the break-even level of income. As income rises
this

1 See page 553 for a definition of disposable income.

 596 THE CIRCULAR FLOW OF INCOME

break-even level, the consumption function lies below the 45° line, indi

eating that consumption expenditure is less than total disposable income so

that households are saving in the aggregate

The average propensity to consume (apc) APC is the proportion of

total income spent on consumption To calculate this value, we take total
consumption expenditure and divide it by total income ‘
If, for example, out

of an income of million, expenditure is ^^95 million, the average

propensity to consume is 95, if income rises to ;^200 million and consump
tion rises to million, the APC falls to 90 We may divide consumption
expenditure either by total national income, in which case we have the

Expenditure

Fig 43 Alternative consumption functions

(0 The MFC declines as income rises
(ii) The MFC remains constant as income rises

average propensity to consume out of total income, or by disposable income,

m which case we have the average propensity to consume out of disposable
income Which of these concepts is being used ts almost always obvious from
the context

The marginal propensity to consume (mpc) MPC measures the rela-

tionbetween changes m
consumption and changes in income It tells us not
how much of total income is consumed, but how much of the last pound’s
worth of income is consumed ^ Say we start from an income of ^(^1 00 million

1 In symbols APC = -
Y
2 In symbols bfPC
’
= —
AY
 HOUSEHOLD CONSUMPTION 597
with consumption expenditure of ^95 million, and move to
an income of
£110 million with consumption expenditure of
^103 million. The APC
moves from -95 to -936; the MPG is -8 (zl7= 10 and 4C=8). This tells us
that when income is j{;il0 million, just
over 93 per cent of it is spent on
consumption, even though only 80 per cent of the last ^10 million of
income
was devoted to consumption. Note that we need only one value for Tand C
to calculate the APC but that, since the MPC is concerned \vith changes,
we need two values for Y and C to calculate it. The MPC may be stated as

Fig 43.2 A graphical representation of the average and marginal propensities to

consume.

a propensity consume out of either total income or disposable income.

to
Which of these concepts - total income or disposable income - we are deal-
ing with will usually be obvious from the context.
Graphically, the average propensity to consume is indicated by the slope

of the line from the origin to the point in question. This is illustrated in a
numerical example in Figure 43.2. Point a indicates a situation in which
income is 60 and consumption 50, while point j3 shows income 80 and con-
sumption 60. The average propensity to consume at a is 50/60= -833.
Graphically, the APC is the vertical distance m = 50)
(
divided by the
598 THE CIRCULAR FLOW OF INCOME
horizontal distance n (
= 60) which is, of course, the slope of the line be-

tween the origin and a The APC at ^ is 60/80= 75 Graphically this is the
distance p (
= 60)
divided by q (=80) which is the slope of the line between
the ongin and ^ Thus the APC is always shown graphically by the slope
of the line between the ongin and the point under consideration (i c , the
ratio of the distance along the E axis to the distance along the Y axis)
Now consider the marginal propensity toconsume This must be measured
between two points because it is concerned with the ratio of changts m
consumption to changes in income In Figure 43 2 the change in income
between a and p is 20 and the change m
consumption is 10 so that the
MPC = dC/d} = 10/20= 50 Graphically dy IS the distance r (
= 20) while
dCis the distance t ( = 10) so that the MPC=//r= the slope of the line join-
ing a and P In general, the MPC is shown graphically by the slope of the
line joining the two points m question
Thus if we take a consumption function such as that shown m Figure

43 1 (u), the slope of the consumption line indicates the marginal propensity
to consume while the slope of a line from the ongin to any point on the
consumption line represents the average propensity to consume at that
point

Summary The hypotheses stated above may now be stated in terms of this
new marginal and average terminology
1 The MPC exceeds zero at all levels of income
2 The MPC is less than unity for all levels of income
3 The APC -exceeds unity at low levels of income

THE CONSUMPTION FUNCTION FOR THE COMMUNITY

AND FOR THE INDIVIDUAL
An aggregate consumption function shows how the commumty’s total con-
sumption expenditure vanes as its total disposable income vanes Con-
ceptually, the society’s function is aggregated from all the functions of the
individual households that compose it in just the same way as the market
demand curve is aggregated from the demand curves of individual house
bolds
Is thisa valid thing to do^ Gan we really expect to find a stable com-
munity relation emerging when changes m
income affect individual house-
holds m different ways^ What are the conditions under which we shall get
a stable community consumption function’ Basically, there arc two con
ditions, either of which is sufficient The first condition is that all households
should have the same marginal propensity to consume In this case, changes
he distribution of income between households will have no effect on the
 HOUSEHOLD CONSUMPTION 599

level of total consumption expenditure.

If, for example, all households have
an MPG ofthen redistributing a given national income among house-
‘8,

holds will leave the aggregate level of consumption unchanged, because any
household that loses a pound of income cuts its expenditure by 16r, while
any household that gains an extra pound of income raises expenditure by
IGl Thus, in this situation, the level of total consumption depends only on
the level of total income; it is independent of the distribution of this income

among households: to every level of national income there corresponds one

and only one level of aggregate consumption.
If households have different marginal propensities to consume, then
aggregate consumption depends not only on aggregate income but also on
the distribution of this income among households. If, for example, ,^1 is
taken from a household with a marginal propensity to consume of -5, and
given to one with a propensity of -9, then total consumption is increased by
8r, although total income is unchanged. In this case a change in the

of income will cause a change in the aggregate level of con-

distribution

sumption expenditure associated with any given level of national income.

When the MFC’s differ, as among households, a sufficient condition for a
stablecommunity consumption function is that the distribution of income
between households does not change. Thus if national income rises or falls
by 10 per cent, each household’s income will rise or fall by 10 per cent, and
total consumption expenditure will be uniquely related to total income. But
this is a very strong assumption, and a weaker one is sufficient. We do not

need to assume that the distribudon of income is constant, as long as

changes in it are related to changes in national income. If for each level of
income there is only one associated distribution of income, then there will
be only one associated level of total household consumption.
Thus the use of a stable community consumption function that relates
national income to total consumption expenditures, in a world in which
MFC’s are known to vary among households, requires that most changes
that do occur in the distribution of income are themselves associated with
changes in the level of incomes. There is considerable empirical evidence
that the distribution of income changes only slowly, so that it is fairly stable
from one year to the next, and also that such changes as do occur are mainly
related to short-run changes in income. This appears to be the explanation
of whyit has been possible to estimate reasonably
stable aggregate con-

sumption functions.

EMPIRICAL OBSERVATIONS OF CONSUMPTION

FUNCTIONS
Over the last thirty years a vast amount of attention has been devoted to

studying the relation between consumption and income. A detailed survey

600 TKE CIRCULAR FLOW OF INCOME

of this beyond the scope of this book We can, however, mention

work is

some of the most important points

The statistical data available are of two mam types cross-section data and
ime-senes data Cross-section data show at any instant of time how different
households’ expenditures have been related to their incomes Time-senes
data show how consumption and income have changed over time, time-
senes data could be for a single household (i e we could observe a house
,

hold’s income and consumption expenditure each year for, say, 15 years)
or for any aggregation of households Much of the timc-series data we work
with IS aggregate data showing total consumption and total income for the
economy as a whole '
Cross-section data are made available by budget studies of samples of
households Many such studies have been made and they all tend to support
the following general conclusions

1 There is a considerable amount of variation between the consumption

expenditures of different households that cannol be associated with differ

ences in incomes between households Of three households with an income

of :C2i000, the first might have total consumption expenditures of ;C2,300,
the second of j(Jl, 900, and the third of ;(^1, 600 This suggests that, in making
their consumption expenditures, households are not solely influenced by the
size of their incomes
2 On the average, household expenditure docs tend to vary in a
remarkably consistent fashion as income varies We can remove individual
variations by grouping households mto income groups (putting all house-
holds earning an income of less than ^^400 m
the first group, all households
earning between ;^400 and second group, and so on) and
averaging the consumption expenditure for all the households each m
group ^ When this is done, we find that average household expenditure on
consumption goods nses steadily as income rises, but not as fast as income
rises Cross-section data very strongly confirm the hypotheses stated on

page 598, the shape of the curve in Figure 43 I(ii) applies directly to
cross-sectional studies
The data from these studies are valuable in suggesting that income does
have a significant effect on. comuxoptma expendiUixe. Cross-section data
show, however, how consumption expenditure vanes as we move up and
down the income scale at any moment of time They do not neccssanly show
how consumption varies as the incomes of all households change over time
For direct evidence of this we must go to time-senes data
Data for total consumption and expenditure and total income are avail-
1 See the Appendix to Chapter 3 for a ducusuon rf cross section and time senes data Data

about an idenucal group of households over time are called panel data
2 See Chapter 3, page 50 for an example in which this procedure is followed
 HOUSEHOLD CONSUMPTION 601

able for many Western economies starting some time around 1930.
Many
analyses of these data have been made. Perhaps the most significant
con-
clusion's that the consumption function is different according to the length
of time over which data are averaged. The principal findings are sum-
marised below.

Long-run consumption function; decade-to-decade changes;

By using five- to ten-year averages for consumption and income, so that
each observation consists of average consumption over, say, a five-year
period and average disposable income over the same period, we can abstract
from many short-run influences on consumption. Such observations reveal
a close relation between income and expenditure. We find that the marginal
propensity and average propensity to consume are less than unity, but that
the average propensity to consume does not decline as income rises. We also find that
APC and MPC are very nearly equal. ^ These data thus confirm the first

and the third of the hypotheses on page 598, but tend to conflict with the
second. We call this function a ‘long-run’ consumption function. Figure
43.3 shows a stylised picture of one.

Short-run consumption function; year-to-year changes; When

we use annual data (i.e., when each observation consists of total income
over a year and total consumption expenditure for the same year), we find
a close relation between consumption expenditure and income. In general,
years with the highest levels of income tend to be years with the highest
levels of consumption. The observed marginal propensity to consume is less

than unity, but larger than that typically found in cross-section studies. The
average propensity to consume income rises. The three hypo-
declines as
theses tend to be confirmed. Figure 43.3 also shows this consumption
function in a stylised way.
The most significant difference between the long- and the short-run con-
sumption functions is in the reladon between the MPC and the APC. In
the long-run function, MPC equals APC, so that households consume a
constant fraction of their income, whatever the level of their income,
whereas, in the short-run function, MPC
is less than APC, so that house-

holds consume a lower fraction of their income, the higher their income is.

Very short-run variations of consumption with income; ^Vhen we

use quarterly figures (so that each observation consists of data for disposable

1 It is instructive to take a moment to prove that in a consumption function of the form

C=aY, (0<a<l), the MPC and the APC are equal to each other, are less than unity, and
shown graphically as a straight line through
do not change as income changes. This function is

the origin.
 )

602 THE CIRCULAR FLOW OP INCOME

income and consumption over a period of three months), we find that the
association between these two vanables is very much weaker than for any
of the other observations mentioned above This suggests that short run
variations in consumption expenditure are not determined mainly by short-
run variations mincome Why are the hypotheses about the consumption
income relationship refuted under very short-run conditions’ Over the past
decade, many economists have tried to answer this question One of their

functions

findings indicates that it is fruitful to divide consumption expenditure into

expenditure on durables (automobiles, refngerators, television sets, etc ) and
expenditure on nondurablcs (food, entertainment, etc ) Such a division
leads to three observations (1) Much of the short run variability in con-
sumption expenditure is in the durable category (This is not surpnsing
since it IS very easy to postpone durable purchases for a few months

(2) Purchases of durable goods seem to respond to a number of influences,

such as interest rates, size of down payment required, and expected changes
in income that do not influence nondurable purchases (3) The factors just
 HOUSEHOLD CONSUMPTION 603

mentioned seem to be more important than income in explaining short-

term variations in the purchase of durables/

Summary of Empirical Findings

These results suggest that income does exert a significant influence on con-
sumption. At any we are not wasting our time building a theor>' based
rate,
on this assumption. They also suggest that factors other than income
influence consumption expenditure. For purposes of building a theory of
very short-run changes from one quarter to the next, our hypotheses may be
too farremoved from the facts to be useful. For explaining such short-run
changes, more subtle consumption functions may be necessary'. The data
also suggest (although this is still subject to debate) that, in the long run,
income and consumption expenditure may change in proportion to each
other (so that CJY is constant in the long run). They also suggest that
although income fluctuates %\dth the business cycle, variations in income are
accompanied by than proportional variations in consumption (so that
less

CjY varies over the cycle). Thus it may be necessary' to think not of the
consumption function, but of one function suitable for year-to-year varia-
tions in income, of another suitable for decade-to-decade changes, and of
yet another suitable for cross-sectional problems, such as the effect on con-
sumption of a redistribution of income. Figure 43.3 illustrates three
consumption functions: the long-run, the short-run and the cross-sectional.^
In the work that follows, we shall use a single consumption funcdon. The
function we shall use is most appropriate to describing changes in income
from one year to the next. The use of this single function will simplify
matters when we are constructing our theory and will also be most appro-
priate when we use the theory to account for year-to-year changes in national
income and employment.

CONSUMPTION AS A FUNCTION OF DISPOSABLE INCOME

AND TOTAL INCOME
In Chapters 41 and 42 we used a graph relating consumption expenditure
to total national income. In the present chapter we have advanced the

hypothesis that in fixing their consumption expenditure, households react

1 These findings, and others, to put fonvard the
have led Professor Milton Friedman
‘permanent-income hypothesis’. Part of this hypothesis says that short-run behaviour in con-

sumption will be more stable than income, because households, individually and in aggregate,
maintain consumption despite either reverses (such as temporary periods of unemployment)
or windfalls. The relative stabiliq- of consumption can be said to be fairly well established.

Whether Friedman’s theory is the best one (there are others) is still a matter of debate among
economists. ....
2 We have drawn no function for the ‘very shon’ period because the relationship, if it

exists, is so unstable as to be misleading.

604 THE CIRCULAR FLOW OF INCOME

to their disposable income Since disposable income is directly related to

total income, we can use a relation between C and y even though the
behaviour of C is related to as long as Yj is itself related to F ‘ We must
remember, however, that anything that changes the relation between total
income and disposable income will shift the relation between consumption
and total income For example, a cut tn income-tax rates wiU mean that,
for any given level of total income, disposable income will be higher than
before, this means that, for any given level of total income, consumption
^
expenditure will also be higher than before

CAN SHIFTS IN THE CONSUMPTION FUNCTION BE CON-

TROLLED ?
There are many factors that can cause shifts the kinds of consumption m
functions we have shown Some of these factors can be controlled by
go\emment policy and others cannot
Typical of the uncontrollable, and often unforeseeable factors, would be
a sudden shift in the deasion to spend on the part of households Typical of
the controllable factors would be a change in taxes If personal income taxes
are lowered, households’ disposable income will nse even though total
national income is (at Erst) unchanged Consumption expenditure will nse
and income will nse On the other hand, a nse in taxes svill shift the con-
sumption line downward and income will be reduced Tax vanations have
been used as an instrument of policy m both the United States and Great
Britain since the Second World War
1 Consider an vxatnple in which households always spend 60 per cent of their disposable
income and in which the withholding from total income (taxes undistributed profiu etc
amounts to 10 per cent This gives us

Ca 8r< (1)

and r, = 9r (2)

Substituting (2) into ( 1 } gives C as a runction of Y

C= 72Y (3)

2 Assume that m the example offootnotc 1 acutmihctaxratemcreasndisposablemcome

to 95 per cent of total income This gives us the following
(!) it unchanged
c= 8r, (
1
)

but a new equation (2) is needed

Yt = 95r (2a)

Substituting (2a) into (1) gives a new relation between C and Y

C = 16Y (3a)

Consumption has increased without any change in households propensity to consume out of
disposable income
CHAPTER 44

INVESTMENT AND SAVING

In 1929, the investment expenditure of firms and households in the

total
US economy was $16.2 billion. This was almost double the amount of
expenditure needed to replace the capital goods that were used up that year
in the process of production. The US economy, in 1929, was thus adding
rapidly to its stock of new capital equipment. Three years later, in 1932,
totalinvestment expenditure was less than SI billion. This was less than
15 per cent of the amount needed to keep the stock of capital intact. The
US economy in 1932 was thus rapidly reducing its stock of capital equip-
ment. Similar behaviour was observed in the British economy and indeed in
most other economies of the Western world. Such enormous fluctuations in
the volume of investment have not been recorded since that dme, but the
variations have nonetheless been large in some cojuntries. In the UK fluctua-
tions have been rather small since the war but they have been larger in the
US. Investment expenditure fell by 15 per cent in the US between 1957 and
1958; it fell by just over 4 per cent between 1960 and 1961 and it then rose
;

by nearly 15 per cent between 1961 and 1962.

Why do these variations in investment expenditure occur and what are
their effects ?
We saw in Chapter 42 the main short-run effect of these changes. Invest-
ment expenditure is an important component of aggregate expenditure, and
changes in it changes in the level of income and
will cause multiplied
employment. If investment is 15 per cent of total expenditure, a 15 per cent
fall in investment represents a 2^ per cent fall in total expenditure; and

given a multiplier of 2-5, this would reduce income by just over 5} per cent
and raise unemployment from per cent to about 8 per cent of the
(say) 2

labour force. The long-run of changes in investment are felt in terms

effects

of economic growth. These problems are discussed in Chapter 56; nothing

further will be said about them at this point. In this chapter, we concentrate
on the short-term effects of variations in investment expenditure on the level
606 THE CIRCULAR FLOW OF INCOME

of activity and employment in the economy We are still concentrating on

the effects on income and employment of changes in total expenditure

Why do these changes occur ^ This is

a question to answer and
difficult

It takes us into the theory of the determinants of investment, which is the

main subject of this chapter

THE FINANCING OF INVESTMENT

Generally, m speaking of the arcular flow we speak of withdrawals and

injectionsm pairs we speak of imports and exports, of government expendi-
ture and government revenue and of saving and investment This pairing
IS no accident and in the case of saving and investment, it reflects the fact

that the savings of firms and households are the major source of finance for
investment When households and firms save funds, the money must go
somewhere, when firms spend on investment, the money must come from
somewhere Generally, the money spent on current investment projects
comes from the current savings of firms and households
Money required for investment expenditure may be raised in several
ways One is for the firm to save the funds itself (by not distributing all its

profits to Its shareholders) A second way to raise funds is by borrowing the

savings of other units in the economy, mainly households This can be
done either directiy by selling bonds to households, or indirectly by borrow-
ing from financial institutions the money that has been deposited by house-
holds in these institutions If the expenditure on investment is equal to the
money saved by households (and by all other units that save out of current
income), then the funds spent on investment can be raised completely from
funds currently saved with nothing left over
If thevolume of investment expenditure exceeds the volume of funds
currently saved, where does the money come from’ Basically there are two
main sources the money may come from funds accumulated in the past by
firms or households, or it may be money newly created by the banking
system In Chapter 49, wc shall study m
detail how the banking system can
create and destroy money within very wide limits In the meantime, we
must note that, if banks can create money, they can lend this money to
firms for investment expenditure without there being any corresponding
saving of funds on the part of households and firms
There is one more possibility investment may Fall short of savings In
this case all of the money saved cannot be passed back into the circular flow
by way of investment expenditure The excess of savings over investment
will pile up in the form of idle funds owned either by households or firms
and held by them or by financial institatiom on iheir behalf
 INVESTMENT AND SAVING 607

WHAT DETERMINES SAVINGS?

Sa\’ings represent income earned and not spent. There are many motives
for saving . firms may save in order to have funds for new investment or for
replacing their capital stock, households may save for emergencies
or for the
education of their children or to protade life insurance and income
for
retirement. These are important aspects of the behaviour of
individual
units. But, when dealing ivith aggregate income, expenditure and saving, if
we have a theory that succeeds in explaining the portion of income spent,
we shall have an implied theorj' of savings as well. That is, since disposable
income is equal to expenditures plus savings, a theory" that
explains the
proportion of disposable income expended will also explain the proportion
not expended (i.e., saved). For that reason we do not require a separate
theor)' of aggregate saving.
Should we ivish to focus on the
between saving and income, we
relation
may readily do so. For households, for example, we could deduce that
savings are a function of disposable income; that as income rises, savings
rise, but not as rapidly as income; and that the percentage of income saved

rises as income rises. These relations are implied by the hypotheses stated on

page 598, of Chapter 43. Once we have a sa\'ings function, we can define
both an average propensity to save and a marginal propensity to save in just
the same ivay as in the last chapter we defined an average and marginal
propensity to consume. \Ve shall have occasion to use these savings concepts
later.

WHAT DETERMINES INVESTMENT?

What determines the amount of investment? In Chapter 42, we treated
investment as constant and autonomous in order to examine the nature of
the equilibrium of the circular flow. We may now relax this assumption.
We first ask what effect the rate of interest has on the level of investment?
In Chapter 35 we explained the relation between the rate of interest and the
demand to borrow moneyinvestment purposes. (You should review
for
pages 471—5 before reading on.) We now draw, in Figure 44.1, a dowTi-
ward sloping investment schedule plotted against the rate of interest.
According to this figure, the amount of investment rises as the rate of
interest (which represents the cost of borrowing money for investment) falls.
In order to explain some of the variations in the amount of investment, it is
necessary' to explain variations in the rate of interest.

Savings and Investment and the Rate of Interest

The rate of interest is toa great extent controlled by the central authorities,
particularly in Britain by the Bank of England. In Figure 44.1, the rate of
608 THE CIRGULAR FLOW OF INCOME

interest is Ot This gives a rate of investment expenditure of Od per year If

the authorities now force the interest rate down to Oj the rate of investment
expands to Ob This change in the rate of investment expenditure in response
to a change m
the interest rate will be shown as an upward shift in the
aggregate expenditure function in Figure 44 2 Thus interest variation
can cause investment variation of the sort we have considered a previous m
chapter
How does the rate of interest respond to the current flow of savings and
investment^ To see how this might work assume that when investment
exceeds current saving interest rates nsc because there is a shortage of funds
to borrow Would be borrowers will outbid each other to offer favourable

ftg 44 J The relation between the

(low of mvesiment expend lure
and the rale of interest

terms to attract the existing funds to themselves Assume also that when
investment falls short of current savings interest rates will fall because some
savers will be unable to lend their money at all and so will be prepared to
ask lower rates of interest rather than leave their money idle This hypo
thesis makes the interest rate a factor helping to restore equihbnum when
savings and investment are not equal

Interest chances as a factor in restoring equilibrium In Chapter

42 we saw that an autonomous nse m
investment woufef raise income untif
new withdrawals sufficient to restore equilibrium had been created The
magnitude of the nse in income depends on the value of the multiplier

investment raises the rate of interest as

If the rise in
we have just hypothesized, then this will tend to re
store equilibrium by choking off some of the extra
investment expenditure
 INVESTMENT AND SAVING 609
This IS shown in Figure 44.2. Suppose the aggregate expenditure
function
is the one labelled E^. The equilibrium income is Oa. Now investment
increases by AI, thus shifting the aggregate demand function to Ej. If ail of
the adjustment were thrown onto income changes, income
would have to
rise to Ob. But if interest rates rise, and investment
falls, the aggregate
demand downward to, say, £3 and the new equilibrium
function will shift
,

will be reached at the lower level of income Oc. If, when investment in-
creases, the rate of interest rises quickly and this rise in costs greatly reduces
investment, then income need not change much to generate the extra with-
drawals to match the (small) increase in investment. On the other hand, if

Fig 44.2 The effects of shifts in

investment or the aggregate
expenditure function
and on the equilibrium level
of income.

the rate changes only a little or very slowly, and if the change in costs does
not affect investment much, then most of the burden of adjustment is
placed on changes in income.
Let us look at these conclusions in another way. In Figure 44.3, the
current investment schedule is assumed to be /j the current rate of interest
,

is Oa per cent per year, and the current quantity of investment is Ob pounds

per year.
Now assume that the investment schedule shifts to 73 If the rate of interest
.

remains unchanged at Oa, investment rises by be. This amount, be, is the
autonomous J/ of Figure 44.2. If interest rates remain unchanged (or if the
schedule I2 is completely inelastic), then the whole of the burden of adjust-
ment is thrown onto national income, and income will rise by the multipher
process until new withdrawals equal to be are generated. If, however, the
by ee. Investment is now only
interest rate rises to Od, then investment falls
above its original level by be, and the change inincome necessar)^ to restore

20
610 THE CIRCULAR FLOW OF INCOME

equilibnum is correspondingly reduced We now have derived the following

predictions

The change in income in response to an autonomous

change in investment will be smaller, the greater is the
change in interest rates in response to the change in
investment and the more sensitive is the quantity of
investment expenditure to changes in the rate of
interest.

Two THE RELATION BETWEEN INVESTMENT AND THE

LIMITING CASES OF
RATE OF INTEREST Thc View about the working of the economy that pre-
vailed (at least in English speaking economics) prior to the publication of

Fig 41 3 The effect* on the volume of

invmment expenditure of ihifti in the
investment function and of changes
in the rate of interest

Keynes’s General Theory of Employment Interest and Money has come to be called
‘Classical’ Historians are quick to point out that there were many dis-
agreements amongst economists of thc time, and that to talk of 'the Classical
theory’ is rather to caricature a complex situation Nonetheless, there was a
more or less common view on many points and, for better or worse, the

term thc ‘Classical Theory’ or the ‘Classical Model’ has come to be used
to express one consistent version of the views prevalent amongst economists
The present book is not a treatise on thc history of economics
at the time
and we would not bother to describe this ‘Classical’ theory of investment
(and saving) were it not for thc fact that it still is regarded as not refuted, at
least in some of its aspects, by many present day economists

The classical theory Thc basic assumption of the classical theory is

that changes in investment and saving cause changes only in the rate of
interest
 INVESTMENT AND SAVING 611

The classical theory can be summarised as follows;

(1) The desired level
of investment falls as the rate of interest rises. (2) The desired level of savings
rises as the rate of interest rises. (3) The rate of interest changes smoothly
and rapidly in such a way as to keep the volume of investment always equal
to the volume of saving. As long as the rate of interest always keeps
sa\'ings
and investrnent equal, there is no reason for changes in either savings or
investment to cause changes in income. Consider one case by way of ex-
ample. Assume that there is an investment boom with a greatly increased
desire to invest. With the increase in the desire to invest, firms will be trying
to increase their borrowings and they
will quickly bid up the rate of interest.
As the rate of interest rises, the quantity of money firms wish to borrow and
spend on investment falls (since the cost of borrowing rises) and the
quantity households are prepared to save and loan to firms increases. The
rate of interest continues to rise until the diminished investment is exactly
equal to the augmented quantity of saving. The whole process happens
quickly enough so that there no significant rise in income generated during
is

the time in which investment exceeded saving. (The student should not
read on before he has worked out for himself the effects of the other three
shifts; a fall in the desire to invest, and a rise and a fall in the desire to save.)

The classical theory of saving, investment and interest is built on two

important assumptions about the real world the assumption that the invest-
;

ment schedule is sufficiently interest-elastic, so that suitable variations in

the rate of interest can bring about investment sufficient to match any
volume of savings that may be forthcoming, and the assumption that the
rate of interest is perfectly free to vary, so that saving and investment are
quickly brought into equality.
Notice that this theory provides a link between two of the most important
withdrawals and injections; the Classical theory provides an automatic link
between savings and investment. It ensures that, except for temporary
fluctuations, the be equal to the volume of invest-
volume of savings will

ment. Thus as long as the central authorities pursue a balanced budget

policy (G—T), the volume of withdrawals can only differ from the volume
of injections by the difference between imports and exports which will
generally be a trivial amount compared to the whole volume of national
income. Thus a major difference between the Keynesian and the Classical
theory is that, in Keynesian theory, income fluctuates in order to bring
injections and withdrawals into equality while in Classical theory such

fluctuations are not necessary.^

The addition of this equilibrating mechanism to our circular flow model leaves the
level
I

in the example
of income indeterminate in just the same way that income was indeterminate
described on pages 553-554. Clearly the interest rate can equate saving and investnient

when income is low just as well as it can equate these magnitudes when income is high.
612 THE CIRCULAR FLOW OF INCOME

Considerable doubt can be cast on the empirical validity of both of these

assumptions First, most empincal studies that have been made seem to
suggest that variations m interest rates over the range aetually experienced do not
cause \ery large variations m the level of investment Other factors such
as expectationsand the level and rate of change of current demand for
consumption goods, seem to exert a major influence on inv estmeht decisions,
with interest-rate variations having only a small eflect In other words, the
investment schedule m
Figure 44 1 is rather mtercst-melastic
*

Second, there is some doubt whether the rate of interest ts free to vary so
as to equate the flow of current savings with the flow of current investment
The central financial authoniics exert considerable influence on interest
rates, manipulating them in many ways, as wc shall sec in Chapter 49
Among the most important of these ways is the sale and purchase of bonds
on the open market It is quite possible that an increase in saving might
coincide with an attempt by the central authonties to raise interest rates
(lower the price of bonds) by selling bonds on the open market If the sale
of bonds was m an amount equal to the new savings, the entire amount of
new savings would pile up as idle balances held by the Bank of England In
this case, the rate of interest vvould not move to equate current savings with
current investment

The Keynesian theory The theory that the interest rate will fluctuate so
as to equate saving and investment is directly challenged by Keynes’s
theory of interest rates By putting forward a plausible alternative theory of
the determination of interest rates in a free market, Keynes showed, at the
very least, that the classical theory could not be accepted as obviously true

In ch« (hcory she level of nationaJ income wat that which with only lemporary
Classical
aberrations produced full emptoyment Is was believed that if there was unemployed labour

wage would fall and ihe demand for labour increase until full employment prevailed
rates
This labour market mechanism kept income at the full employment level it then did not
matter how much people wished to save at this equilibrium level of income because the interest
rate would fluctuate until firms wished to invest exactly what households wished to save
1 Much of the discussion both tn the Classical literature eg CttieXi Nature arid Ntctsnt;/ if
fnlertsi and in the modem post Keynesian controversies, is concernetS with long tun com
parative static analysis The question piosed in such literature is what would happen to the
demand varying rates of inleresl if alt forces worked themselves out and the
for investment at
economy settled down indefinitely m a completely static position The possible desire to invest
at an interest rate of tay I pec cent under such conditions tells us very little about what would
happen to the desire to invest if under Ihe tmpael national
«f an excess iff samng and a low level of
income and employment interest rates were to fall to 1 per cent for a year or two Yet if we wish
for a theory which is able to handle year to year vanaiions in national income we need a
theory that can handle such short run prohlems rather than one which tells us what should
happen if only (he economy would stay put for a longer time than it ever actually does
 INVESTMENT AND SAVING 613

There were now two contradictor)' theories to choose between, and

the
choice between
them had to be made on the basis of empirical evidence.
The extreme version of the Keynesian th eory was that th e-iii.tgrj?ci-r3 .tp
was completely by the speculative actions of .b ond hnIHprs;; bond -
stabilised
holdersJiayfi.aji_id e_a of the normal rate of interest, and, whenever fluctua -
t ions in current savings and investment cause
even small changes in the
price_o£bonds,J±Ley„would-bu-y_or..sellJrom-their-existing-sE6cteT5f'bonds,
msp reMenting-tfie-actTtal-Tate4rorn-diver.ff-ing far from
tT what they hplip ^rp

If bondholders could prevent the rate of interest from changing rapidly,

then what would restore equilibrium when there was a large shift in either
savings or investment? The equilibrium-restoring mechanism in the
Keynesian theory is income fluctuations of the kind analysed in Chapter 42.
Although the theory that all adjustments take the form of changes in the
level of income is now generally regarded as too extreme, it did focus
attention on what has come to be understood as an empirically ver)' im-
portant mechanism for adjusting savings and investment (even though there
may be others) fluctuations in the level of income and employment are
:

observed to occur in response to shifts in the schedules of saving and invest-

ment.
The balance of empirical evidence seemsto most economists to support a
modified version of Keynes’s According to this theory, fluctuations
theor>'.
in savings or investment cause fluctuations both in interest rates and in
income. Although most economists agree that, as a general rule, more of the
burden of equilibriating the system falls on changes in income rather than
on changes in interest rates, the relative importance of these two mech-
anisms varies between times and places according as the interest rate is free
to change in response to changes in savings and investment and as the
volume of savings and investment reacts to changes in interest rates.

Investment and the Level of Income

Empirical ex'idence tends to indicate that investment plans are more

responsive to the level of demand for goods than to interest rates. When
national income, and hence the demand for consumers’ goods, is high,
businessmen are likely to spend a great deal on investment; when national
income, and hence the demand for goods, is low, businessmen may be un-
willing to spend heavily on investment. This makes investment a function
of the level of national income.'
The theory that investment will be related to the level of national income
1 In symbols, /=I(f ), where I is the volume of current investment and Y is the Itvel of

national income.
614 THE CIRCULAR FLOW OF INCOME

has two aspects the higher the level of demand and income, the more
First,

willing businessmen will be to invest in new tisky enterprises because they

will have favourable expectations about the future Second, the higher the
level of demand, output and hence of profits, the more businessmen wilt

be able to invest This aspect of the theory assumes that most businessmen are
not able to borrow all the funds they require at the current rate of interest,
in fact, so goes the hypothesis, they are severely limited m
the quantity of
funds they can borrow at any moment of time In technical language, they
do not borrow funds in a perfect market As a result of this, the businessman
IS forced to look within his own firm for funds to finance many ofhis desired
investment projects These funds can be obtained by not distributing profits
to shareholders If we now add to this the hypothesis that profits will tend
to be high when demand and income arc high, ue obtain the hypothesis that
*
investment will depend on the level of income
The theory that investment is influenced by the level of profits has been
and bas occasioned much controversy The
subjected to considerable testing
discussionIS complex, and much of it concerns the statistical difficulties m

determining whether the observations conform to the predictions of the

theory One of the main problems is that the observation that investment is
high when profits are high is not necessarily evidence supporting the theory
The causal connection could be m
the opposite direction high investment
causes a high level of income (by the muliipiicr process), which causes high
profits The argument has not yet been settled, and all that wc can say at
this stage is that there is no really compelling evidence to date that would
lead one to reject the theory

Investment and the Rate of Change of Income The Accelerator Theory

According to the ‘accelerator’ theory, it is not so much the level of national

income that affects investment but, rather, the rate of change of national
income The theory is based on thclbllowmg line of argument whenincome
is increasing, it will be necessary to invest in order to increase capacity to
produce consumption goods, investment may also be high because business
expectations based on the rising trend of sales may be favourable On the

other hand when income is falling, it may not be necessary even to replace
capital equipment as it wears out, furthermore, expectations based on the
may be unfavourable For both of these reasons invest
falling trend of sales
ment expenditure need not be great JThuS according to the accelerator
theory, investment is a function of change in income V
The basic idea of the accelerator theory is not a difficult one, it can be
1 In symbols /=f(/l!) where
nands For profits if we add the hypothesis that profits vary
with income /f =
R( F) we obtain which ti the result assumed in the previous footnote
2 In symbols /=I(JF)
 1

INVESTMENT AND SAVING 615

elaborated as follows : when income is constant, it wall not be necessarj' to

expand existing plant and equipment; investment will be limited to re-
placing existing machinery as it wears out (and installing improved
machiner>' for producing the existing level of output, which we ignore
at
this stage for simplicity’s sake) Investment will be constant and
.
will be equal
to that fraction of the existing stock of capital
equipment that wears out and
that needs replacing each year. If, however, income starts to rise, it will be
necessary to invest in new plant and equipment in order to expand the
existing capacity to produce (once the expansion of demand is sufficient to
employ all of the existing capacity). After this point, every additional rise in
income will give rise to additional investment expenditure in order to
increase productive capacity. Further, an increase of ;^1 in income may lead
to an increase of more than £\ in investment expenditure; this is because
a machine has a long life and may have a value considerably in excess of the
value of its output over one year. A machine that has a value of ;^40,000
may produce an annual output of, say, ^^10,000. The entrepreneur who
wishes to increase his output by using these machines must therefore spend
on investment for every that he wishes to add to his annual output.
(This is not necessarily a losing proposition, because he only pays once for
the machine, which goes on producing goods over all the years of its life.)

In our example, once existing capacity is used to its fullest, eveiy ;^1 increase
in income would entail ^4 of additional investment expenditure. In more
general terms, new investment is said to be some muldple, a, of the change
in income. The multiple a is called the accelerator coefficient.^ This may
be expressed in symbols as follows:
/„=ad7 (1)

when I„ is new investment, dF is the change in income, and a is the

accelerator coefficient, which depends on the ratio of the value of capital
equipment to the annual value of its output.
The accelerator theory has two aspects: (1) that the level of new invest-

ment will be a multiple greater than one of the change in income (because
the value of a machine is usually well in excess of the value of its annual pro-
duction), and (2) that the level of new investment varies with the change in
income (because new capital is needed to expand rather than to maintain
output).
1 The basic assumption of the accelerator is that there is an optimal relation between the

stock ofcapital [B) and the level of income, i.e., 5=ar. Now assume we start from equilibrium
relationship to income, i.e.,
so that at period t the actual stock of capital bears the desired
Bj= 0iYi, Now let income rise between period / and /+ 1. If the capital stock is adjusted to its
new desired level over the same period, 5,+ i = aI^+i. Subtracting these ttvo expressions from
each other gives - B, = a(Y,+ - Y,). The change in the capital stock is new investment,
,

B,. and Y,+ — Y, is the change in income. If we denote 7,+,

— Y, as df, this gives
us the relation set out in equation- (1).
 INVESTMENT AND SAVING 617

More subtle tests have found considerable e\'idence of an accelerator-ty'pe

relation, but one complicated by the factors named above. The simple
studies that have failed to find evidence of such a relation are, therefore, not
conclusive. The existing e\adence does point to a fairly strong accelerator-
type relation betiveen investment and changes in demand, but one compli-
cated by the fact that the relation is subject to a distributed lag of quite
long duration.*

Investment and Expectations

Investment takes time. A businessman who decides this year to expand
capacity may not see the fruits of his investment for several years. The
decision to invest now is thus to a great extent an act of faith concerning the
future. If the businessman guesses wrongly, the penalties can be great. If he
decides not to expand capacity and the market for his product expands, he
can fall irrevocably behind his more foresighted competitors. If, on the
other hand, he decides to expand capacity and his market does not expand,
he can be saddled with unused plant and equipment, the fixed costs of
which may ruin him. The businessman does his best to predict the extent
of his market, but many things can influence it other than the tastes of
households. A new Government may adopt different taxing and spending
policies that affect him profoundly; the apparent success, or failure, of a
disarmament conference may cause some lines of production to look more
profitable and others less profitable. The rise of a new method of trans-
portation, a revolution in South America, or a relapse in the health of
President De Gaulle can all influence him in important ways that are
hard to predict. Occasionally, mob psycholog)' can predominate investment
decisions and a feeling of pessimism about the future can snowball into a
general cut in investment expenditure, or a feeling of optimism can snow-
ball into an investment boom based on expectations that later turn out to
be false.

There no doubt that the state of business expectations affects invest-

is

ment expenditures and that a general psychological reaction can cause

major shifts in investment. We shall see, in Chapter 45, that such changes in
expectations are capable of setdng off major expansions or contractions in
the level of business acti\'ity.

1 See, for example, D. Smyth, ‘Empirical Evidence on the Acceleration

Principle’, Rninv of
Economic Studies, 1964.

20*
CHAPTER 45

FLUCTUATIONS IN THE LEVEL

OF ECONOMIC ACTIVITY

The general trend of most industrial societies, is upward output, employ-

ment, and living standards all tend to nse over time If you compare any
year in the 1960’s with any year in the first decade of this century, your
overwhelming impression will be one of growth, even if you choose a year
of low activity from the I960’s and compare it with a boom year from the
1900’s Living standards are vastly higher now than then
If, however, you take each year of the 1950’s and 1960’s and compare it

with the year following, you will find that economic activity proceeds in an
irregular path, with forward spurts followed by pauses and even relapses
The irregular and often violent movements of the economy over short
periods of time have long occupied the attention of economists These move-
ments were once commonly known as business cycles or trade cycles, the
word cycle suggesting a regular oscillation of good times and bad At some
times and places, these movements have been remarkably regular
Figure 45 1 shows a time senes of the percentage of the labour force un-
employed in the United Kingdom from 1861 to 1950* The cyclical
fluctuations arc immediately apparent In the nineteenth century there was
a quite regular cycle of varying amplitude but with a duration of between
eight and ten years The level of unemployment varied continuously, there
were no prolonged penods either of full employment or of heavy unemploy-
ment Here then is a regularity in the data that requires explanation ‘Why’,
the student should immediately ask, ‘did the economy show such regular
cyclical fluctuations^’
The period between the two World Wars presents a dismal picture of

1 The data for the nineteenth century refer only to trade unions which probably covered
the more volatile parts of the economy Thus the actual varialions for the whole labour force
— ttiay have been somewhat less than recorded for the unionistd sector
 1940

-1950.

IRGl

1930

Britain

Great

1920

in

unemployed

1910

Force

Labour

1900

Unionised

the

1890
of

percentage

1880
The

45.1

Fii;

1870

1862
620 THE CIRCULAR FLOW OF INCOME

heavy unemployment The unemployment of the late 1920’s v,zs a local

British phenomenon associated with the return to the gold standard and the
long term decline in some of Britain’s staple export mdustnes This hea \7
unemployment was not matched elsewhere in the world, in the United
States, for example, the middle 1920’s was a period of boom The 1930 s,
howe\er, saw heavy unemployment throughout the world At the wont
point in the depression almost one person in every four was unemployed in
the United Kingdom, a similar situation ruled in America and in many
other industrialised countries The data displayed in Figure 45 I show that
the Great Depression of the 1930’s was unmatched m
seventy and duration
by any other depression in the last 100 years It is important to note, how-
ever, that unemployment did not remain at a constant level throughout the
period, there was considerable vanation from one year to the next

Fig 45 2 The percentage of

the total labour force
unemployed m the UK and
the US 1945 1964

194S 1950 1955 I960 1965

During the Second World War unemployment fell to a very low level

indeed It is instructive to note that, m

spite of the efforts of a total war, it
was not until 1942 that all of the unemployed could be put to work In 1940,
one year after the war had begun, 7 per cent of the unionised labour force
was still unemployed Since the war, unemployment has fluctuated, but
over a very much narrower range than in any other 15-year period over the
last hundred Even when all possible allowances are made for changes in the
definitions of the unemployment figures, the post-war experience represents
a substantial reduction in the average Icv'cl unemployment Has this been
by accident or design’
Unemployment rates since the Second World War have displayed much
lessregular fluctuations than those of the pre First-War penod But regular
or irregular the fluctuations have always been present Figure 45 2 shows
the percentage of the labour force unemployed m
the United States and the
United Kingdom for each year since the Second World War Fluctuations
are presentm both economics, but they are not as regular as the nineteenth-
 FLUCTUATIONS IN THE LEVEL OF ECONOMIC ACTIVITY 621

centur)' ones. Indeed,from 1959 to 1964, the US unemployment rate was

remarkably steady at the disturbingly high rate of about 5-5 per cent.
The
only exception ivas in 1961 when the rate rose to 6-7 per cent.
Some students of industrial fluctuations have thought that they ivere able
to discern several t^^ies of cycles in the level of economic activity.
One such
cycle, ivhich is clearly obser\'able in the British nineteenth-centuty'
employ-
ment had a duration of about nine years from peak to peak. This
series,

nine-year cycle was the one usually identified in the past as the trade cycle.
A second ty^pe of cycle, for which there is considerable ewdence, is one of
much shorter duration, lasting an^^vhere from 18 to 40 months. This cv'cle
is sometimes associated with variations in business ini'entories; ivhen
inventories are being built up, purchases by firms Mill exceed their sales;
when inventories are being reduced, purchases will be less than sales. W'e
shall see that the building up and running down of inventories can give rise
to cyclical oscillations in the economy. Finally, some economists have
thought that they perceived a very long cycle of about 50 years’ duration.
These cycles ivere thought to be associated with, among other things, major
on some fundamental innova-
flourishes of investment acti\ity consequent
tion,each burst being followed by a long pause in investment, once all the
most obvious new lines have been exploited. Of aU the ‘c)'cles’, this long-
w'ave one is the most conjectural, and w'e shall say nothing further about it

in this book.
From time to time, many have been put forward to
different theories
explain these fluctuations in the economy. In the present chapter we can do
little more than pro\ide a very general introduction to this interesting and
difficult subject.

PHASES OF THE CYCLE

Figure 45.3 shows a stylised cycle and diUdes it into four periods. We shall
first briefly describe the general characteristics of each of the phases of the

cycle and then describe one possible theory of the cycle.

Depression : A depression is characterised by hea\w unemplo)Tnent and a

level of consumers’ demand that is low in relation to the capacity of industrj'
to produce goods for consumption. There is, thus, a substantial amount of
unused industrial capacity. Some prices may be falling while others wiU be
unchanged, but few' if any wall be rising. The average level of prices wiU tend
to drift sloivly downw'ard. Business profits as defined by the businessman,
let alone as defined by the economist, w’ill be low and,
in many indiri-
the future will be lacking
dual cases, they will be negative. Confidence in

and, as a result, businessmen will be unwilling to take risks in making new’

investments. Banks and other financial institutions will have surplus cash
622 THE CIRCULAR FEOW OF INCOME

thatno one whom they consider to be a reasonable credit nsk washes to

borrow

Recovery When something sets ofF a recovery, we say, as a matter of

terminology, that the lower tuning pevtl has been reached Once begun, the
pace of recovery is likely to quicken Old and obsolete machinery will be
replaced Employment, income and consumers’ spending all begin to nsc
Expectations will become more favourable as a result of increases in pro
duction, sales and profits Investments that once seemed nsky may now be
undertaken as the climate of business opinion begins to change from one of
pessimism to one of optimism As demand expands, production will be

expanded with relative ease merely by rc*emp]oying the existing unused

capacity and unemployed labour Prices will stop falling and will generally
tend to stay constant or else to nsc slowly

Boom As the recovery proceeds, bottlenecks begin to occur in various

industries capacity will be fully utilised first in some industries
all existing

then in others, labour shortages begin to occur, particularly m

certain key
skilled categories,and shortages of certain key raw matcnals develop It
now becomes increasingly difBcult to increase output merely b> putting
unused resources to work, since the supply of unused resources is rapidly dis-
appearing, output can 6e raised lurfficr only 6y means of mvestmeiTi*,
which raises the productivity of already-employed labour Further rises in
demand arenow met more by increases m prices than by increases m pro-
duction As shortages develop in more and more places, a situation of
general excess demand for labour develops Costs nse but pnees nse also,
and business remains generally very profitable Losses arc infrequent, since
a money profit can be earned merely by holding on to those goods whose
prices are rising over time Investment expenditure will be heavy, invest-
 FLUCTUATIONS IN THE LEVEL OF ECONOMIC ACTIVITY 623

ment funds wll be in short supply, and interest rates may rise in the face of
a hea\y excess demand for loanable funds. Expectations of the future
are
favourable, and much investment may be made that is not justified
on the
basis of current prices and sales and that requires further rises in prices
and
sales to render it profitable.

Recession: The point at which the boom turns into recession is called the
upper turning point. Once a recession sets in, it tends to gather its own
momentum. Consumption demand falls off. Inv'estments that looked profit-
able on the expectation of continuously rising sales and prices suddenly
become unprofitable. High interest payments, which seemed easily bearable
when sales and prices were rising steadily, now become a hea\y burden.
Business failures, which ^vere ver^' infrequent in the boom period, now be-
come more common. Production and employment fall as a result and, as
employment falls, so do income and expenditure; as demand falls, more and
more firms get into difficulties. Prices and profits fall, and new investment
is reduced to a very low level. It is very often not even worth replacing
capital goods as they wear out, since unused capacity is increasing steadily.
When the decline is spent, a period of full depression sets in, and we are
back where we began.
This discussion is brief and stylised, but it gives some picture of the t^qjical

elements found in the various phases of most cycles, and thus in the theories
that seek to explain them. No two cycles are exactly the same. In some, the
recession phase is short and the resulting depression is not severe; in others,
a full-scale period of stagnation sets in. In some c)'cles, the boom phase
develops into a severe inflation ; in others, the pressure of excess demand is
hardly felt.

A THEORY OF THE CYCLE

Why is it that market economies do not settle down into some position of

equilibrium, maintaining more or less full employment at all times? In

order to answer this question, we shall develop a theory of the cyclical be-

ha\dour of the economy — a theor)' that attempts to account for the fact that
such economies tend to progress cyclically rather than smoothly. This ele-
mentary theory, which brings together a number of ideas and theories that
we developed throughout Part VII, is dmded into three parts: first, a

theory" of cumulative upswings and dowmswings, explaining

why, once
started, booms and slumps tend to carrj^ on of their own momentum,
second, a theory' of floors and ceilings, explaining why cumulative
upward
to a halt, and, third, a
and dow'nward movements are eventually brought
theor>' of instability, explaining how a process of upward or downward
62t THE CIRCULAR FLOW OF INCOME

movement, once U brought to a halt, tends to rcvcne itself The theory

is

here presented is many theones that have been offered

but one of the
^V'e ha\e chosen to present this one because it has a reasonable intuitive

appeal A formal dev elopment of any theory of c> clical behav lour requires
mathematics This is true of the present theory as with any other, but the
present one can at least be apprcaated at a more impressionistic level The
theor> IS also interesting because it does stress the importance of investment
expenditure m the cyclical process and most economists arc agreed that, no
matter how it may actually affect the process, investment is an important
influencing factor It must be stressed, however, that this is only one of
several competing theories, the choice between which depends ultimately
on the balance of empirical evidence

CuMULATiv E sfovEMENTS Why does a period of expansion or contraction,

once begun, tend to go on of its own momentum ^ First, the multiplier pro-

cumulative movements As soon as a revival begins, some

cess tends to cause

unemployed labourers find work again These people, with their newly
acquired income, can afford to make much-needed consumption cxpcndi
lures This new demand causes an increase in production and creates new
jobs for other unemployed workers As incomes rise, demand rises, as
demand nses, incomes rise Just the reverse happens in a downswing Un-
employment in one sector causes a fall m demand for the products of other
sectors, which leads to a further fall in employment and a further fall m
demand
A second major factor is to be found in the accelerator theory of invest-
ment demand, discussed m Chapter 44 New investment is needed to expand
existing productive capacity and to introduce new methods of production
When consumer demand » low and there w excess capacity, investment is
likely to fall to a very low level, once income starts to nse and entrepreneurs

come to expect further rises, investment expenditure may rise very rapidly
Further, when full employment of existing capacity is reached, new invest-
ment becomes the only way available for entrepreneurs to increase tiicir
output Since a capital good lasts many years, the value of a machine wall
generally be greatly in excess of the annual value of consumption goods that
It produces If a machine costing £4,000 produces £1,000 worth of goods

every year, then it will be necessary to spend £4 on investment for every

desired increase of £1 in the annual production of consumers’ goods For
these reasons, investment demand may nse by very much more than con-
sumer demand, and even a moderate fall off in consumer demand may
reduce desired new investment almost to zero The volatility of investment
expenditure is one of the important causes of fluctuations in market
economies
 FLUCTUATIONS IN THE LEVEL OF ECONOMIC ACTIVITY 625

A third major explanation for cumulative movements may be found in

the nature and importance of expectations. All producdon plans take time
to fulfil. Current decisions to produce consumers* goods and investment
goods are very strongly influenced by the expectations of businessmen. Such
expectations can be volatile. Expectations can, for quite long periods of
time, lead to their own realisation. If enough people think, for example,
that stock-market prices are going to they will all buy stocks in
rise,

anticipation of a price and these purchases ivill themselves cause prices

rise,

to rise; if, on the other hand, enough people think stock-market prices are
going to fall, they will sell now at what they regard as a high price and
thereby actually cause prices to fall. This is the phenomenon of self-realising
expectations. If enough businessmen think the future looks rosy, they may all

begin to invest in increasing capacity; this wall create new employment and
income in the capital-goods industries, and the resulting increase in demand
wll help to create the rosy conditions, the expectations of which started the
whole process. One cannot lay down simple rules about so complicated a
psychological phenomenon as the formation of expectations, but it is often,
if not always, true that they show a sort of band-wagon effect. Once things
begin to improve, people expect further improvements, and their actions
based on this expectation help to cause yet further improvements. On the
other hand, once things begin to worsen, people often expect further
worsenings, and their actions based on this expectation help to make things
worse.
Variations in the burden of fixed money debts when changes occur in the

price level also cause cumulative movements. Debts and other contractual
payments are fixed in terms of money. Practically all borrowed money is
obtained at the cost of an interest payment fixed in money terms. Prices
and costs rise during a boom, but interest payments do not rise in pro-
portion. Thus, if prices and other costs double, profits will more than
double, because one element of costs, fixed interest payments, will not rise.
On the other hand, if prices and costs fall, profits will fall further, because
fixed interestpayments \vill not fall at all. Thus, on the ups^nng of a cycle,
the real burden of fixed money payments diminishes, making it easier to
earn profits, whereas, on the downswing, the real burden increases, and
profits fall off.
There are other reasons too why recoveries and recessions, once started,
tend to build up a momentum of their own, but we shall not go into them
here. We note in passing that this tendency complicates government
anti-

cyclical policy. It is not sufficient to know what measures will stop a

boom; it
is necessary to appl)' these measures at
just the right time and in just the

right strength to curb the boom without setting off a downward spiral that

%vill gather momentum of its o\vn accord.

 FLUCTUATIONS IN THE LEVEL OF ECONOMIC ACTIVITY 627

limits, floors, and slow up or stop upward and downward

ceilings that
processes, and (3) that,once such a process is stopped or slowed down, it
will reverse its direction, vill be complete.
The accelerator theor)^ of investment
is probably the best known of the

theories accounting for this reversal of direction. This theor>' was introduced
in Chapter 44, and )nu should how' re-read pages 614—16.
The accelerator makes the desired level of new (not replacement) invest-
ment depend upon the rate of change of income. If income is rising at a
constant rate, then investment irill be at a constant level. If the speed at
which income is rising slackens, then the level of investment will decline.^
This is illustrated in Table 45.1 for a hN-pothetical example where I—?>AY.
If income merely levels off to a constant level, then new im'estment may be
reduced to a very loiv level. This means that a levelling off in income at the

Table 45.
1

Period T AY /

1 100 '

2 200 100 300

3 300 100 300
4 375 75 }
225
5 425 50 150
6 450 25 75
7 460 10 30
8 465 5 15
9 465 0 0

top of a cycle ivill lead to a decline in the level of investment.^ The accelerator
thus provides a theory of the upper turning point, although not one for the
lower turning point. The decline in investment at the upper turning point
will cause a decline in the level of income that wall be intensified through

1 The concept of a declining rate of increase alwap causes trouble when it is first en-

countered. The student familiar with the differential calculus will recognise that it refers to a
time.
function with a positive first derivative and a negative second derivative with respect to
increasing
Others will recognise the values of 1^ in Table 45.1 as an example of a series that is
at a declining rate.
relation is /=3dr. The
2 Let us consider a further example. Assume that the accelerator
accelerator coefficient is 3 and the total of new investment occurs in the
year in which the rise

in income takes place. Now assume that income is rising at a constant amount of 100 each year.
constant lerel of 300. If income now continues to rise, but only by 50
Investment will be at a
at a constant amount,
each year, the level of investment falls to 150 a year. If income levels off
the level of investment falls, in this very simple example, to zero.
620 THE CIRCUI.AR FLOW OF INCOME
^

!
the multiplier process the fall in income continues, the floor c\cmu
ally be reached After a while, investment may nse exogenousl) If it does
not, then, once existing capaat) falls to the level suitable to current output,
there will be a revwal of replacement demand, and new machines will now
be bought as old ones wear out This nse in the level of activity in the
capital goods industnes causes, by way of the multiplier, a further nse in
income in response to which new investment will take place, leading to yet
further rises in income So a multiplier and an accelerator combined with
‘ceilings and may be sufficient to set up an endless cyclical process
‘floors’

m the economy To present and develop such a theory formally requires the
*
use of some elementary mathematical tools

THE INVENTORY CYCLE

Business firms all hold inventories of materials and finished goods These
inventories fluctuate widely, as Figure 45 4 illustrates Such sharp fluctua

i-ig -fj 4 Fluctuates in the level of inventories about their trend value in the
the US 1939-1962

tions are amajor cause of the short term vanations in the level of activity
How does this inventory cycle’ occur’ The theory of inventory cycles is
very similar to the accelerator mechanism, only we now emphasise invest
ment in tnvaitories of goods rather than in plant and equipment
Start by assuming national income to be in equilibrium, with withdrawals
equal to injections at the full employment level Now assume, m order to

get the process started that there is an autonomous nse in the propensity to

save (a fall m the propensity to consume) The first result of the fall in
demand will be a piling up of unsold goods on dealers’ shelves After some
time, dealers will reduce their orders so as to prevent imentones from
1 The wi5h« to see »och a thcMy handled mathematical!) should consult
siudttit \*Vio
William H Baumol Economic Dynamul ilii /alraAnrtioa (2nd rev ed Macmillan 1959)
 fluctuations in the level of economic activity 629
increasing indefinitely: retailers will reduce purchases
from wholesalers and,
after wholesalers stocks have risen, they in turn will reduce their purchases
from manufacturers. Manufacturers may maintain production for
a while,
adding the unsold goods to inventories, in the hope that the fall in
demand
isonly temporary. If this proves not to be the case, manufacturers will
cut
back on production, laying off some tvorkers and reducing the hours worked
by the remainder. Thus income and output ivill begin to fall and, at this
isdll have risen to an abnormally high level. Once pro-
stage, inventories
duction a level equal to the new (lower) level of consumers’ demand,
falls to

there be no further rise in the level of stocks. Unfortunately, however,

w'ill

matters will not remain at this point. Stocks will now be too high on Uvo
counts : first, because sales will be at a lower level than they were originally,
and, second, because stocks MU have increased during the transitional pro-
cess. In order to work off excess inventories, retailers %vill buy less from
wholesalers than they are selling to consumers, wholesalers will buy less
from manufacturers than they are selling to retailers, and manufacturers
will produce less than they are selling to wholesalers. Thus the current
level of output, and hence of income earned by households selling factor
services to firms, will fall below the current level of sales. This fall in income
will reduce the level of demand still further. As long as production can be
held below the level of current sales, then inventories \vill be falling, even
though the level of sales is itself falling. Once inventories are reduced to the
desired level, the retailers and wholesalers will increase their orders, so that
they are equal to current sales, thus keeping inventories constant; manu-
facturers will also increase the level of output until it is equal to the
(increased) level of sales, thus keeping their own stocks at a constant level.
But means that production, and hence income earned by households,
this
is increased. As this happens, the demand for goods MU rise. The initial

impact here wiU be on inventories, which MU be run down as sales rise

unexpectedly. Now the whole process is set into reverse. For a while, every-
one’s inventories MU be run down, but then orders will increase, first from
retailers, then from wholesalers. Finally, the output of manufacturers MU
increase. T his means that incomes, and Mth them the level of demand, will
rise. Once production is increased to the level of current sales, inventories

MU no longer be decreasing. But now the level of inventories is too low on

two counts: because the level of sales is higher than it was when
first,

inventories w^ere at the correct level, and, second, because inventories have
been run down during the transitional phase. In order now to build up their
inventories, retailers wiU order more from wholesalers than they seU to con-
sumers, wholesalers will buy more from manufacturers than they sell to
retailers, and manufacturers MU
produce more than they sell to wholesalers.
This rise in production ^viU raise incomes and thus raise the level of demand
630 THE CIRCULAR FLOW OF INCOME

sull further As long as production is kept above the level of sales, however,
inventories will be rising in spite of the
fiutt that sales are also rising Once

the level of inventories brought up to the desired level, orders will fall off
is

Retailers and wholesalers will reduce orders to the level of current sales, and
manufacturers will reduce output to that level as well But this fall in output
will reduce incomes and with it demand For a while, inventories will pile

up, but orders and output will be cut back, thus reducing the lev e! of income
and demand If yougo hack to the bcginmng of this paragraph and te-rcad
It, you will find out what happens next the whole downward process is set
in motion again
Although this kind of verbal anal>sis can provide some general ideas of
the cyclical process, mathematical tools are essential if wc want to carry our
analysis much further In some branches of economics, one can get a long
way by means of careful verbal and geometrical analysis In the field of
dynamic fluctuations, one can get practically nowhere We will certainly
want to investigate a number of questions like the following What are the
effects of varying the time lags with which firms react to changes in their
sales’What if the reaction does not occur suddenly but is through
time’ Under what circumstances will such a cycle die out rapidly so that
income converges on its equihbnum level’ In what circumstances will the

self-excitmg process continue indefinitely, so that the cycle will itself carry

on mdefimtely by consaous government policy’ What will

unless stopped
be the effect of government controls built into the system in an attempt to
dampen these fluctuations’ What difference will it make if the government's
control mechanism itself acts with a time lag’ To answer such questions,
mathematics is indispensable particularly the sort of mathematics that
electrical engineers use to analyse self-exciUng {closed loop) control systems
For the student interested in dynamic fluctuations, a knowledge of such
'
techniques has become essential

l One of the b«« surveys of the formal economic vheory of flutluaiions and of iHe mathe-
matics used therein is to be found in the first half of RGD Allen, Mathimatical Economics
(2nd ed Macmillan 1959)
 CHAPTER 46

FOREIGN TRADE AND

NATIONAL INCOME

In Chapter 42, we developed predictions about the effects of imports and

exports on national income. Before continuing, we briefly summarise the
analysis of that chapter.

IMPORTS
Imports are regarded as withdrawals from the circular flow ofincome. When
a domestic household purchases a commodity manufactured abroad, it
createsincome for foreign firms. Imports thus represent money earned by
domestic households, but not passed back to domestic firms through con-
sumption expenditure. In that sense, imports withdraw money from the
domestic circular flow. If households switch some of their expenditure from
domestically produced to foreign-produced goods, the level of national
income will fall in just the same way as it will fall if households reduce their
expenditure on domestically produced goods. In both cases, the volume of
withdrawals from the circular flow of income increases. The effects of the
withdrawals are magnified by a multiplier process.
To see this consider the effects of a change in the purchasing patterns of
British households so that they decide at each level of income to buy more
Volkswagens and fewer British cars. Fewer British cars will be produced, so
that fewer workers will be needed in the car industry, thus Bridsh national
income will fall. But this is not the end of the process. Since UK
car workers
will have less money to spend, they can buy fewer goods and services: fewer
dresses, fewer radios and fewer holidays in Brighton.
The incomes of the firms producing these products will fall and their
payments to their employees will fall correspondingly as they reduce their
outputs. This will induce a further contraction in demand as dressmakers.
632 THE CIRCULAR FLOW OF INCOME

radio workers and salesmen, dishwashers, waitresses and fairground opera-

tors in Brighton find their incomes falling This multiple process continues
until the original fall in sales of the car industry has magnified into a much
larger general fall m sales and incomes throughout the economy

EXPORTS
Exports provide an injection into the domestic circular flow of income
When a foreign household purchases a good manufactured in the UK, it

income for a UK firm Exports thus represent income earned by

creates
UK firms that does not arise from the spending of UK households In this

sense, exports inject expenditures into the domestic circular flow of income
When foreign households buy British goods, the level of UK national in-
come rises in just the same way as it rises when UK firms spend more
money on investment In both cases, the volume of injections into the
circular flow of income increases The effects of the injections are magnified,
both directly and indirectly, by a multiplier process If German households
decide to buy fewer Volkswagcns and more British cars, British national
income will nse More British cars will be produced and more workers will
be needed in the UKcar industry This, of course, is not the end of the
process Since UK
car workers will have more money to spend, they will
purchase more goods and services, thus increasing the income of the firms
making these Firms providing such goods and services as night clubs, hotel
accommodations, and radios will find their incomes rising, since their sales
to car workers will nse The employees of these firms will increase their
spending, and so the multiplier process will continue until the original rise
m sales and incomes has multiplied throughout the economy
When we speak of a foreign-trade multiplier, we mean the multiplied
on national income of a change in either imports or exports A change
effect
m exports has thesame effect on national income as has any change in
autonomous expenditure A change in imports has the same effect on
national income as has any change in withdrawals from the income stream

THE FINANCING OF IMPORTS'

Imports must be paid for in foreign currency The German firm selling
Volkswagcns to British households must in the end receive marks pay- m
ment so that It can pay for the factors of production it has hired to make its
cars The foreign money needed to pay for these imports comes from British
exports Exports are sold in the first instance for foreign currency, say,
marks But the Bntish exporter requires payment in pounds in just the same
I This problem is discussed m more detail m Chapter 53
 FOREIGN TRADE AND NATIONAL INCOME 633

way as the German exporter required payment in marks. British

exports
thus produce a supply of foreign currency that people are trying to trade
for
pounds while British imports produce a demand for foreign currency. If, at
the current exchange rate, the value of imports equals the value of exports,
there will be exacdy as much foreign currency earned by exporters as is
needed by importers. Thus, imports rvill be ‘financed’ by exports.
But what if the value of imports does not equal the value of exports ? If
imports exceed exports, more foreign currency is being spent than is being
earned through exports. In this case, one of two things must be happening.
Either the difference is coming out of exchange reserves, or it is being pro-
vided by foreign lenders. On the other hand, if exports exceed imports, then
more foreign currency is being earned than is being spent, and the difference
between what is earned and what is being spent must be being used either
to add to reserves of foreign exchange or to lend money abroad.

A Formal Analysis of the Foreign Trade Multiplier*

We now wish to study the effects of various changes in imports and exports
on domestic national income. To make the analysis as simple as possible
we shall ignore the government sector and deal iwth a model wth only two
injections, investment and exports, and two withdrawals, savings and im-
ports. Our behavioural hypotheses are the same as those used in earlier
chapters: savings and imports depend on domestic imports, while invest-
ment and exports are exogenously determined constants. These can be
summarised as follows;

M=Z+mY (1)

S = W+sY (2)

Z=J (3)

1=1
These hypotheses are shown graphically in Figure 46.1. Z and are W
constants in the imports and savings functions, Z being positive and W
negative. They represent the quantity of imports and saving that would
occur if income were zero. A change in the value of Z and will shift theW
two functions without changing their slope^ representing a change in the de-
sire to import or save at each level of income ivithout
any change in the
marginal propensities to save or import. The values of m and s indicate the

1 This section may be omitted.

not obvious to you, review section 8 on page 38 of the
Appendix to Chapter 2.
2 If this is
634 THE CIRCULAR FLOW OF INCOME

marginal propensities to import and to save because they indicate the

fraction of any change in income that will go on imports and savings *
X
and / are the constant flows of exports and investment expressed as amounts
per period of time
In equilibrium, injections must be equal to withdrawals so that

M+S = X-hl (5)

Substituting our behavioural hypotheses from equations (!) to (4) into this
equilibrium condition, we obtain

Z+mY+lV+sY= X+I
which solves for Y as follows^

{X-z)+{/-yv)
(
6)
s+tn

This gives us the equilibrium level of income

Note that the terms for savings and imports (withdrawals) enter the ex-
pressionsymmetncally with each other as also do the terms for investment
and exports (injections) Inspection of equation (6) leads to the following
conclusions

1 A change in autonomous exports and a change in

autonomous investment have identical effects on
national income. A change in either changes national
income in the same direction. To see this raise first X
and then I by one vuiit in equation (6) and see that both
changes will increase Y by the same amount.
2 Changes in savings and imports have identical
effects on national income. A rise in either lowers
national income. Shifts in the savings and import
functions can be accomplished by changing the values
of Z and If'. Clearly a change of one unit in the value of
Zhas the same effect on the value of the expression as
a change of one unit in the value of fK
1 The marginal propensity to import is AAIjAY while the average propensity is MIY
2 The algebraic steps arc as follow*
 FOREIGN TRADE AND NATIONAL INCOME 635

Fig 46.1 Derivation of the aggregate withdrawal and injection functions in a

model with only two withdrawals, savings and imports, and only ttvo
injections, investment and exports.

3 An upward shift in the savings or the import func-

tions will lower national income in exactly the same
manner as would a downward shift in the export or
investment functions. An increase of X or / by one
unit combined with an increase of or fV by one unit
leaves the value of the expression unchanged. The
common sense of this result being that if more is
 FOREIGN TRADE AND NATIONAL INCOME 637
thus the American national income fairly well insulated
is from fluctuations
in the income of other countries.

IMPORTS AND EXPORTS: GOOD OR BAD?

It is very commonly argued that exports raise national
income and hence
are ‘good’, whereas imports lower national income and hence are
‘bad’. The
view that exports are beneficial and imports harmful goes back at least as
far as the eighteenth century.
When we say exports raise national income we mean that they add to the
value of output. But they do not add to the value of domestic consumption.
In fact, exports are goods produced at home and consumed abroad, while
imports are goods produced abroad and consumed at home.
The issue involved in this section is a most important one, and we shall
approach it indirectly by thinking not of exports and imports, but of a
family that sells produce to the outside world (‘exports’) and buys goods
from the outside world (‘imports’). Suppose the farm family sells every-
thing it can grow, but buys nothing. The money will roll in, but the family
will get hungrier and hungrier, the car will run out of petrol, their clothes
will become tattered and their radio will eventually stop working. Clearly,
;

they cannot be better off than if they used some of their cash to buy needed
goods. Of course, this is an extreme case, but not without some interest or
some modern international parallels.
The standard of living of a person or of all persons in a country depends
on the goods and services that they consume, not on what they produce.
The may be
average material standard of living of the residents of a country
thought of in terms of the following equation

total of goods and services consumed

Average standard of living
number of people

If exports are really good and imports really bad, then a fully employed
economy that obtains an increase in its exports without any corresponding
increase in its imports ought to be made better off thereby. This change will,
however, result in a reduction in current standards of living, because, when
more goods are sent abroad and no more are brought in from abroad, the
total goods available for domestic consumption must fall. The view that
exports are good and imports are bad implies that the best of all possible
worlds would be the situation in which a country exports all its production
and imports nothing. No matter how ‘good’ or ‘bad’ this situation may
appear to you, you should not fail to notice that all the inhabitants of the
country would soon starve to death, since there would be no goods and
services of any kind available for domestic consumption.
638 THE CIRCULAR FLOW OF INCOME

^Ve must nosv ask in more general terms what happens if a countn
achIe^cs a surplus of exports over imports for a considerable penod of time ^
It unll be accumulating claims to foreign currenc) There are three possible
‘

uses for the foreign currency earned by exporting it may be used to buy
foreign goods, to make in\cstments abroad, or to add to foreign exchange
reserves Foreign-exchange reserves are required for the smooth functioning
of a system of fixed exchange rates Such reserv cs do not help the functioning
of the system if they arc accumulated beyond the level necessary to reduce

the chance of running out of reserves to a very low level After that, foreign
currencies are useful only if they are spent to purchase useful goods and
services We cannot eat, smoke, drink or wear US dollars or Indian rupees
But vse can spend them US dollars and Indian rupees can be used to buy
American and Indian goods that can be eaten, smoked, drunk or v,om
When such goods are imported and consumed, they add to UK living
standards Indeed as we shall see in Chapter 52, the mam purpose offoreign
trade is to take advantage of international specialisation, trade allows us to
consume more than would be possible if all goods were produced at home
From this point of view, the purpose of exporting is to enable one to import
goods that can be produced more cheaply abroad than at home Exporting
more than one is importing makes more sense if the purpose is to obtain
funds to invest abroad (Such foreign investment only makes sense if the re
turn for an equal nsk investmentis as high as or higher than the rttuni that

can be obtained on an investment made at home ) But such foreign invest

menu add to li\ ing standards only when the interest earned on them is used
to buy imports that do not have to be matched by currently produced
exports i c when in the future, they produce an excess of impwrts over
,

exports From this point of view, the purpose of exporting more than one is
importing in order to make foreign investments is tvmtually to be able to
import more than one is exporting'

The living standard of a country depends on the goods

and services consumed in tbat country. The import-
ance of exports is that they permit imports to be made.
This two-way international exchange is valuable be-
cause more goods can be imported than could be
obtained if the same goods were produced at home.

An Economy with Unemployment an Exception?

\\ hat about an economy wnlh sut»tantial amounts of unemployment, such
as (he US economy m
the early ]960*sor the Bntish economy the 1930's’ m
Assume there is a nsc in that country’s c-xports without any corresponding
Th« mfchanum Tor this u evpUtned in more detail in Chapter 53
1
 FOREIGN TRADE AND NATIONAL INCOME 639
nse in its imports, perhaps because the
government has put a subsidy on
exports and increased the rates of tariffs charged on
imports. As rve sa\v in
the last chapter, this rise in exports will increase income
and employment.
Surely, in a time of unemployment, this is to be regarded as a ‘good thing'.
Two points need to be made about such a policy. In the first place, the
goods being produced by the newly employed workers in the export sector
are not available for domestic consumption and so do not raise domestic
standards of living.Would it not be better if, instead of subsidising exports,
the central authorities subsidised the production of goods for the home
market, so that the initial rise in employment would also contribute to a
rise in domestic living standards ? Or if one objects to the gov'ernment sub-
government could create new employ-
sidisation of private firms, then the
ment by building more roads, schools and research laboratories. Again,
income and employment would go up, but, again, there would be something
more tangible to show for it than the smoke of ships bearing the subsidised
exports to foreign markets disappearing over the horizon. ‘
The second point to be made concerns the foreign effects of such a policy
of fostering exports and discouraging imports in a situation of general world
unemployment. Although the policy raises domestic employment, it will
have the reverse effect abroad - it will create unemployment abroad. Such
a policy may therefore be referred to as one of ‘exporting one’s unemploy-
ment’. The foreign countries will suffer a rise in their unemployment, be-
cause their exports rvill fall and their imports rise. This will set up a
multiplier process that reduces their levels of income and employment. Even
if the governments of these countries are prepared to see their unemploy-

ment rise, cannot long continue, because these countries will

this policy

begin to have balance-of-payments deficits. They will soon be forced to

take steps to remove these deficits. If they do this by restricting imports, the
original country' will lose the stimulus that it originally obtained by
encouraging exports. If all countries tr)' such a policy of expanding exports
and discouraging imports, the net effect is likely to be a large fall in the
volume of international trade without much change in the level of employ-
ment in any country. ^
There is, however, one set of circumstances in which the policy could
provide a long-run solution. In the early 1960’s, the US had heav')' un-
employment and a balance-of-payments deficit, rvhile Germany had full
employment mth heavy inflationar>' pressure plus a balance-of-payments

the export
1 We are here considering the effects of the initial rise in employment in either
or the investment industries. Of course there will be multiplier effects of an increase in either
X or I and these will contribute to an increase in domestic living standards.
2 The experience of the 1930’s, when many major trading countries did adopt such a

policy, is discussed in Chapter 55.

640 THE CIRCULAR FLO^\ OF INCOME
surplus In these circumstances an Amcncan policy of discouraging imports
and encouraging exports to Germany tvould have raised US employment
and alleviated the US balance of payments deficit, while at the same time
reducing inflationary pressures in Germany and reducing her balancc-of
payments surplus This result would have been mutually satisfactory for the
US and Germany In general these mutually satisfactory conditions exist
when one country requires a nse in expenditure on its products because it is
suffering from unemployment while the other country requires a fall in
expenditure on its products because it not only has full employment but
substantial excess demand and consequent inflationary pressure as well
CHAPTER 47

GOVERNMENT AND THE

CIRCULAR FLOW OF INCOME

In Chapter 42, we studied briefly the effects of government taxing and

spending policies on the circular flow of income. We treated government
expenditure (G) as an injection into the circular flow of income it creates :

income for those firms and households that sell goods and services to the
government, but the income does not arise out of the expenditure of house-
holds. Government tax revenue ( 7^ was treated as a withdrawal from the
circular flow it is income earned by households and firms that cannot be
:

passed on by way of spending because it is removed from the circular flow by

government, A rise in the level of government spending or a fall in the rate of
taxes so that less will be raised in taxes at each level of income will raise
national income, for the former represents a rise in the injections schedule
and the latter a fall in withdrawals schedule. A fall in the level of govern-
ment spending or a rise in the rate of taxes will lower national income, for
the former represents a fall in the injections schedule and the latter a rise
in the withdrawals schedule. In all cases, the original change in government
expenditure or tax revenue be magnified through a multiplier process
will

so that the final change in national income will be larger than the initial
change in G or T.^

THE FINANCING OF DEFICITS AND SURPLUSES

The above discussion implies that have situations in which
it is possible to
government expenditure does not equal government tax revenue. When
revenues exceed expenditures, we say that there is a budget surplus; and
when revenues fall short of expenditures, we say that there is a budget
deficit. If the government spends more than it raises, where does the money

1 If you have any trouble with this argument you should re-read Chapter 42.

21
642 THE CIRCULAR FLOW OF INCOME

come from’ If the government raises more than it spends, where does the
monc> go to’ Basically, the difference between G and T is reflected m
changes m the level of the government’s debt If the expenditures exceed
revenues, the balance must be borrowed from someone, if revenues exceed
expenditures, the balance goes to pay off some of the loans that were made
m the past
A deficit requires an increase m borrovvang, for which there arc three
mam sources the central bank,* the commercial banks and the public The
money may be raised from someone who would have spent
by borrowing it

It for other purposes, or the money may represent a net addition to pur

chasing power Ifthe government sells bonds to the central bank, the central
bank can pay for these by creating new money The same thing can happen,
but within government sells its bonds to commercial banks In
limits, if the
both be a net addition to expenditure iti the economy once
cases, there will
the government spends this money If, on the other hand, the government
sells bonds to commercial banks or to households and obtains money that

would otherwise have been lent to private firms, the expenditure u mertlj

being rechannelUd from the pmate to the public sector of the economy The new
government expenditure still represents an increased injection into the

circular flow, but it is matched by an exactly parallel reduction in the

injection of private investment expenditure

A surplus allows the government to reduce its outstanding volume of
debt Bonds m the hands of the central bank, the commercial banks, or
households may be redeemed If the money is neither spent by those who
receive it nor lent to someone else to spend, there is a net withdrawal from
the circular flow If, on the oihcr hand, the money is spent by thosewho
obtain It, there will be a parallel increase in private spending and no net
change m total spending
In what follows, we shall study the common case in which budget deficits
cause net increases in the volume of spending, whereas surpluses cause net
decreases Other special cases can easily be analysed when their particular
circumstances are specified

Tbft •sf,

conceptually to divide government expenditures into currently

It IS possible

consumed items and capital items TTic former, such as the costs of ninning
the civil service, the
army or the police force, the costs of social security
payments, and the costs of the national health service, arc used up and add

I Tht Ctninl BinV of most cotmtnes n the controller of the lupply of money and W can
lend ihe government money without limit tiRipIy by creating new money TTm rnalter n
iM-uued in detail in Chapter <9
 GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 643

toliwng standards in the current period. Capital items, such as dams, roads
and schools, last a long time and yield their benefits to firms and households
over the whole of their lifetime. Assume for the moment that these capital
items are paid for by borrowed money
be repaid over the working
that is to
life of the asset. As a capital item wears
we will, if it has proved useful,
out,
wish to replace it with a new one. If the value of the capital assets owned and
operated by the government remained constant over time, then we would
expect the portion of the national debt required to finance capital expendi-
ture to remain constant. Their value does not remain constant, however,
and therefore the national debt does not remain constant either. There are
several forces that lead to a rising national debt.
First, if prices rise, we would expect that the national debt will have to
rise in order to keep the real resources that it represents constant. For
example, if the government issues bonds to cover the cost of building a
school, and if prices then rise, the cost of replacing the school will rise, as will

the money value of the debt necessary to cover the cost of replacement.
Second, if the government kept its relative importance in the economy
constant, we would expect the absolute size of the national debt to grow as
the productive capacity of the economy and the size of the population grow.
Economic growth of the kind we studied in Chapter 21 raises the total out-
put of the economy. If the Government’s share in total output is to be held
constant and if the consequent capital expenditures such as new schools,
roads, etc. are to be financed by borrowing, then we would expect the
national debt to rise as total output rose. Finally, if, as we get richer, we wish
to spend a larger fraction of our total income on those things produced by
governments, education, roads, parks, defence, we would have to increase

capital expenditures on the these, and we would thus expect the national
debt to grow faster than the national income is growing. Indeed there is
much evidence that many of the goods and services produced by govern-
ment have high income elasticities of demand so that we would expect the
proportion of the GNP produced by the government to rise as total output
rises.

In some countries the size of the national debt is tied to government

capital expenditures, borrowing is permitted for certain specified capital
expenditures but not for current expenditures. By analogy with an indivi-
dual firm this seems a not unreasonable thing to do, but there are major
objections to the practice. Perhaps the most important objection, is that it
is. Is an invest-
is extremely difficult to agree on what a capital expenditure

ment in the education of children and scientists a current expenditure, or

an investment in the future ? Is money allocated to research on the causes

of disease a current expenditure, or a capital expenditure? Because these

things do not correspond to physical assets, many people regard them as
not
644 THE CIRCULAR FLOW OF INCOME

being capital assets But they represent more of an m^ estment in the future
than a nuclearsubmanne, which, though it ts tangible, is providing a current
service (defence), not a future one If we segregate government expenditures
into tvso groups and make it one and harder to finance the
easier to finance
other, we of resources between the two groups
affect the allocation
The question of how to finance puUic expenditure has from time to lime
been a hot public issue It is a subject of great debate in the United States
although It is not in the forefront of political debate in Bntain WTiatever
the current state of public opinion, there is no doubt that when presented
with the logic of Keynesian income theory, most students feel that surely

there must be something wrong with spending what you have not earned
This reaction is sometimes a purely normative one about what one ought
to do But as often as not the reaction is the product of the feared conse
quenccs of a nsing debt, and it is within the province of positiv e economics
to stud) whether these consequences are indeed likely to occur Specificall),
It IS often feared that a rising debt transfers a burden to future generauons
and that, if deficits continue )ear after year so that the debt nscs without
limit,an impossible burden will eventually be placed on the country’s
economy But what « the nature of the burden of the national debt’ To
what level must it nse before senous economic consequences ensue’

The Burden of the Debt*

The national debt represents money owed by the government to those

households, firms and institutions that hold government bonds In this

debt is owed by all of us to some of us The

sense, the national existence of
the debtmeans that households have saved in the past by lending their
money to the government and that this money has been spent by the
government
In a very real sense, the cost of government activity cannot be postponed
to future generations W’e saw in Chapter 4 that the cost of doing something
can usually be measured in terms of the things that might have been done
instead If the government uses the economy’s scarce resources to build
dams schools, nuclear submarines or tanks, the opportunity cost is
measured by the other goods and services that the resources rrught have
produced This cost m terms of foregone alternatives is incurred, no matter

how the money to pay for the project is raised by taxes, by borrowing from
the public, or by creating new money The method of financing a project
1 There u a substantial current coRinnmv about the burden of the debt For a view
different from the one presented here see WiUiam G Bov'tn Richard G Das'u and Da«d
H K.opf The Public Debt A Burden on Future Generaiions'* ’ Anencm Econonat Rtriru:
''
tember 1960
 GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 645
dictates how the cost is distributed throughout the community, but it does
little or nothing to affect the total current cost.
The problems can best be seen by considering a wartime economy such
as that of the United Kingdom between 1939 and 1945. During a major
ivar, the government spends vast sums on war materials. A high proportion
of the nation’s resources are used to produce these materials. The real cost
of using these resources to produce war goods is measured in terms of the
consumption goods that might have been produced instead. This cost is
necessarily borne by the wartime generation, because it as a group has less
to consume than it could have consumed if there had been no ivar. The
total cost can be determined only when it is determined how many re-
sources are to be used for war production and therefore not used for the
production of consumers’ goods or for investments that would yield con-
sumers’ goods in the future. During the Second "World ^Var, for example,
virtually no durable consumers’ goods such as cars, refrigerators, radios, etc.
were produced in the UK, because the resources were diverted to war
production.
The method of finance cannot significantly affect this total cost, but it

does determine w'ho bears the cost and it does allow some group to postpone
the burden by persuading other groups to assume the burden in return for
some advantage in the future. If the cost of the war is met solely by taxes,
then current taxpayers bear the burden by ha\ang their consumption of
goods and senuces reduced by taxes. Consumer expenditure will fall by
just as much as the output of consumers’ goods falls when resources are
transferred to the production of war goods. If, on the other hand, the ivar
expenditure is financed solely by borrowing from households and firms, the
reduction in current consumption is suffered by those ivho lend their money
to the government rather than spending it on currently-produced goods and
services. People who do not buy government bonds do not postpone current
consumption and thus do not bear any of the real cost of the war effort.'
To the extent that the war is financed by current taxes, the matter is
finished once the ^varis over. Resources can then be transferred back to the

production of consumers’ goods, and households’ disposable income can be

allowed to rise by reducing taxes. To the extent that the war is financed by
borrowing, the debt remains after the war. It is necessary to pay interest
each year to the bondholders and eventually to repay the bonds as they
reach maturity. To the extent that interest payments and eventual redemp-
tion of the bonds are made from tax revenue, the post-war taxpayers are
suffering a reduction in their consumption below' what it \vould otherwise

1 A method of financing is by creating new money by selling bonds to the central

third
bank. In government purchases bid up prices, and the costs are borne by the people
this case,

through the reduced purchasing pou’er of their money incomes.

646 THE CIRCULAR FLOW OF INCOME

have been, and postwar bondholders are obtaining a nsc m their con-
sumption above what it could otherwise have been The transfer is now
reversed in return for bearing the wartime reduction m consumption,
bondholders or their heirs now enjoy a postwar nse in consumption, and
taxpayers who are not bondholders suffer a postwar reduction For a com
munity, the full cost m
terms of foregone output is all borne dunngthe war,
after the war, total production goes back to normal The oppoituiuly cost
could not be postponed, but some individuals (taxpayers) must now pay for
the war by transfemng their claims on current production to other
individuals (bondholders)
Exactly the same analysis apphes to the peacetime activities of a govern-
ment, assuming we are in periods of full employment If the government
builds dams, roads, schools, and nuclear submannes, the opportunity cost
of these is the cars, television sets, beer and innumerable other consumers’
goods that could have been produced instead This cost is necessarily
incurred by the current generation But the disinhutton of this cost can be
affected by the method of finance If all of the expenses are met by tax

revenues, then taxpayers are forced to bear the reduction m current con-
sumption If the costs are met byborrowing, the lenders voluntarily agree to
take on the current costs by sacrificing consumption Taxpayers are then
forced to bear the cost slowly over time as funds are raised to pay interest
to the bondholders and to provide for the eventual redemption of the bonds
The face value of the bonds represents the consumption originally foregone,
and the interest represents the additional payment necessary to persuade
bondholders to accept the whole of the burden the first placem
How do mtcrgcncrational issues enter the picture’ If you do not choose
tobuy bonds now, your heirs may be faxed to pay your share of the interest
payments to those who did buy them But this kind of debt is no different
from any other kind of debts that your heirs may inhent from you The
interest will be paid to other memben of their generation who have
inherited the bonds
Up to now, we have assumed the resources used by the government
would otherwise have been used by the private sector to produce goods and
services for current consumption There arc two other alternatives First,
the resources might have been used lo produce capital equipment to increase
output at a future date Tins would be the case if the government borrows
from households money that would otherwise have been loaned to firms for
investment in capital equipment by the firm In this case, the sacrifice is
piostponed to future generations No reduction in current consumption need
occur (The current opportunity cost of the government activity is measured
m terms of foregone output of capital goods for the private sector ) If the
* produced by the government add less to future national income than
 GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 647
the capital goods the private sector would have produced, there is an added
cost that borne by future generations in terms of a national income lower
is

than it would have been if the government had not borrowed the money,
but had left it to be loaned to private firms. ^
on the other hand, theIf,

capital goods produced by the government add

or more to national
as much
income as would have been added by the capital goods private firms would
have installed had they been able to obtain the money that the government
raised, then there is no additional opportunity cost of the government
activity over and above the consumption goods that could have been pro-
duced instead of the capital goods.
The second alternative is that the government borrows money to put
unemployed resources to work. If these resources would have been idle in
the absence of government spending, then there are no opportunity costs
in terms of current production.^

Summary: The above discussion of the national debt may be summarised

The opportunity cost of any government expenditure is measured
as follows.
in terms of the goods and services that could have been produced by the
resources used by the government. If the foregone alternatives ^v•ere to pro-
duce consumers’ goods, then the cost is necessarily borne by the com-
munity at the time at which the government activity takes place. The total
current cost cannot be avoided, but its allocation betv'een groups in the
society can be determined. If the expenditure is financed by taxes, then the
current cost is borne by taxpayers. If the expenditure is financed by
borrowing, then the current cost is borne by lenders, and the cost to

taxpayers postponed and spread over time. The taxpayers bear the cost
is

when their consumpdon is reduced by future taxes levied to pay interest

on the debt and to provide for its eventual repayment.

GOVERNMENT FISCAL POLICY

The inflationary and deflationary gaps: We have considered the
effects that government policy, with and expenditures,
respect to revenues
has on the national debt. We must now consider government policy with

1 This is one of the major points argued by those that hold that there is a real and important
burden imposed by the existence of the national debt.
2 Throughout this discussion we have neglected one real cost to future generations of
debt

financing. The existence of national debt requires a transfer of income. The government must
raise money by taxes and then pay out this money as interest payments to
persons holding

government bonds. There will be some real cost of collecting taxes and paying out interest.
Resources in terms of tax collectors and inspectors and accountants and clerks to look after
bond issues, will be used for this purpose. These resources would otherwise have been employed
to make goods and serv'ices for general consumption. Such real costs are but a very small
fraction of the interest payments on the national debt.
618 THE CIRCUl:.AR FLO%N OF INCOME

respect to re\ enues md expenditure as a means of influencing the behaviour

of the economy Fiscal policy may be defined as the policy of the central

authorities that attempts to influence the behaviour of key macro \anablcs

m the economy by working through the government budget
Before we discuss this aspect of policy, we need to define two important
concepts the inflationary gap and the deflationary gap In order to intro-
duce a clear cut division between these two, we shall use a simplifying
assumption ihere u a lettt of nalumal output that ton be produced u.hen all resources

are fully employed If aggregate expenditure is less than this level of output, prices mil

be constant and fluctuation in aggregate expenditure will call forth proportionate

fuctuahons in output, uhen the full'employmenl Itiel of output has been reached, out
put cannot be expanded no matter kou, great ts aggregate expenditure, and if total

expenditure exceeds the lalue offull employment output, all that mil happen ts that

prices uill rise and lariations tn aggregate expenditure will make their effect fell solely

on lanations m the rate of inflation

These assumptions arc illustrated graphically in Figures 47 I and 17 2 In

Figure 47 1 v\e draw the familiar aggregate expenditure functions with

equilibrium income at OF, In Figure 47 2 we show the corresponding

(i; The rquilibnum level of national income

Ftg 17 /
 Expenditure

Expenditure

(iii) An inflationary gap.

21 *
650 the circular flow op income

(t) The equihbnum level of national income

(ii) The deflationary gap
(ill) The inflationary gap
Fig 47 2
 GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 651

withdrawl and injection schedules. In parts (ii) and (iii) of both figures
we
add to this basic diagram a constraint to total output dictated by the level
that can be produced when all resources are fully employed.
In Figure 47.1(ii) the equilibrium level of income OY^ is less than the

full employment OYp. There will be unemployed factors sufficient to

level
produce, if they could be employed, Y^Yp of additional national income per
year, (This is the situation we have been studying to date;
you put Yp in if
any of the diagrams drawn in the last seven chapters,
must be to the right
it

of all equilibrium positions.) Let us now consider what would happen if

full-employment national income were attained for a short time. In Figure
47.1(ii) we see that expenditure would fall short of income by ab per year.
In Figure 47.2 (ii) we see the same thing in terms of withdrawals and in-
jections with the volume of withdrawals exceeding the volume of injections
by ab per year at the full employment level of income. We saw in detail in
Chapter 42 that this would lead to a fall in national income until income
reached the equilibrium level, OY^.
We now say that in situations such as the one illustrated in Figures 47. 1
(ii)

and 47.2 (ii) there exists a deflationary gap. We define this gap to be the
extent to which the volume of expenditure would fall short of the volume of income if
full employment income were achieved. An alternative statement is to say that the
deflationary gap measured by the extent to which the volume of with-
is

drawals would exceed the volume of injections if the full-employment

income were achieved.
Next consider Figures 47.1 (iii) and 47.2(iii). In this case the equilibrium
level of income is to the right of the full employment level. This level of
real income cannot be achieved. If we start at some level below OYp, then
income will rise. Once it reaches OYp, however, all resources will be fully
employed and further rises in output will be impossible. But expenditure
will be in excess of current income and this must cause prices to rise there ;

will be an inflation. A study of how the economy will behave in such periods
of inflation must be left to Part 8 but there can be no doubt that at income
OYp there will be inflationary pressure in the economy.
We now say that in situations such as the one illustrated in Figures
47.1 (iii) and 47.2(iii), there is an inflationary gap. We define this
gap

to be the extent which aggregate expenditure exceeds aggregate income when

to

full-employment income obtains. An alternative statement is to say that the

inflationary gap is measured by the extent to which the volume of injec-
tions exceeds the volume of withdrawals at full-employment national
income.
The inflationary gap measured in parts (iii) of both Figures by the
is

distance cd. The deflationary gap is measured in parts (ii) of both Figures
by the distance ab.
652 THE CIRCULAR FLOW OF INCOME

Balanced Versus Unbalanced Budgets

Not so many years ago it was generally accepted, and indeed many people

still prudent government should balance its budget on all

believe, that a
occasions The argument is usually based on an analogy with what seems
prudent behaviour for the individual It is a foolish person who lets his
current expenditure for such things as food, clothing and entertainment
consistently exceed his current revenue, so that he gets steadily further into
debt It was then argued that what is good for the individual must be good
‘
for the nation
When the government followed a balanced-budget policy, as most

governments did until the I930’s, it restricted its expenditure dunng a

depression because its tax revenue would necessarily be falling at that Ume
On the other hand, dunng a boom, when its revenue was high and nsmg
It increased its spending In other words it was rolling with the economy,

raising and lowering its expenditures in step with everyone else

To some people, the government, by going along with the crowd, did
not seem to be making the most of its potential to control the economy m a
beneficialmanner Why should the government not try to stabilise the
economy by doing just the opposite of what everyone else was doing - by
buying so to speak when everyone was selling, and selling when everyone
was buying^ This idea seemed particularly sensible in view of the fact that
government revenues and expenditures were such an important part of the
whole national income of the country In the United Kingdom, approxi
mately 35 per cent of the total national income is accounted for by the
government sector In the United States the figure is about 25 per cent The
potential of so large a class of expenditures to influence the economy seems
too obvious to ignore
As far as the government budget is concerned, taxes represent a with-
drawal from the circular flow of income, and expenditures represent an
injection into it If, m
times of depression, the government runs a budget
deficit, spending more than it is raising in taxes, then the government will

be adding to the arcular flow If the government spends more than it

raises, then there will be a net increase in the demand for resources If, on

the other hand, the government raises considerably more than it spends,
vt. vi'JJ. hie m \Vit foi lesomcts.
Thus we would expect a government deficit or surplus to have an effect
on the level of economic activity Furthermore, our multiplier analysis
allows us to predict that the final effect on the level of income will be
greater than the actual amount of the budget deficit or surplus, assuming it
to be a continuing one If the government spends million in excess of
1 Notice that this analogy doe* not usually hold at the firm level The debts of most successful
and rapidly growing firms will be found to be increasing steadily over time
 GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 653
revenues, this will have an effect similar to that of million new invest-
ment or any other kind of new injection of £X
million; it will start off a
multiplier process that will cause national income to rise by some
multiple
of^^ million.^
Trying to stabilise the economy by running deficits in periods of slump
and surpluses in periods of boom is called a countercyclical fiscal policy.^
We can describe countercyclical budgeting in its simplest possible terms
using Figure 47.1. Assume that the aggregate expenditure function is
shifting up and down due to changes in investment activity by firms. At
some times investment expenditure is very low and the expenditure function
is as illustrated in part (ii) of the Figure. At other times investment expendi-

ture is very high and the expenditure function is as illustrated in part (iii)
of the Figure. If the government does nothing to offset this, the economy
will suffer bouts of unemployment and depression alternating with bouts of
full employment and its simplest form counter-
inflationary pressure.^ In
cyclical budgeting government to attempt to remove both the
calls for the

deflationary and the inflationary gaps by running a deficit in situations such

as shown in part (ii) and surpluses in situations such as shown in part (iii)
of Figures 47.1 and 47.2. One possible way of doing this is to vary govern-
ment expenditure (keeping tax rates constant) so as to offset exactly the
fluctuations to the investment schedule, thus keeping the total expenditure
schedule constant.
In an ever changing and uncertain world, the policy is not as easy to
follow as might seem from the above discussion. Nevertheless, there is no
doubt that very large flcutuations can be damped down by the policy of
offsetting them rather than going along with them.

Built-in Stabilisers

The idea of countercyclical budgeting requires that each year a conscious

decision must be made to adjust the budget in a manner designed to
stabilise the economy. The political problems involved are too obvious to
need enumeration. There are time lags as well. To institute a change in the

government deficit takes time, even if there is no political problem. If the

1 This assumes that the government goes on with its deficit, for the simple theory of the
determination of national income spelled out in Chapter 42 predicts that, as soon as the deficit
the deficit spending
spending ceases, national income will fall back to where it was before
began.
2 Countercyclical budgeting independent of the question of the long-term change in size
is

of the debt. A countercyclical policy can be imposed

on a situation in which the long-term
trend of the debt is to rise, fall, or remain the same.
3 Notice that the losses suffered during depression
can never be made up. If resources lie
idle their potential output is lost for ever.
654 the circular flow of income

Cabinet proposes to the House of Commons tomorrow that the government

increases its road or school-buildmg programme by £10 million per year,
It will take a minimum of six months to a year before construction starts

even if the House passes the necessary l^islation the same day it receives

the request It will be longer yet before the construction is completed and
the government funds paid out full Cuts m m
taxes may work more quickly,
but here too lags between a change in tax policy and the reduced collections

may be substantial Where taxes are withheld, more rapid adjustment

well
IS possible than where taxes are payable at the end of a year or a quarter

Thus a cut in personal income taxes comes into play more rapidly than
does a cut in the business proBls tax
These problems suggest the idea of building certain automatic stabilising
devices into the system For instance, the government could commit itself
in advance to a fairly stable level of expenditure and to a policy of IctUng
revenues fluctuate over the cycle This would ensure deficits in recessions
and surpluses in booms
Steeply progressive taxes ensure that as incomes nse tax revenues nse
more than in proportion, and as incomes fall lax revenues fall more than m
proportion Since expenditures do not adjust instantaneously to revenue,
these tax»revenue changes tend to be stabilising
The mere fact that a much larger part of total investment is done by
governments than iti the past has a stabilising influence, for government,
investment is not nearly so volatile as is private investment Thus at least
part of this potentially very unstable element m
national income is made
relatively stable
The stabilising effect of government activity can be even further increased
if some part of government expenditure can be made to vary inversely with

income, rising when incomes fall and falling when incomes rise Important
built-in stabilisers that work in this way are unemployment benefits
schemes When incomes fall, expenditure on unemployment benefits nses,
when incomes rise, the expenditure falls Unemployment benefits ensure
that, when workers become unemployed, their consumption expenditure
will not fall to zero, since their disposable incomes will not fall to zero In
general, the higher the payment made to the unemployed in relation to the
Wrtycffn ‘iVitry caTTi wkien entfAtiyed, fne iimaklcr VneiaVi in V/ieir ctms-aTTiYA’/W-
when their employment falls

Another stabiliser is provided by agncultural policy When a slump

occurs and the demand for agncultural goods falls off, government price
support policies come into action and prevent farm incomes from falling as
drastically as they would without such aid Since agncultural income is
more or less stabilised,the expenditure of the agncultural sector is likewise
stabilised Thus, that part of the multiplier process that would have worked
 ;

GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 655

through the agricultural sector is frustrated by policies that automatically
support agricultural income. These stabilisers are important in
countries
such as the United States, in which agricultural sectors are large,
but are
unimportant in countries such as the United Kingdom, in which agri-
cultural sectors are small. ^
Most of these built-in stabilisers are the unforeseen by-products of policies
originally adopted for other reasons. The progressive income tax arose out
of a concern to make the distribution of income less inequitable. The
growth of the government sector has been the result of many factors other
than a desire for cyclical stability. Social-assistance and agricultural-support
programmes were adopted more because of a concern with the welfare of
the indi\dduals and groups involved than %vith preserving the health of the
economy. But, unforeseen or not, they w'ork. Even governments can be
lucky.

Dynamic Problems of Stabilisation Policy

All of the fiscal policies that stabilise the economy, whether they are auto-
matic or the result of conscious policy decisions, work to create negative
FEEDBACK. Negative feedback is a technical term that means that, when
any system deviates from its target level, forces are set in motion that push
the system back toward its when demand is too high so
target level.^ Thus,
that inflationary conditions prevail, demand is reduced; when demand is
too low so that unemployment prevails, demand is increased. It would be
grossly misleading to leave the reader with the impression that providing
negative feedback is sufficient to solve all stabilisation problems. Negative
feedback is a necessary but not a sufficient condition for stability.^ If any
control system operates with delays that are large relative to the period of
fluctuations it is seeking to control, it can do the very opposite of stabilising;
the controls can make the system less stable than it would otherwise be;
they can actually accentuate rather than check fluctuations.
Controls operate ^vith lags for two main reasons. The first reason is that
our knowledge of what is happening is always somewhat out of date. At a

1 This paragraph assumes that the government agricultural policy does more than merely
income from one group to another with the same propensity to
redistribute a given level of real
spend. While this is theoretically a possible situation, there is abundant empirical evidence that
that a fall in farm prices would other-
suggests that these policies do reduce the adverse eflTecl

wise have had on national income.

2 The system may be anything from an economy to an airplane using an automatic pilot
the target level may be anything from a set of equilibrium values to a set of values desired and
consciously sought after by the central authorities.
illustrated this point in the case of a single competitive
market. See page 102, note 1,
3 We
and the discussion of the cobweb theory in Chapter 12.
656 THE CtRCWLAR FLOW OF INCOME

minimum, it takes a month or so, and often very much longer, to gather
data about current happenings Our current information thus tells us not
what happening today, but what was happening anywhere from a month
IS

to SIXmonths ago The second reason is that it takes time for any policy
change to affect the beha\'iour of firms and households Such lags may vary
from a few weeks to a year, depending on the particular measure
A simple explanation of this problem can be developed along the
following lines Consider a system that is oscillating around a full employ
*

ment level of output as illustrated in Figure 47 3(i) Time is measured along

the horizontal axis, and on the vertical axis we measure the difference
between full employment output and current expenditure The fluctuations

Fii 473

arc such that a which aggregate expenditure estcetds aggregate

boom in
output is followed by a slump m which aggregate expenditure falls short of
aggregate output Assume that the government plans to vary its own
expenditure so as to offset these fluctuations exactly The government wants
Its plan to ha\e the impact pictured by the solid line in Figure 47 3(u) At

first, It plans a surplus that will reduce total expenditure, later, in period 5
It plans a deficit that will raise aggregate expenditure If the government s
plans are fulfilled instantaneously, then the addition of the solid lines m
Figure4? l{i) andFiguie47 l(n) will produce aggregate excess expenditure,
for the private plus the public sectors, always zero The govern-
which is

ment’s deficit or surplus exactly counterbalances the difference between

aggregate expenditure and full employment output in the private sector so
that Its stabilisation policy is completely successful
Now, assume that a time lag is involved, further assume, to make the
I This particular formulation of the problem u taken from
A W Ph 11 ps Fmi tovinrni
Inflation and Growth Economtea Februarv 1968
 GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 657
point as clear as possible, that the time lag is equal to half the
period of the
cycle. Now the planned balanced budget at period zero will
not actually
occur until period 2; the planned maximum surplus of period 1 will not
occur until period 3, when the economy is already in a slump; and the
planned maximum deficit of period 3 will not occur until period 5, when
the economy is already in a period of boom. Although planned government
expenditure still follows the solid line in Figure 47.1 (ii), actual government
expenditure now follo\vs the dotted line. Instead of stabilising the economy
as planned, the ‘stabilising’ policy actually destabilises it. The combination
of public and private demand wall give rise to larger fluctuations than
would have occurred if the government had done nothing!
This simple example is sufficient to show that the problem of controlling
the economy is not so simple as it sometimes seems. In general, policies
designed to stabilise individual markets or the whole economy will have
quite widely differing effects, depending on the time lags both in the actual
working of the economy and in the functioning of the stabilisation scheme.
Professor Frank Paish of the London School of Economics has compared
controlling the economy to driving a car with blackened front and side
windows and only a rear-view mirror from which to see. The car has brakes
and an accelerator, but they take effect only a long time after the driv'er
has used them. Thus the driver often has to have the courage to apply the
brake when he estimates that he is going uphill and the accelerator when he
estimates that he is going dowmhill, just as the government, in attempting to
influence the economy, often has to make plans to increase its spending
during a boom and to decrease it during a slump.

THE HYPOTHESIS OF SECULAR STAGNATION

So far we have been considering the appropriate governmental policy in
the case of cyclical ups and downs in the economy. A
much more drastic
problem arises if we have a continuous long-term tendency for withdrawals
to exceed injections at the full-employment level of national income. Assume
that, with international payments in balance and a balanced budget, there
is a continuous tendency for saving to exceed investment at the
full-employ-

ment level of national income so that there is a persistent deflationar)' gap

in the- economy. There is nothing logically contradictory in such a situation,
all that it requires is that the public should desire to
save more than

businessmen should wash to invest when the economy is at full employment.

It also requires that the gap between full-employment saving and full-
employment investment cannot be removed by driving down the interest
rate to a very low level. This situation could arise, for example, if as real
658 THE CIRCULAR FLOW OF INCOME

incomes over the years, households wished to save an ever increasing

rise

friction of their income, whereas businessmen did not wish to make invest
ment expenditure an cvcr»increasing fraction of national income If such
circumstances were to arise, and if the government insists on following a
balanced-budget policy, then there would be a continual tendency for the

economy to settle down in an equilibrium {with S=I) at less than full

employment Of course, occasional spurts of investment may sometimes
raise the economy to full employment, but if the average level of investment
IS that shown say, m Figure 47 l(ii) and 47 2(ii), the economy will tend lo

display a chronic tendency towards unemployment

If chronic unemployment did become a serious possibility, what could the
government do to remedy this situation^ Our national-income theory pro
vidcs an answer If the government departs from its balanced-budget policy
and adopts a policy of conlinuing budget defcilt, national income can be raised

to the full-employment level In this case, resources will not be lying

continuously idle, instead, they will be being used to create such public
goods as schools, roads, opera houses, univenities, defence weapons and a
host of other goods Clearly, the policy of continued budget deficits creates a
higher level of employment and a higher level of real income than does the
balanced budget policy Yet the national debt will be increasing year alter

year Surely, /Aw must matter*

To answer this question, let us consider what the government is doing It
IS borrowing that amount of full employment savings that private businesses
will noi borrow The total amount of borrowing is the same as it would have
been if full employment had come about because the volume of private

investment was higher by the amount ab in Figure 47 I(iO than the level
that It actually achieved All that has happened is that some of the money
saved is being channelled into the public sector The amount that is being
so channelled is the amount that the private sector is unwilling to utilise for

Us own purposes
There no reason why such a situation cannot go on forev cr There is no
is

reason why
the government cannot perpetually borrow and spend those
borrow Indeed, if the pnvate sector
savings that the private sector will not
will not borrow all the money the public wishes to save, the alternative to
budget deficits is that income and cmpIoYment will shrink and the surplus
savings will be removed by the reduction of income and employment to
below the full employment one that actual savings are
levels sufficiently
reduced to the same volume as that of the investment that the private sector
ISwilling to undertake
Conditions of a chronic excess of desired private saving over desired
pnvate investment have been labelled conditions of secular stagnation We
^e not saying here that such a situation is likely to occur the near future m
 GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 659
There has been a long debate in economics about the possibility
of such a
situation developing. By and large, current opinion is
that there is no evi-
dence that such a situation is imminent. The long-run estimates of the
consumption function (see pages 599-603) suggest that one of the main
postulates of the hypothesis may be factually incorrect. These estimates
suggest that, over the long run, the proportion of national income devoted
by households to current consumption and the proportion saved tend to
remain fairly constant. Furthermore, there is ample evidence that new
investment opportunities develop more or less as fast as old ones are utilised.
At least, they have so far.
It is important
to note, however, that such a situation could arise (i.e.,
there nothing logically contradictory in the hypothesis of secular stag-
is

nation; it describes a world that could exist). If it did arise, and if the govern-
ment had a balanced-budget policy, this pohcy would lead to a higher level
of unemployment and a lower level of real income than would a policy of
continued budget deficits.

WHAT IF PEACE BROKE OUT?

Arms expenditures are a significant part of total expenditures in almost
allWestern countries. They are highest in the United States where defence
and related expenditure amount to almost 10 per cent of total national
expenditures. It a basic tenet of the Marxists that these expenditures alone
is

bolster the American economy and that without them its collapse would
be sure and swift and would bring down with it the economies of most of the
other Western countries. What are we to make of this Marxist prediction
which was first mentioned in this book on page 3 ? What does the national-
income theory that we have developed predict about the effects of reductions
in defence expenditure in the United States?
It is perfectly clear that, if the need for defence expenditures fell from its
present 560,000 million 56,000 million (the level that they were in
to, say,

1940 in terms of 1964 prices), there would be problems. Much would de-
pend upon what else happened and how suddenly the change occurred.
If the government maintained its tax rates, reduced its expenditures by
554,000 million overnight, and allowed the extra receipts to pile up in idle
balances, our theory predicts that the results would be disastrous. With a
multiplier of, say, the reduction in national income would be on an order
3,

of 25 per cent, and unemployment might rise as high as 25 per cent of the
labour force.
Suppose, instead, that the government reduced both taxes and expendi-
predicts an adverse
tures equally by 554,000 million. Again, our theory
660 TilE CIRCULAR FLOW OF INCOME

impact on the economy ' The reduction in national income would be m the
order of 8 per cent, and iinemploymenl might rise to 11 or 12 per cent of
the labour force
These two situations of course do not exhaust the possibilities In a
country in which households desires arc unlimited and cannot be fully
satisfied because of resource limitations it is perfectly clear that peacetime
government expenditures could be substituted for defence expenditures
dollar for dollar if necessary Tlic Icsel of national income could thus be
maintained at the same time as the standard of living rose in response to

more goods per capita produced by the government

civilian
Another is to let private spending take over from public spending
policy
In the short run there would be dangers in returning too much money to
the private sector because the capacity to produce civilian goods would not
expand overnight factories that build rockets do not convert to civilian uses
in a hurry but could we make over time a phased reduction in both
taxes and expenditures with governmental deficits if needed tosmooththe
transition* The long run evidence about consumption functions and invest
ment opportunities suggests no fundamental problems in doing this The
evidence is that if consumers disposable income were raised by lOpercent
the long run elTcct would be for expenditure to nse by 10 per cent
The greatest problems would be transitional First there would be
problems in adjusting the productive capacity and the distnbulion of
resources geographically and among uses to a new set of demands made
upon them Given vvhat we know about immobility in the economy we can
predict that there would be surpluses m some skills and shortages in others
and that the adjustments would take time and would cause heavy fnctional
problems Tlie second problem v>ould be the lack of wisdom of policy
makers to say nothing of economic advisen ^\ e know much less about the
dynamic properties of our economic system than about its equilibnum
properties yet the short penod adjustments arc necessanly dynamic \Nith
out both wisdom and substantial discretion to act quickly many problems
would persist We have had litile experience of economics faced vvnth sodden
changes of such magnitude in peacetime Such changes m the pattern of
demand have however occurred dunng wan and the evidence seems to
suggest that as long as the level of aggregate demand is held high sub
stantial reallocations of resources can occur quite rapidly Although there
would be problems there i$ no evidence to suggest that they
transitional
could not be overcome within the space of a few years
1 The reajon for the advene effect ih s t me b deter
bed by ihe io*called balanced budget
mui pi er The problem ar set because the go eminent normally spends all s tax ncome
wh Ic households save some of the r inramea Thtu f the government cuts G and T by $ V

households expend ture w 11 nse by less than SJTso the total effect will be to reduce aggregate
expend ture
 GOVERNMENT AND THE CIRCULAR FLOW OF INCOME 661

Peace is, of course, unlikely to break out all at once. A gradual relaxation
of international tensions and a gradual reduction in required defence
expenditure would create less critical transitional The American
problems.
economy has often had experience with changes of S5,000 to 810,000 million
per year, and there seems litde doubt that increased public nondefence
expenditures combined with tax reductions could accommodate a gradual
shrinkage in the size of government spending Avith gain rather than loss in

standards of living.
 PART 8

MONEY, BANKING AND

THE PRICE LEVEL
 CHAPTER 48

THE NATURE AND HISTORY

OF MONEY

Up to this point we have taken the supply of money for granted. We have
assumed that there was sufficient money to finance any desired level of con-
sumption and investment expenditure. We now wish to study the generation
of money and credit and the possible effects of these factors on the circular
flow of income. A considerable portion of both consumption and investment
expenditure is made with borrowed money, and one would expect that the
availability and terms of credit would have an influence on both of
these
flows of expenditure. Also, a rise in output requires an investment of funds
in goods being manufactured but not yet ready for sale. If funds are not
available for the purpose, firms may be unable to expand output, even
though there is a rise in thedemand for their goods.*
In this chapter vve describe the functions of money and give a brief out-
line of its history. There is probably more folklore and general nonsense
believed about money than about any other aspect of the economy. The
purpose of this very stylised bit of history is to remove some of these miscon-
ceptions.^

WHAT IS MONEY?
We use the term money to refer to any generally accepted medium of
exchange — to anything, that is, that will be accepted by virtually everyone
1 Production does not occur instantaneously; a firm purchases raw materials and factor
services at one time and only at a later time docs it sell the goods that the factors produce. If
there is a rise in demand, the firm must increase its flow of expenditure on factor sendees and
raw materials; and only later, when the finished goods are sold, svill the firm’s flow of revenues
increase.
2 From the point of view of economic theor\', this historical analysis represents a digression.
My reason for including it is that a reader who holds one of the major misconceptions about

money understanding of the whole of macro-economic theory tvill be dis-

will find that his
torted. In my own experience, this particular approach provides one of the most effective
means of conveying an appreciation of some of the basic facts about money. .
666 MONEY, BANKING AND THE PRICE LEVEL

in exchange for goods and services It is customary to distinguish bctv,een

money’s several different functions The reason for doing this is that dif
ferent kinds of money \ar> in the degree of efficiency with which they fulfill
each of these functions, in assessing the efficiency of a particular money
system, we have to assess it in relation to each of these functions The major
functions of money are to act as a medium of exchange, as a store of wealth, and
as a umt of account

A MEDIUM OF EXCHANGE An important function of money « to facilitate

exchange Without money, our complicated economic system, which is
based on specialisation and the division of labour, would be impossible, and
we would have to return to a \ery primitive form of production and ex
change not without justification that money lias been called one of the
It is

great inventionscontnbutmg to human freedom

were no money, goods would have to be exchanged by barter, one
If there
good being swapped directly for another Tins system is a cumbersome one
m which every transaction requires a double coincidence of wants If I have a
donkey example I must search not only for a man who wants a
to trade, for
donkey, but also for one who has something that I Mould like to acquire
Furthermore, there is no way to give change on the transaction If I find
someone who wants my donkey, wc must agree on a swap rale If we decide
that a donkey is worth nine chickens, then wc arc m trouble if my trading
partner has only six chickens I can hardly give him tMO*thirds of my
donkey Thus, goods that are not readily divisible make poor subjects for
barter transactions If we were restncied to barter, we would have to spend
a great deal of time searching for saiisractory transactions, we would be
unable to specialise m
producing some single commodity, for we could not
be certain that wc could obtain when we wanted them all the other goods
that we required
The use of money as a medium of exchange removes these problems, as
long as money IS readily accepted by everyone If I wish to trade my donkey,
then all I is to find someone who wants a donkey I then hand
need to do
over take money m return, it matters not that the individual
my animal and
who my donkey has no goods that I require I now take my money
takes
and search for someone who has chickens that he wishes to trade When I
find him, I hand over my money and take his chickens, it does not matter
that my provider of chickens has no use for a donkey The difficulties of
barter force people to become more or less self sufficient, with money as a
medium of exchange, everyone is free to specialise and, with specialisation
m the direction of one's natural talents and abilities, there comes a great
increase in the production of all commodities

^ If money is to serve as an efficient medium of exchange, it must have a

 ^ ;

THE NATURE AND HISTORY OF MONEY 667

number of must be readily acceptable; it must have a

characteristics: it

high value for its weight (otherwise

it would be a nuisance to carry around)

it must be divisible, for money that comes only in large denominations

is
useless for transactions having only a small value; and it must not be readily
counterfeitable, for if money can be easily duplicated by individuals it will
lose its value.

A STORE OF wealth: Money is a handy way to store wealth; with barter,

one must take some other good in exchange. With money, you can sell goods
today and store the money until you need it. This means that you have a
claim on someone else’s goods that you can exercise at some future date.
The two sides of the barter transaction can be separated in time with the
obvious increase in freedom that this confers. To be a satisfactory store of
wealth, money must have a stable value. If prices are stable, then one knows
exacdy how much command over real goods and services has been stored
up when a sum
money has been accumulated. If prices change
certain of
rapidly, then one has litde idea how many goods one will be able to com-
mand when previously accumulated money is .spent. Clearly, rapid fluctua-
tions in the general level of prices reduce the usefulness of money as a store
of wealth.
Money can serve as a perfectly satisfactory store of accumulated wealth
for a single individual, but not for the society as a whole. If a single indi-
vidual stores up money, he will, when he comes to spend it, be able to
command the current output of some other individual. The whole society

cannot do If all individuals saved their money and all simultaneously

this.

retired to live on their savings, there would be no current production to

purchase and consume. The society’s ability to satisfy wants depends on
goods and seirdces being available; if some of this want-satisfying capacity
is to be stored up for the whole society, then goods that are currently
pro-

duced must be left unconsumed and carried over to future periods.

A UNIT OF account: Money may also be used purely for accounting

purposes without having any real physical existence of its o^vn. For instance,
a government store in a truly Communist society might say that everyone
had so many ‘dollars’ at his disposal each month. Goods could then be given
allowed
prices and each consumer’s purchases recorded, the consumer being
exhausted. This money
to buy all he wanted until his supply of dollars was
as entries in the store s books, but it
would have no existence other than

This point should be reconsidered after you have studied

Chapter 49. In the meanume,
1 last
the money can be counterfeited at \cry
you should try to understand why it does not matter if
high cost. _ P . .

sense in which money is a unit o

2 use of dollars in this context suggests a further
The
unit with which they are fami lar.
account. People think about values in terms of the monetary
 »

668 MONEY, BANKING AND THE PRICE LEVEL

\%ould be serving as a perfectly satisfactory unit of account (although it

could serve as a medium of exchange between individuals only if the store
agreed to transfer credits from one customer to another at the customers’
^
request)

THE ORIGINS AND GROWTH OF METALLIC MONEY

The origins of money are lost in antiquity, most primitive tribes known to

day make some use of it The abihty of money to free people from the
cumbersome necessity of barter must have led to its early use as soon as

some generally accepted commodity appeared All sorts of commodities

have been used as money at one lime or another, but precious metals must
soon have asserted their ascendancy as the most satisfactory ones They
were m heavy and permanent demand by the rich for ornament and decora
tion,and they were in continuous supply (since they do not easily wear out)
Thus they tended to have a high and stable pnee They were easily recog
nised and generally known to be commodities which, because of their stable
pnee, would be accepted by most people They were also divisible into
extremely small units (gold (o a single gram)
Precious metals thus came to circulate as money and to be used in many
transactions Before the invention of coins, it was necessary to carry precious
metals about in bulk When a purchase was to be made, the requisite
quantity of the metal would have to be weighed out carefully on a scale A
sack of gold and a highly sensitive set of scales were thus the common equip-

ment of the merchant and trader

Such a system, although better than barter, was still rather cumbersome
Coins eliminated the necessity of weighing the metal at each transaction
The prince or ruler weighed the metal and made a com out of it to which he
alKxed his own seal, guaranteeing the weight of precious metal m the com
Thus a certain com was stated to contain exactly of an ounce of gold K
^
a commodity was priced at ^ of an ounce of gold, two coins could be given
over without any need of weighing the gold This is clearly a great con-
venience, as long as traders know that they can accept the coin at its ‘face

value’
The prince’s subjects, however, could not let a good thing pass, and some-
one had the idea of clipping a thin slice off the edge of the com If he collected
a com stamped as containing J of an ounce of gold, he could clip a slice off
the edge and pass the com off as still weighing the
I of an ounce (‘Doesn’t
I A founh function w jomctimw djtingniihcd that of a iiandard of deferred payments
Payments that are to be made in the future, oo account of debts etc , arc reckoned in money
Money is being used as a unit of account with an added dimension m time for the account
settled until the future
 ,

THE NATURE AND HISTORY OF MONEY 669

Stamp prove he would argue.) If he were successful, he would have

it?’,

made a profit ec^ual to the market value of the clipped metal. If this practice
became common, even the most myopic of traders would notice that things
were not quite what they used to be in the coinage world. Mistrust would
grow, and would be necessary to weigh each coin before accepting it at
it

its ‘face value’ back would come the scales and most of the usefulness of
;

the coins would be lost. To get around this problem, the princes began to
mint their coins with a rough edge. The absence of the rough edge would
immediately be apparent and would indicate that the coin had been
clipped. This practice still survives today as an interesting anachronism to
remind us that there were days when the market value of the metal in the
coin (if it were melted down) was equal to the face value of the coin. The

coin itself was nothing more than a guarantee that a certain weight of
metal, the value of which did not depend on its being stamped into coins,
was contained therein. The subjects, when presented with an opportunity
of getting something for nothing, were ingenious enough to surmount even
the obstacle of the rough edge; they invented the practice oi sweating. Sweat-
ing involved placing a large number of coins in a bag and shaking the bag
vigorously. The dust that flaked off the coins was their reward. This practice
seems never to have been as disruptive to the money system as that of clip-
ping, possibly because it was difficult to remove very much metal without
defacing the coin, but possibly also because the disruptive effects were
eclipsed by the upset caused by the prince’s periodic debasement of the
coinage.
Not be outdone by the cunning of his subjects, the prince was himself
to
quick to seize the chance of getting something for nothing. Since the prince
was empowered to mint the coins, he was in a very good position to work a
really profitable fraud. When he found himself with bills that he could not

pay and that it was inexpedient to repudiate, he merely used some suitable

occasion — a marriage, an anniversary, an alliance - to remint the coinage.

The subjects would bring their coins into the mint, where they would be
melted down and coined afresh with a new stamp. The subject could then
go away the proud possessor of one new coin for every one old coin that he
had brought in. Between the melting down and the recoining, however,
the prince had only to toss some inexpensive base metal into the works to
earn himself a handsome profit. If the coinage was debased by adding, say
one part base metal to every four parts of the melted-down metal, then five
coins
coins would be made for every four turned in. Thus, for every four
four and have one left over for himself
brought in, the prince could return
as profit. With these coins, he could pay his bills.
The result would be an inflation. The subjects would have the same num-
ber of coins as before and hence could demand the same number of goods.
670 MONEY, BANKING AND THE PRICE LFVEL

Once the pnnce paid his bills, however, the recipients of the extra coins
could be expected to spend some or all of them, and this would represent a
net increase in demand The extra demand would bid up prices Debasing
of coinage thus led pretty certainly to a nse in prices After observing this
process in action, early economists propounded the quantity theory of money
They argued that there was a relation between the average level of pnees
and the quantity of money in circulation, such that a change in the quantity
of money would lead to a change in the pnee level in the same direction
We shall have more to say about this theory m Chapter 51

THE EVOLUTION OF PAPER MONEY

Another important step in the history of money was the evolution of paper
currency Goldsmiths - craftsmen who worked with gold ~ naturally kept
very secure safes m
which to store their gold ‘ The practice grew up among
the public of storing their gold with the goldsmith for safe-keeping In
return, the goldsmith would give the depositor a receipt promising to hand
over the gold on demand If the depositor wished to make a large purchase,
he could go to the goldsmith, reclaim his gold, and hand it over to the stUn
of the goods Chances were that the seller would not require the gold but
would carry it back to the goldsmith for safe-keeping Clearly, if people
knew the goldsmith to be reliable, there was no need to go through the

cumbersome and nsky business of physically transferring the gold The

buyer need only transfer the goldsmith’s receipt to the seller, who could
accept It secure in the knowledge that the goldsmith would pay over the
gold whenever it was needed If the seller wished to buy a good from a third
party who also knew the goldsmith to be reliable, this transaction too could
be effected by passing the goldsmith’s receipt from the buyer to the seller
The convenience of using these bits of paper instead of gold is obvious Thus,
when paper money first grew up it represented a promise to pay on demand
so much gold, the promise being made first by goldsmiths and later by
banks, as long as these instituDons were known to be reliable, such pieces
of paper would be ‘as good as gold’ Such paper money was backed by
precious metal and was convertible on demand into this metal
In the nineteenth century, paper money was commonly issued by banks,
in Bntain, the commercial banks issued their own notes backed by their
own reserves Also, the Bank of England issued notes backed by the coun
try’s gold reserves, for, although it was nominally a pnvate institution until
1947, the Bank always had close links with the government Since th«c
1 paper money can be displayed by concentrating on the gold
All the basic ideas about
smiths although there were earlier sources of paper money in various negotiable evidences
fdebt.
 THE NATURE AND HISTORY OF MONEY 671

notes were convertible on demand into gold, the country was said to be
on a gold standard.

Fractionally backed paper money: For most transactions, individuals

were content to use paper currency; it was soon discovered, therefore, that
it was not necessary' to keep an ounce of gold in the vaults for every
claim
to an ounce circulating as paper money. It was necessary' to keep some gold
on hand, because, for some transactions, paper would not do. If an indi-
vidual wished to make a purchase from a distant place where his local bank
was not known, he might have to convert his paper into gold and ship the
gold. Further, if he was going to save up money for use in the distant future,
he might not have perfect confidence in the bank’s ability to honour its
pledge to redeem the notes in gold at that time. His alternative was to
exchange his notes for gold and store the gold until he needed it. For these
and other reasons at any one point in time, some holders of notes would be
demanding gold in return for their notes. On the other hand, some of the
bank’s customers would be receiving gold in various transactions and would
be wanting to store this gold in the bank for safe-keeping. They would
accept promises to pay (i.e., bank notes) in return. At any one time, there-
fore, some of the bank’s customers would be withdrawing gold, others would

be depositing it, and the great majority would be carrying out their trans-
actions using the bank’s paper notes without any need or desire to convert
them into gold. Thus the bank would be able to issue more money redeem-
able in gold than it actually had gold in its vaults. This is a profitable thing

to do, because the money can be used to make profitable investments. It

can be used either purchase securities that yield a return, or to make
to
interest-earning loans or advances to households and firms.
This discovery was made early on by the goldsmiths from that time down
;

to the present day, banks have had many more claims to pay cash outstand-
ing against them than they actually had cash available. In such a situation,
we say that the currency is fractionally backed by gold. A rough rule of thumb
is that a per cent backing for these claims is more than sufficient: if a
10
bank holds 10,000 worth of gold and has issued ,7; 100,000 in notes, it
would be perfectly safe in normal dmes and would be able to convert into
gold all of its notes that were presented for conversion.
The major problem of a fractionally backed currency is that of maintain-

ing its convertibility which it is backed. In the

into the precious metal with
past, the imprudent bank that issued too much paper money found
itself

unable to redeem its currency in gold when the demand for gold was even
slightly higher than usual. This bank would then have to suspend payments,
and all holders of its notes would suddenly find them worthless. The prudent
bank, which kept a reasonable relation between its note issue and its gold
672 MONFY, BANKING AND THE PRICE LEVEL

reser\c, found that it could meet the normal everyday demand for gold

without an> trouble It was always the case with fractionally backed

currency, however, that, if all note-holders demanded gold at once, they

could not be satisfied Thus, if c\cr the public lost confidence and cn masse
demanded redemption of their currency, the banks would be unable to
honour and the holders of their notes would lose eicrjthing
their pledges,
The history of nineteenth- and early IwenUcth-ccntury banking on both
sides of the Atlantic is replete with examples of banks ruined by momentary
runs on their cash and gold reserves When this happened, the bank’s de
positors and the holders of its notes would find themselves holding worthless
pieces of paper Future social historians may wonder how it was possible,
such a system, that economists could have believed that frcc-
in the face of
market capitalism provided evidence that the hidden hand of perfection
was guiding the economic affairs of mankind

The development of fiat currencies A.% time went on, note issue by

commercial banks became less common, and central banks took over a
steadily increasing share of this responsibility The paper currency was, as
Italways had been, freely convertible into gold It was also only fractionally
backed by gold The commercial banks retained the power to create money,
but this was no longer done by printing paper money, instead, deposit
money was created
During the period between the two World Wan, virtually all the coun
tnes of the world abandoned the gold standard Tlie reasons for this arc
historical and cannot be gone into here {They were mentioned briefly in
Chapter 55 The result of abandoning the gold standard vvas that currency
)

was no longer convertible into gold

Some countries (including the United States, until 1965) preserve the
fiction that their currency is backed by gold, but none allow's it to be con-
verted into gold as a right The past is recalled by the following statement
on the British pound note, ‘I promise to pay to the bearer on demand the
sum of One Pound [signed] Chief Cashier, Bank of England’ (A similar
statement appears on the currency of most countnes ) If anyone takes this
seriously today and demands his ‘One Pound’, he can hand over his one
pound note and receive in return a new one pound note'^ln fnc daysxfi'dft
gold standard, paper money was valuable because everyone knew it could
be converted into gold on demand ^ Today, paper money is valuable be
cause it IS generally accepted Because, by habit, everyone accepts it as

1 Originally the gold market value independent of lU use as money Later howeter
had a
as large stocks of gold accumulated gold itself came to have value tecaust it was a generally
acceptable medium of exchange There is little doubt that if at any time in the last few
centuries gold had ceased to have value aa money (because say some supetvor metal was
 THE NATURE AND HISTORY OF MONEY 673

valuable, it is valuable; the fact that it can no longer be converted into any-
thing has no effect on its functioning as a medium of exchange.
This fact, that present-day paper money is not convertible into any-
thing - that nothing but bits of paper whose value derives from common
is is

acceptance through habit - often disturbs the student. He feels his money
should be more substantial than that; after all, what of ‘dollar diplomacy’
and the ‘prestige of the pound’? Well, his money is only bits of paper. There
is no point in pretending otherwise.

Once it is accepted that modern money is only bits of paper, the next
question that comes to mind is: does it matter? Gold derived its value
because was scarce relative to the heavy demand for it (the demand being
it

derived from its monetary and its nonmonetary uses). Tying a currency to
gold meant that the quantity of money in a country was left to such chance
occurrences as the discovery of new gold supplies. This was not without
advantages, the most important being that it provided a check on the
prince’s ability to cause inflation. Gold cannot be manufactured at wall;
paper currency can. There is little doubt that, if the money supply was
purely paper, many governments would have succumbed to the temptation
to pay their bills by printing new money rather than by raising taxes. Such

increases in the money supply would lead to inflation in just the same way
as did the debasement of metallic currency. Thus, the gold standard pro-
vided some check on inflation by making it difficult for the government to

change the money supply. Periods of major gold discoveries, however,

brought about inflations of their own. In the sixteenth century, Spanish
gold and silver flowed into Europe from the New World, bringing inflation

in its wake. On usually desirable to increase the money

the other hand, it is

supply in a period of rising trade and income. On a gold standard, this

cannot be done — unless, by pure chance, gold is discovered at the same
time. gold standard took discretionary powers about the money supply
The
out of the hands of the central authorities. Whether or not one thought that
authorities
this was a good thing depended on how one thought the central
probably better
would use this discretion. In general, a gold standard is

irresponsible govern-
than having the currency managed by an ignorant or
ment, but it is worse than having the currency supply adjusted by a
well-
by
informed intelligent one. Better and worse in this context are judged
the needs
the criterion of having a money supply that varies adequately with
as to cause violent inflations or
of the economy, but that does not vary so
deflations.

have fallen very ^eat y, sin

discovered), the market value of the world’s stocks of gold would
been very low in re ation
the demand for gold for use as a commodity would have
supply of it.

22
674 MONEY, BANKING AND THE PRICE LEVEL

DEPOSIT MONEY
In most countries today, the money supply consists of notes and coins issued
by the government and the central bank, and of deposit money Notes and
coins (the market value of the metal m
the coinage is but a minute fraction
of the face value of the coin) are the inconvertible moneys that we haie
already discussed Deposit money is created by the commercial banbng
system We have already explained how bank notes promising to pay gold
on demand circulated as money and how, because most people did not
require the actual gold, the banks were able to create money by printing
and putting into circulation many more notes than they could redeem in
gold at any one time When Vie banks lost the right to issue notes of their own, the
form of money creation changed but the substance did not Today, banks have money
m their vaults (or on deposit with the central banks) just as they always did,
only the money is no longer gold, it is the legal tender of the times, paper
money Banks’ customers sometimes deposit paper money with the banb
for safe keeping just as, in former times, they deposited gold The bank takes
the money and gives the customer a promise to pay it back on demand
Instead of taking the form of a printed bank note, as in the past, this promise

to pay recorded as an entry on the customer’s account If the customer

IS

wishes to pay a bill, he may come to the bank and claim his money m
pound notes, he may then pay the money over to another person, and this

person may redeposit the money with the gold transfen,

in a bank Just as
this is a cumbersome procedure, particularly for large payments, and it
would be much more convenient if the bank’s promise to pay cash could
merely be transferred from one person to another This is done by means of
a cheque If individual A deposits X^IOO in a bank, the entry, a ;(^100 credit
in his account, the bank’s promise to pay £100 cash on demand If^ pays
is

B £100 by giving him a cheque that B then deposits in the same bank, the
bank merely reduces A'i deposit by £100 and increases B’s by the same
amount Thus the bank still promises to pay out on demand the £100
originally deposited, but it now promises to pay it to B rather than to A
IfB now paysC£100 by cheque andCdeposits the cheque, then the promise
to pay (i e the credit entry in someone’s account at the bank) will be
,

transferred from B to C
The modern deposit is the equivalent of the old bank note a promise on
the bank’s behalf to pay out on demand the money of the time The passing
of the bank’s note from hand to hand transferred ownership of the claim
against the bank, this is now done by means of a cheque, which is merely

an order bank telling it to transler from one individual to another its

to the
obligation to pay cash It is true today, just as it was m the past, that most
of the bank’s customers arc content to pay their bills by passing among
themselves the bank’s promises to pay cash, only a small proportion of the
 THE NATURE AND HISTORY OF MONEY 675

bank’s customers want cash at any one time. Thus, just as in the past, the
bank can create money by issuing more promises to pay than it actually has
cash to pay out. The bank can grant a loan by giving the customer a credit
on his account equal to the amount of the loan. If the borrower uses the loan
topay bills by cheque, then the deposit is transferred from person to person.
In most circumstances, the bank can have liabilities greatly in excess of the
amount of cash that it has in reser\'e. These deposits can be used to buy
goods and services through the medium of cheques. Since they are a
generally accepted means of exchange, they are money. The great pro-
portion of transactions (by value) take place by cheque and only a small
proportion by notes and coin. Thus, in the modem world, the greater pro-
portion of the money supply is the deposit money that is created by com-
mercial banks. The banks can, if they wish, contract the money supply by
not creating deposits, or they can expand by creating deposits up to the
it

limit of prudence or law

(so that there is Just enough cash to meet the normal
demands of customers who do not wish to pay by cheque) It is, of course,.

in the bank’s interest to expand the supply up to the safety limit because
every pound created can be used to grant a loan, to purchase a bond, or
to acquire some other asset that pays a return to the bank.

Demand deposits and time deposits: If a customer has a deposit in a

bank, he can keep it in one of two forms; in a current account or a deposit
account. Current accounts are often referred to as demand deposits and
deposit accounts as time deposits. The holder of a current account deposit
may immediately obtain cash for it without giving any notice of his intention
and he may pay his bills by writing a cheque on it. Such a cheque instructs
his bank to pay without delay a stated sum of money to the person to whom
the cheque is made payable. Banks in most countries do not pay any interest
to customers who have money deposited in current accounts.
holder of a deposit account must legally give notice (thirty or sixty
The
days) of his intention to withdraw his money. Although banks often do not
enforce this law, they could at any time do so if they wished. Furthermore,
the holder of a deposit account cannot pay his bills by writing a cheque
ordering his bank to pay someone out of his time deposit. Banks usually pay
interest to customers who have money deposited with them on time
.

AN OPERATIONAL DEFINITION OF MONEY

Earlier in this chapter we defined money as any generally accepted means
what
of payment. We must now make this definition operational by saying
will be included and what excluded when we reckon the money supply of

a country. Clearly, notes and coins are part of the money supply. Current
account deposits also fit pretty well into our definition. You can pay for most
676 MONEY, BANKING AND THE PRICE LEVEL

things by cheque, although cheques are not quite so generally acceptable as

notes and coins You may have trouble, for example, buying a pacltage of
cigarettes in a small country store if you ofTcr a cheque in exchange, you
w ill almost certainly has c trouble should > ou try to walk out of a fur store
with a £2,000 mink coat if you have offered a cheque in payment and you
are not knovm by the store manager You will not have much trouble ifyou
offer pound notes in cither case Since demand deposits are a means of
exchange and since cheques are widely (if not quite universally) acceptable,
It seems reasonable to regard demand deposits as part of the money supply

Time deposits are not so regarded since cheques cannot be drawn on them
and since, should the bank insist on its l^al rights, the deposits canpot be
turned into cash with which to make payments until after a lapse of time
(the length of which depends on the period of notice) The money supplyis
therefore, usually said to consist of three parts, metallic currency (coins),
paper money (notes), and demand deposits

Near Money
Although it IS not conventional to include deposit accounts in the money
supply, they are clearly %ery close to being money In practice, they can be
turned into money without notice merely by out a withdrawal slip at
filling
one’s bank We thus talk of deposit accounts as being ‘near money’ - some
thing that is almost but not quite ’ Once we have developed the con
cept of near money, we realise that there arc ‘liquid assets’ other than
deposit accounts, which, although themselves not generally acceptable as
means of exchange, are easily convertible into money Such liquid assets
include certain short-term government sccunties, deposits with savings and
loan societies, and a host of other assets that arc readily convertible into
money The phenomenal nsc in the US, and recently in the UK, of credit
cards has helped to make ‘trade credit’ an important near money in these

countries
There is no need to get into an argument about the definition of money
The way in which we wish to define money and near money depends on the
theory we are developing For many purposes, a narrow definition that
includes only currency and demand deposits is useful On the other hand,
near moneys may m
many cases be virtually perfect substitutes for money
The matter is an empirical one will a theory give generally better predic-
tions usingone concept of money rather than another^ If so, the theory
should be formulated using the relevant concept There is no right and
wrong about definitions, and the relevant questions to ask about them are
‘Are they consistent with each other’’ and ‘Do they define classes of things
that will be useful in relating our theones to real-world observations’’
CHAPTER 49

THE BANKING SYSTEM

AND THE SUPPLY OF MONEY

Our primary concern in this chapter is with the factors that determine the
supply of money. In order to study these, we must look at the nature of the
banking system, at the way in which banks create deposit money, and at
theway in which attempts are made to regulate the money supply through
The principles involved are the same in all Western countries
public policy.
although the institutional arrangements through which these operate do
differ significantly from one country to another.

THE BANKING STRUCTURE OF THE UNITED KINGDOM

Today, there are five major commercial banks in the United Kingdom
and another six of significant size. Each bank has innumerable branches
throughout the country. In the United States, on the other hand, there are
approximately 14,000 separate banks each with only a small number of
branches. These commercial banks accept deposits, make loans, and
provide a variety of services to their customers. They are privately owned
profit-seeking corporations. They accept savings deposits (often called dme
deposits) and checking deposits (often called demand deposits). There are
also a variety of other kinds of private financial institutions, including
savings banks, building societies, finance companies, and other institutions
that accept only time deposits.
The basic unit of the banking system is the ordinary commercial bank.
In the UK there are five large commercial banks that dominate the whole
banking system Midland, Barclays, Lloyds, National Provincial and West-
:

minster, and a dozen or so much smaller banks. The second main element
in the banking system is the Bank of England which is the central bank. We
shall first consider the activities of the commercial banks and then consider
the central bank.
678 MONEY, BANKING AND THE PRICE LEVEL

THE CREATION AND DESTRUCTION OF DEPOSIT MONEY

BY COMMERCIAL BANKS
One part of the money supply, deposit money, is under the control of
pnvately-owned banks How do these banks create deposit money’ Wc
have already seen that the ability ofbanks to create deposit money depends
on the fact that bank deposits need to be only fractionally backed by notes
and com If all deposits had to be backed 100 per cent, banks ivould be
nothing more than safety deposit vaults for their customers’ money If a
customer deposited ;C100 in a bank, that bank would give the customer
;{^100 credit on his deposit account, thej(^100 would go into the bank’s vault
to ‘back’ the deposit and nothing further would happen Because the bank
does not need to keep 100 per cent reserves, it can use some of the £100 that
was deposited to purchase income-yielding investments Just how does the

bank do this’

A Single Bank
Consider first a country with only one bank (with as many physical branches
as IS necessary) and assume that someone makes a new deposit of £100 in
cash Table 49 I shows how this transaction will be recorded on the books

Table 49 1

A NEW DEPOSIT OF £100 IS MADE

Assets Lsahibtits

Cash £100 Deposit £100

of the bank The balance sheet will show new assets of £100 in the form of
cash and new liabilities of £100 m the form of the customer’s deposit This
deposit, and all it, is a liability of the bank, since the bank
others like owes
this money customer and must pay it to him whenever he demands it
to the
Since there is only one bank in the whole amntry, the bank can immediately
create new deposits by some multiple of £100, depending on the legal
reserve requirements of the country ‘ Let us say that the law requires a
10 per cent reserve The bank could immediately create further deposits of
£900 Assume, by way of example, that the bank loans £500 to a customer
and buys £400 worth of bonds m the open market The bank does this by
1 If therewere no legal reserve requirements the bank would be able to create deposits up
to the ratio proved prudent by experience In most countries the banks are required by law
to hold reserves to an amount larger than would be dictated by normal prudence Thus the
effective limit on deposit creation is set by law not by experience
 ;

THE BANKING SYSTEM AND THE SUPPLY OF MONEY 679

permitting the borrower to write cheques on his account to the amount of

;^500, and by writing ;^400 to the credit of the account belonging to the
person who sold bonds to the bank. Table 49.2 shows how these transactions

Table 49.2
£900 IS INVESTED INLOANS AND
BONDS WITH NO CASH DRAIN

Assets Liabilities

Cash £100 Deposits £1,000

Loan £500
Bonds £400

will appear on the bank’s books once the borrower has written cheques to
the allowable amount. The bank’s assets include the /^lOO cash of the
original deposit, the loan of £500 (it is an asset of the bank, since the
borrower owes this money to the bank and must repay it at some stated
date), and the bonds to the value of ,(^400 (these are an asset, since they can
be sold again for cash). The bank’s liabilities are now 1,000 in deposits,
£100 to the account of the original depositor, £500 to the account of the
persons who have received payment from the customer who borrowed from
the bank, and £400 to the account of the person who sold the bonds to the
bank. Note that by a few strokes of the pen the bank has created ;0900 in
deposit money. The customers of the bank are now able to spend ^(^900 more
than they could yesterday and no one else is forced to spend any less.
The persons who borrowed the money and sold the bonds can be expected
to spend their money. In most cases, they will do so by writing
cheques. The

bank honours these cheques by reducing the deposit of the person writing
the cheque and increasing the deposit of the person in
whose favour the
cheque is written. Thus, if everyone pays by cheque, the bank can effect
individual customers
these payments merely by changing the accounts of

no cash ever leaves the bank and the total of the bank s deposit liabilities

does not change.^

are observed to hold some of their

1 This introductory' treatment ignores the fact that people
per cent of their money n
money in cash and some in deposits. If, for example, people hold 10
cash drain when it creates new money. This cash
the form of cash, then the bank can expect a
drain would mean that deposits could not safely be
expanded by £900. If such an «pan^n
£810 in new deposits. (This
were to occur, £90 would be withdrawn by the public, leaving
per cent of their new money would be held m cash
would leave the public satisfied, since 10
cash and £90 m deposits.
and 90 per cent in deposits, but the bank would have lost £90
in
680 MONEY, BANKING AND THE PRICE LEVEL

Many banks, a single new deposit The whole system is somewhat

more complicated when there are many banks If a depositor in Bank A
wntes a cheque to someone in Bank B, then a mere book transfer will not do,
because Bank A now owes money to Bank B By wnting the cheque, the
depositor in Bank A is saying, ‘1 claim the money owed me and ask thalti
be passed over to the man indicated on the cheque’ when the recipient of ,

the cheque deposits it Bank m

he is saying, ‘I want my money held for
me by Bank B thus Bank A must pay the money over to Bank h It is
'
,

exactly the same as if one individual withdrew cash from Bank A and gave
It to thesecond individual who deposited it in Bank B, when the transaction
is done by cheque, however, the banks, rather than the individuals, transfer
the money
There are, of course, many such transactions in the course of a day If the
banks are staying the same size in relation to each other, these transfers
between banks will tend to cancel each other out If, for example, Mr
Brown who banks with A gives a cheque for ,^100 to Mrs Smith who banks
with 5, and same time, Mrs Jones who banks with B gives a
if, at the
cheque for Green who banks with A, then these two trans-
actions cancel each other out Bank A loses ^100 to B on account of the
first transaction, but gams a like sum from B on account of the second No

money need move from bank to bank all that needs to happen is for A to
,

reduce Mr Brown’s account by X^IOO and increase Mr Green’s by the same

amount, and for B to reduce Mrs Smith’s account by and raise Mrs
Jones’s by the same amount
Multibank systems make use of a clearing house where interbank
debts are cancelled out At the end of the day, all of the cheques drawn by
Bank ^’s customers and deposited in Bank B are totalled and set against the
total of all of the cheques drawn by Bank B’s customers and deposited in
Bank ^4 It is only necessary to settle the difference between these two sums
The actual cheques are passed through the cleanng house back to the bank
on which they are drawn The bank is then able to adjust each individual s
account by a set of book entries, a flow of cash between banks is necessary
only if there is a net transfer of cash from the customers of one bank to those
of another
m
What would happen if, a multibank system, one bank received a new
deposit of;^100 in cash’ In this case, the bank could not immediately create
another ^{^900 m
deposits because, when cheques were written on these
deposits, the majority would be deposited in other banks Thus, the bank

leaMtig It with £10 m cash and ,^10 in depouu which gives it a cash deposit ratio of only
1 23 per cent ) There does a smaller deposit expansion that will leave the banks with a
exist
10 per cent cash reserve, in spite of the dram of cash to the public
 THE BANKING SYSTEM AND THE SUPPLY OF MONEY 681

must expect much of its £100 in cash to be drained away to other banks as
soon as it creates new deposits for its owm customers.
If the bank that obtains the new deposit has only 10 per cent of the total
deposits held by the community, then 90 per cent of any new deposits it

Table 49.3
BANK A EXPANDS DEPOSITS AS FAR AS POSSIBLE
WHILE OTHER BANKS DO NOT
Bank A

Assets Liabilities
1

Cash ;,^10-98 Deposits ;^1 09-89

Loans ;()98-91

/:i09-89 /:i09-89

A// other banks in the system

Assets Libiliiies

New cash ,f)89-02 New deposits ;089-02

creates will end up in other banks.^ If other banks are not simultaneously
creating new deposits, then this one will be severely restricted in its ability

to expand deposits. The reason for this restriction is that the bank will suffer

a major cash drain as cheques are written to the favour of individuals who
deal with other banks.
If the bank illustrated in Table 49.1 was only one bank in a system (say
with one-tenth of the total deposits in the system) and if the other banks
refused to expand deposits, then the final situation would be as illustrated
in Table 49.3.
Whathas happened between Tables 49.1 and 3 is that the bank has
created j098-91 deposits by granting loans.^ But 90 per cent of these have
ended up in other banks, so that the original bank ends up with only
;010'98 of the new deposits held by its own customers. The remainder is

1 In general, bank has X per cent of the total value of deposits in the community,
if the
then 100 — X per cent of any newly created deposits can be expected to end up in other banks.
of
2 Throughout the following calculations we shall deal for simplicity in decimal fractions
pounds and will not convert these into shillings and pence.

22 *
682 MONEY, BANKING AND THE PRICE LEVEL

distributed the other banks The Wtal increase

among deposits is onlym
98 9 per cent of the original cash deposit, as opposed to 900 per cent in the
monopoly-bank case illustrated in Table 49 2 This leads us to the following
conclusion

One bank in a multibank system cannot produce a

multiple expansion of deposits based on an original
accretion of cash, unless other banks also expand
deposits.

If all other hanks arc willing to expand deposits whenever they gam
extra cash, the situationshown in Table 49 3 will not represent an cqui
librium position All other banks in the system will have excess cash the
;C89 02 in new cash and new deposits uill be spread about equally among
them Thus all banks will have their new deposits backed 100 per cent by
cash This should lead all banks to expand deposits simultaneously and
produce a different situation from the one shotsm in Table 49 3, since there
will no longer be a cash dram from one bank to another

Many banks, man% new deposits Assume that, in a system with many
banks, each bank obtains new deposits m cash {possibly because ofa general
increase in themoney supply due to a change m
govcmmcnl policy) ‘ Say,
forexample, that the community contains ten banks of equal size and that
each receives new deposits of/^lOO m
cash Now each bank is the position m
shown in Table 49 I and each can begin to expand deposits based on the
;£100 of reserves (each bank does this by granting loans to customers and
by buying bonds and other income-earning assets) Since each bank does
one-tenth of the total banking business, an average 90 per cent of any newly
created deposit will findits way into other banks as the customer pay’s by

cheque people in the community This will represent a cash drain

to other
to these otherbanks On the other hand, 10 per cent of each new deposit
created by each other bank should find its way into this bank Thus, if all
banks receive new cash and all start creating deposits simultaneously no ,

bank should suffer a significant cash drain to any other bank Thus all
banks can go on expending deposits without losing cash to each other, they
need only worry about keeping enough cash to satisfy those depositors who
occasionally require cash Thus the expansion can go on, with each bank
watching its own ratio of cash reserves to deposits, expanding deposits as

long as the ratio exceeds 1 10 and ceasing when it reaches that figure The
process will come to a halt when each bank has created ;^900 in additional
deposits, so that, for each initial £100 cash deposit, there is now £1,000 m
l The ways m which such changes in the money supply can be effected are discussed later

m this chapter
 THE BANKING SYSTEM AND THE SUPPLY OF MONEY 683

Table 49.4
EXPANSION OF CREDIT IN EXPECTATION OF A
90 PER CENT CASH DRAIN TO OTHER BANKS WHEN
NO CASH DRAIN ACTUALLY OCCURS
Assets Liabilities

Cash £100-00 Deposits £198-91

Loans £98-91

£198-91 i
£198-91

deposits backed by 100 in cash. Now each of the banks will have entries in
its books similar to those shoivn in Table 49.2.
We can think of this process taking place in a series of steps. During the
first day, each bank gets ;(^100 in new deposits and the books of each bank

show entries similar to those shown in Table 49.1. During the second day,
each bank makes loans, expecting that it will suffer a cash drain on account
of these loans. Indeed, 90 per cent of the new loans made by Bank A do
find theirway into other banks when the borrower pays money to people
who are customers of other banks, but 10 per cent of the new loans made by
each other bank finds its way into Bank A as those borrowing from other
banks pay money to people who are customers of Bank A. Thus, there is no
net movement of cash between banks. Instead of finding itself in a position
such as that shown in Table 49.3, each bank’s books at the end of the day
contain the entries shown in Table 49.4.
Cash is now just over 50 per cent of deposits instead of being only 10 per
order
cent as is desired. Thus each bank can continue to expand deposits in
income-earmng assets. As long as all banks
to grant loans and to purchase
any
do this simultaneously, no bank will suffer any significant cash drain to
other bank, and the process can continue until each bank has created £900
in Table
worth of new deposits and then finds itself in the position sho%vn
49.2.^
which one bank creates a deposit on the basis of an
accre-
1* Textbooks often take a case in
second bank then creates deposits,
tion of cash, and all of this ends up in a second bank, and
the

all of which end up in a third bank, and so on. Two

objections can be raised a^inst such a case.

in which all banks get extra cash is much

more common in t e rea
First, the situation
of extra cash, econ even
than the situation in which one bank gets a significant amount
,

creates should end up by being distn ute even >

one bank did get extra cash, the deposits it

throughout all other banks, and thus the cash drain should
be evenly spread among the other
immediately in the case of many banks and many new
banks. Thus, after round one, we are
deposits, rather than in a case in which Bank B is the sole holder of a new epos t.
 ,

684 MONEY, BANKING AND THE PRICE LEVEL

CENTRAL BANKS
Banks of the kind we have been discussing arc private, profit-seeking firms
Most advanced free market economics also have a ceniral bank whose pnmary
purpose IS to regulate the flow of mopey and credit the economy The m
central bank is always an instrument of the central government, whether it
IS mfact owned publicly or not The Bank of England, ‘the Old Lady of

Threadneedle Street’, is the oldest and most famous of the central banks It
began to operate as the central bank of England m
the sixteenth century
Although It was not officially nationalised until 1 947 this represented only a
legal change, its function in the economy had long been to act as the execu-
tor of monetary policy of the British central authorities In the United
States, the central bank is the Federal Reserve System which was organised
in 1913
All central banks perform the same functions, but they have different
forms of organisation In this introductory treatment, we shall concentrate

on the general functions of central banks, leaving the details of the institu-
tion of a particular country to be filled m
by the student’s own reading

1 Banker to Commercial Banks

Commercial banks need a place to deposit their funds, they need a mecha-
nism for transferring funds to other banks and they need a place to borrow
,

money when they are short of liquid funds The central bank accepts de
posits from commercial banks and will, on order, transfer these funds to the
account of another bank In this way, the central bank provides commercial
banks with the equivalent of a chequing account and with a means of
settling debts to other banks
Commercial banks often need liquid funds, and one way of getting these
IS to borrow from the central bank Such borrowing takes one of two forms,

either ‘rediscounting’ a note (or lOU), which represents a loan made to

some individual or firm, or ^
giving the central bank the commercial
bank’s own note, usually backed by the secunty of government bonds The
rate of interest the commercial bank must pay to the central bank is known
in Britain as the bank rate and more generally as the rediscount
RATE ’
The central bank is on occasion a stern banker, and can refuse to

1 The discount rate is merely the rate of iRteren If a bill promising to pay £100 in thre«
months time is sold (at a discount) for £99 then the rate of interest is clearly I per cent for
three months or 4 per cent per year If at the end of two months, the bill is resold (i e
rediscounted) for £99 lOr, then the rale being chaiged is lOr for one month or 6 per cent per
year The original lender of course, receives only IQt for parting with £99 for two months
or 3 per cent per year, although, had he been able to wait a further month he would have
received his full £100
 THE BANKING SYSTEM AND THE SUPPLY OF MONEY 685

lend money commercial bank, or it can demand that loans made in

to a
the past be paid off now. Such actions are particularly important because
the central bank is the ‘lender of last resort’. Central banks to some extent
arose in order to provide a lender of last resort - to prevent a bank that had
sound assets (such as government securities) but not enough ready cash
from being forced into failure by a sudden demand for cash from its deposi-
tors. The histor>^ of banking in the 19th century' is replete with ‘panics’ in

which bank after bank failed because depositors demanded cash that the
banks could not provide. The other side of the coin is that the central bank
can, as we shall see below, use its lending power to regulate the commercial
banks.

2 Banker to the Government

Governments, too, need to hold their funds in an account into which they
can deposit and on which they can write cheques. Most government funds
are on deposit with the central bank. When the government requires more
money, it too needs to borrow, and it does so by printing bonds. Some of
these are sold directly to the public, but the more usual way in which the
government raises funds is to sell bonds to the central bank, which ‘buys’
them in the sense of giving the government a deposit in the amount re-
quired.^

3 Controller and Regulator of the Money Supply

The major function of central banks is to control the money supply. They
do so in a number of different ways: by controlling the issue of currency, by
regulating what banks may do, and by selling or purchasing government
bonds on the open market. We shall discuss these different techniques at some
length later in this chapter.

CURRENCY control: In most countries, the volume of currency is wholly

determined by the central bank, which has no limits on its power to issue
money. In fact, in Britain as in most countries, the quantity of notes and
households,
coins is contracted or expanded to meet the demands of firms or
its quantity is not controlled in an effort to control the total money supply.

RESERVE requirements: In most countries the central bank forces the

to hold cash resert'es weU in excess of what normal
commercial banks
for that amount and turns
I Suppose the government wants £100 million. It prints bonds
by £100 million the deposit of the government.
them over to the central bank, which increases
‘As easy money’, yon may say, and, indeed, it is the
same thing, if \ou remem er
as printing
that deposit accounts are money.
686 MONEY, BANKING AND THE PRICE LEVEL

prudence would dictate In Britain the reserve requirement is that 8 per

cent of total deposits be covered by cash reserves (held either in the bank s
own vaults or on deposit with the Bank of England) In the US the average
reserve requirement is about 15 per cent A rise in this reserve requirement

can force a multiple contraction of deposit money on the commercial banks

To sec this in its simplest possible form, return to Table 49 2 and see what
would happen if the single bank were suddenly required to hold 15 per cent

cash reserves instead of the 10 per cent required in the example

Such changes in the reserve requirements are not commonly used as a

policy instrument But the institution in Britain of Special Deposits was

meant to fulfill the same purpose The Bank of England was empowered to
require that the commercial banks made special deposits with it in addition

to their normal 8 per cent reserve requirements This possibility has been
used, but for various reasons that we cannot go into here, there is debate

about the success of the policy

Open market operations The most important tool that the central
bank has for influencing the supply of money is the purchase or sale of
government bonds on the open market What is the effect of such purchases
and sales ^
If the central bank wishes to increase the cash reserves of the private banks, it buys

bonds in the open market bank buys a bond from a pnvate citizen it
If the
pays for it with a cheque drawn on the central bank and payable to the
seller The seller will deposit this cheque m his own bank The commercial

bank will present the cheque to the central bank for payment, and the
central bank will make a book entry increasing the deposit of the commer
cial bank at the central bank At the end of these transactions, the central

bank will have acquired a new asset m the form of a bond and a new liability
in the form of a deposit by the private bank The individual will have
reduced his bond holdings and will have raised his cash holdings The com
mercial bank will have a new deposit equal to the amount paid for the
bond by the central bank The commercial bank will find its cash assets
and ns deposit liabilities increased by the same amount
The books of the three parties concerned will show the changes indicated
in Table 49 5 after ;{^100 worth of open market purchases have been
completed
The commercial banks are now in the position illustrated in Table 49 1
They have received a new deposit of £100 cash and they can engage a m
multiple expansion of deposits of the kind already studied the earlier m
parts of this chapter Notice that the whole business has been accomplished
by a set of book transactions The commercial banks have extra cash to their
credit on the books of the central bank No new notes or coins have been created
 THE BANKING SYSTEM AND THE SUPPLY OF MONEY 687

Table 49.5

CHANGES RESULTING FROM THE PURCHASE OF A

^100 BOND BY THE CENTRAL BANK FROM A
PRIVATE HOUSEHOLD
Central Bank

Assets Liabilities

Bond +^100 Deposits of Commercial Banks +;01 00

Commercial Banks

Assets Liabilities

Deposits with Central Bank +/) 100 Deposit of Private Household +^100

Private Households

Liabilities
Assets

Bonds -
No Change
f 100
Deposits \vith Commercial Banks
+£100

some cash dram to the pubhc ^see

If the subsequent credit expansion causes
banks wthdraw some o
note 1, page 679) only then will the commercial
then will new notes and com
their deposits for the central bank and only
have to be created by the central bank. , j n
banks, it sell
bank wishes to reduce the cash reserves of the private
If the central
bank sells a bond to a private citizen
bonds in the open market. If the central
dr^v
it gives the bond to the individual
and receives in return his cheque
payable to the central bank.
against his own deposit in his bank and
central bank presents the cheque to the
private bank for Payment.
^
reducing the private bank
payment is made merely by a book entry
held on deposit at the central bank. u
assets by the value of the bond it
Now the central bank has reduced its
688 MONEY, BANKING AND THE PRICE LEVEL

Table 49 6

CHANGES RESULTING FROM THE SALE OF A

£100 BOND BY THE CENTRAL BANK TO A
PRIVATE HOUSEHOLD
Central Bank

Assits Liabilities

Bonds— £100 Deposits of Commercial Banks— £100

Commercial Banks

Assets Liabilities

Deposits with Central Bank— £100 Deposits of Private Households -£100

Private Households

Assets '

Liabilities

Bonds -l-£100
Deposits with commercial bank No Change
-£100

sold and reduced its liabilities m owed to private banks

the form of cash
The individual has increased his holding of bondsand reduced his cash on
deposit with his private bank The commercial bank has reduced its deposit
liability to the individual and reduced its cash assets (on deposit with the
central bank) by the same amount Each of the asset changes is balanced
by a liability change Indeed, everything balances' But the private bank
finds that by suffering an equal change in its cash assets and its deposit
liabilities, Its ratio of cash to deposits falls If this ratio was previously at the

legal minimum, it will now have fallen below the minimum and the private
bank will ha\e to take immediate steps to restore us cash ratio The neces-
sary reduction in deposits can be effected by not making new investments
when old ones are redeemed (e g , by not granting a new loan when old
ones are repaid) or by selling (hquidatmg) existing investments (eg, by
 THE BAXKIXG SYSTEM AXD THE SUPPLY OF MOXEY 689
selling bonds to the public and recei\-ing payment in cheques which reduces
the deposits held by the public).
The booksof the three parties concerned will initially show the changes
indicated in Table 49.6 after ;^100 worth of open market sales have
been
accomplished. This is not the end of the stoiq" for we have already
seen that
when these initial changes have occurred the commercial banks uill have
suffered an equal loss of cash and of deposits. Their ratio of cash to deposits
will have been upset and theywill have to effect a multiple contraction of
deposits in order to restore the required rado of cash reser\'es to total deposits.

Other tools; A Central Bank can affect the supply of money and credit
through a variety of other desdces that operate through interest rates and
through so-called selecdve credit controls. Although these de\-ices are much
less important than open-market operations or reserve requirements, they
are used from time to time.
One such tool is control o\'er the bank rate. Generally, the central bank
can exercise an immediate influence on all rates of interest charged on
up to, say, a year or so) by changing
short-term loans (i.e., loans for periods
the bank rate.^ The reason ^vhy this can be done is partly custom and partly
the fact that, if those making short-term loans should hat'e to borrow from
the central bank, they do not ssish to be in a posidon of charging their
customers too much less than the bank will charge them (for to lend money
for any length of dme at a lotver rate than one is paying for borrowing the
same money is a sure w'ay to financial ruin). Furthermore, the central bank
can force the commercial banking system to borrow' from it by engaging in
sufficiently \igorous open-market operations. These changes in the short-
term rate of interest ^'ill ha\'e certain effects on the economy; in particular,
they will influence the flow’ of foreign funds. A rise in the short-term rate
in London ^vill induce people throughout the W’orld w'ho have money to

lend for short periods to transfer the money to London and lend it out there.
This means that these lenders must turn their currencies into sterling and,
hence, there \vill be an increased supply of foreign funds available in the
London market. A fall in the short-term interest rate has the reverse effect.
1 The whole business of rediscounting is complicated in
Britain by the existence of dis-

count HOUSES, a curiously anachronistic institution whose existence is probably better

borrow
accounted for by the anthropologist than by the economist. The discount houses
money at call (i.e., repayable immediately on demand) from the commercial banlis and lend it

for three-month periods, at a higher rate of interest to the govemmenL Should

the commercial

banlis need cash they can always recall their loans to the discount houses.
The dfrcoimt houses
who have loaned the money out to the government are forced to borrow from the Central Bank
in order to repay their loans to the commercial banks.
movements is analj-sed in the Appendix to ,
2 The significance of these international capital
Chapter 53.
690 MONEY, BANKING AND THE PRICE LEVEL

Whether or not rediscount-rate policy can have an effect on industnal

development, which is usually financed by long-term loans, depends on the
relation between short- and long-term rates of interest Here the evidence is
not conclusive, but, on balance, it tends to suggest that the long-term rate

does not respond very much to changes in the short rate as long as these
changes do not persist for very long periods of time
Open-market operations designed to affect the quantity of money also
have aneffect on interest rates Buying large quantities of bonds tends to
up their price This, as we have already seen, is equivalent to forcing
force
down the rate of interest *
Selling large quantities of bonds, on the other
hand, tends to force down the pnee of bonds and force up the rate ofmterest
Thus, an open-market policy designed to expand the money supply also
tends to lower interest rates, and a policy designed to contract the money
supply tends to raise interest rates
There is a vanety of selective credit controls designed not to regulate the

over-all volume of money or credit, but rather to limit (or encourage)

particular forms of it Stock-market fluctuations can be controlled to some

extent through margin requirements,* consumer credit, which can be
highly volatile, can be controlled in several ways Minimum necessary
initial payments and maximum terms for hire-purchase contracts can be
set
The final tool that we need to mention here is direct order or suggestion
If thecommercial-banking system is prepared to cooperate, the Central
Bank can operate a tight money policy merely by asking banks to be con
servative in granting loans, when the restrictive policy is removed, the
commercial bankers can then be told that it is all right to go ahead granting
loans and extending deposits up to the legal maximum
In all of these ways, the Central flank seeks to control the money supply
in a country

A THEORY OF THE DETERMINATION OF THE MONEY

SUPPLY
We can now develop a theory of what determines the size of the money
supply We do this by making a senes of assumptions about the behaviour
of the various units we have already studied Our theory is based upon four
behavioural assumptions

1 The Central Bank can control the supply of cask reserves in the hands of the
private banks We have seen a number of ways in which the Bank of England
1 See pages 446-9

2 A margin requirement u the proportion of the purchase price of any secunty that must
be put up by the buyer, the remainder being loaned to him
 THE BANKING SYSTEM AND THE SUPPLY OF MONEY 691

can control reserv^es. In the following discussion, we shall assume that they
use open-market operations for this purpose.
2 Commercial banks are proJit-maximisers} Profits are maximised by expand-
ing deposits up to the legal limit as long as there are reasonable-risk invest-
ments available, since newly created deposits can be used to purchase
income-earning assets. For instance, the bank could grant a loan to a

customer (‘purchasing’ the customer’s promise to pay) and in return receive

an interest payment from the customer. This assumption of profit-
maximisation gives private banks an incentive to expand their deposits up
to the legal limit.
3 The public's demand for loans fluctuates widely and is influenced by a number
of factors, one of which is the state of and
activity in the economy. When sales
profits are high and rising, the demand to borrow money will be high and
rising, but when profits and sales are low, the demand to borrow money

will be low. This means that the bank’s ability to expand deposits by making
loans to the public will be limited by the public’s desire to borrow money
and that this desire will fluctuate considerably over time.
4 Banks will refrain from purchasing bonds if they think the price of bonds now is

too high relative to what it will be in the future. If the current price of bonds is

very high (the rate of interest is very low), banks may prefer to hold cash
rather than to buy bonds. If, for example, the price of bonds is expected
to fall by 3 per cent within 6 months and the bond pays its holder an
interest of 2 per cent over a 6-month period, it might be better to hold the
cash now and buy bonds in 6 months’ time, when their price has fallen.
This means that at times when the interest rate is thought to be unusually
low (the price of bonds unusually high), banks may not be willing to ex-
pand deposits by purchasing bonds.

The theory can now be stated as follows: the Central Bank can, through

its control of banks’ reserves, expand or contract the maximum amount

banks may lend. Banks will usually be motivated to expand their loans and
deposits to thismaximum amount. Whether they in fact do so will depend
upon the state of demand. Let us see how this works out.
Assume that the Central Bank wishes to reduce the level of aggregate
demand in the 'economy because it believes there is too much inflationary
pressure at the moment. What is the sequence of events?
1 The Central Bank enters the open market and sells bonds.
The Central Bank receives cheques from the public m
payment for
2
this causes a re uc
the bonds. When it presents these cheques for payment,
tion in the cash reserv'es of the private banks.

really needed is the weaker assumptio

1 This is a rather strong assumption. All that is

ceteris paribus, private banks prefer more profits to less profits.

692 MONEY, BANKING AND THE PRICE LEVEL

3a If cash reserves are already at their legal minimum, then com

mercial banks will have to contract deposits by a multiple of their loss of
reser\es in order to restore their reserve ratio
3b If cash reserves are above their legal minimum, then no contraction
of deposits need occur until cash reserves are reduced below their legal
requirement However, any desired contraction of deposits can always be
achieied by sufficiently large sales of bonds on the open market
4 The reduction in the supply of deposit money will affect total spend
mg m the economy People who would have borrowed money in order to
spend Iton consumer goods or on investment projects will now be unable
to do so Total spending will be less than it otherwise would have been and
this will have a contractionary effect on the economy

Now assume that the Central Bank wishes to increase the level of demand
in the economy, possibly because ihcrc is heavy unemployment The bank
buys bonds, thus increasing the cash reserves of the private banks The
private banks are then able to expand deposits by lending money to mem
bers of the public however, no one wishes to borrow further money, the
If,

expansion of deposits will not occur ‘

Our theory of the control of the money supply has led to two important
predictions

1 A contractionary monetary policy can always force

banks to reduce the supply of deposit money.

2 An expansionary monetary policy permits banks

to increase the supply of deposit money, but does not
force them to do so

THE PROBLEM OF NEAR MONEY

The validity of the theory outlined above has been questioned in recent
years Critics point to many highly liquid assets held by banks, firms, and
households that can easily be converted into money Thus, say the critics,
a contractionary monetary policy will have its effects offset by a liquidation
of these near-monc)S in order to restore desired cash positions If this
happens none of the other predicted effects will Ibffow, because the mone-
1 In these circumstancM Ihe private bank rosM expand deposits by purchasing bonds The
pnet of bonds will, howeser already have been driven very high by the Central Banks
purchasing policy The prisate banks may well decide that they do not wish to buy Urge
quantities of bonds at what may seem to them an abnormally high pnee In this case the extra
cash that accrues to them because of the Central Bank s open market operations merely swells

their cash reserves it does not result man expansion of deposit money This is where
assumption 4 becomes important, because banks can always expand deposits up to the limit
viding they do not mmd whal pnee they have to pay for the bonds they buy
 THE BANKING SYSTEM AND THE SUPPLY OF MONEY 693

tary authorities will not succeed in permanently reducing cash ratios of

banks, firms, and households. This theory is too complicated to spell out in
detail in an elementary textbook and it is still the subject of debate and
testing. At this stage, however, it does serve to warn against accepting as
definitely proved the predictions of the orthodox theory of the control of the
money supply that we have developed above. Choice among competing
theories must rest on evidence, and much work is yet to be done in tliis

field. As yet, it is too early even to guess at the conclusion that \vill emerge;
at any rate, it is not yet possible to dismiss the views of the critics of the

orthodox theory who state that near moneys have greatly altered the
workings of a tight money policy and reduced its effectiveness, at least in
the short run.
CHAPTER 50

THE DEMAND FOR MONEY

Ira household or a firm holds a money balance cither in the form of cash
or of a demand deposit, it incurs an opportunity cost, the cost being
measured in terms of foregone alternatives The minimum cost of holding
cash is measured by the rate of interest on low-nsk secunties, which in most

countnes is something in the order of 5 per cent This means that for every
£1,000 held m
cash by the household or the firm at least £50 is sacnficed
per year in terms of earnings foregone on alternative uses of the money In
the case of a firm subject to credit rationing,* the foregone carnmgs may be
very much higher than the rate of interest on low risk secunties If the firm
ISunable to raise all the money it would like to use, the return on additional
money invested in the firm may be as high as 15 or 20 per cent In this case,
each £1 ,000 held by the firm as cash or demand deposits costs the firm £150
to £200 per year in foregone profits Why then do firms and households
hold any cash at all ’ If we are to answer this question, we must develop a
theory of the demand for money

REASONS FOR HOLDING MONEY

Virtually all transactions m our economy are settled with money Money is
passed from firms to households and back again to firms in order to pay for
the factor services supplied by households to firms and the goods and
'jtTj’Ats, '(Kwi’iK.t'i TfrA VwcortreWa TVivs vs, “zK the

circular flow of income, which we studied in Part 7 These transactions

force both firms and households to hold cash balances
If we let Fnday stand for the day on which the wage bill is paid to house
holds, then the cash balances held by firms will build up through the week
as money is received and will fall sharply on Friday when wages are paid
to households The balances held by households will run down through the
''
I See Chapter 35 page 475
 THE DEMAND FOR MONEY 695

week as purchases are made and will rise sharply on Friday when wages are
received. Firms must make payments other than wages, and the timing of
these is not as predictable as is the timing of periodic
wage payments. The
receipts of cash from the sale of goods are not perfectly predictable, because
one never quite knows either how much rvill be sold or when the goods sold
will actually be paid for. The receipts and disbursements of firms are, there-
fore, subject to considerable random fluctuations. It is necessary to hold
cash balances in order to be able to carry on business when disbursements
are unexpectedly large, or when receipts are unexpectedly small.
It is important to notice that the above discussion implies that cash
balances are held for two different reasons. One reason is the nonsyn-
chronisation of receipts and disbursements ; the holding of balances on this
account is unavoidable. The second reason is the uncertainty about the exact
timing of receipts and payments; the holding of balances on this account is

avoidable (provided the firm is willing to be temporarily embarrassed when

receipts are unexpectedly delayed or payments unexpectedly advanced).
With this general introduction in mind, let us now consider in more detail
each of the reasons for holding cash.

Unavoidable balances: The minimum level of cash balances that must

be held depends on the pay period and the size of the wage bill. Assume, for
purposes of illustration, that firms pay wages weekly and that households
spend all their wages on the purchase of goods and services with the expendi-
ture being spread out evenly over the week. Thus, on Friday morning,
firms must hold balances equal to the weekly wage bill, while on Friday
afternoon households will hold these balances. Over the week, households
balances will run down as a result of purchasing goods and services. By the
same token, the balances held by firms must build up as a result of selling
goods and services until, on the following Friday morning, firms will again
have amassed balances equal to the wage bill that must be met on that day.
On average, over the week, firms will hold balances equal to half the
wage
bill and so will households ; thus total balances held will be equal to the total
weekly wage bill.

The size of these unavoidable balances thus depends on the size of the
wage bill. If the wage bill doubles, either because twice as many people are
employed at the same rate or because the same number is employed at
v/age
the wage rate, then the balances held must double. The size of the
i

tends to vary directly with the level of national income.

holdings, ^sume
In order to see the importance of the pay period on cash
above. ^
that wages are paid daily, instead of weekly, as we assumed
total daily wage i
average, the total balances required will be equal to the
,

which is, of course, only a fraction of the weekly wage bill.

696 MOTLEY, BANKING AND THE PRICE LEVEL
Balances of the kind we have been
discussing must be held, because pay.
ments and receipts are not perfectly synchronised both for firms and house
holds The more often wages arc paid, the more nearly synchronised pay
ments and receipts will be and thus the smaller the balances will be ihat
need to be held
These balances are virtually unavoidable* and would exist m a world in
which everyone was perfectly certain of the timing of all receipts and paj-
ments as long as such inflows and outflows were not perfectly synchronised
We have conducted the argument in terms of the wage bill, but a similar
analysis holds for payments for all other factor services

Avoidable balances The second reason for holding balances anses

because of the uncertainty about the exact timing of receipts and payments
Most goods and services are sold on credit, and the seller can never be quite
certain when these goods will be paid for, while the buyer can never be
quite certain of the day of delivery and thus of the day on which payment
due, nor can he be certain of the degree to which his suppliers will
will fall
be pressing for prompt payment at the time at which such payment is due
In order to be able to continue m
business during times in which receipts
are abnormally low and/or disbursements are abnormally high, firms carry
cash balances that enable them to weather such periods by making pay'
ments out of cash reserves The larger the cash reserves, the greater is the
degree of insurance against being unable to pay bills because of some tern-
porary fluctuation in either receipts or disbursements Cash balances held
for this purpose are avoidable If the firm is pressed for cash or has other
very profitable uses for its funds, it may run down these balances and thus
take a higher risk of being caught by some temporary fluctuation in receipts
and disbursements How serious this nsk is depends on the penalties of being
caught without sufficient reserves by some temporary fluctuation A firm is

unlikely to be pushed into liquidation, but it may have to incur considerable

costs when it is forced to borrow money for short periods in order to meet
such temporary enses The cost depends on the lines of short-term credit
that areopen to the firm The firm might also lose goodwill if it is unable to
pay some of its bills until short-term credit has been arranged
The size of the balances held depends on the degree to which payments
and receipts are subject to random fluctuations and on the volume of pay-
ments and receipts If the volume of transactions rises, then a given cash
holding will provide less and less protection To provide the same degree of

I Firms could loan out the money as it accrued through the week by making very short term
loans repayable on Friday, but, given the present institutional setup and the cost of making
loans such a means of reducing cash balances can be adopted only by a few of the very lai^est
 THE DEMAND FOR MONEY 697

protection as the volume of business rises, more cash is necessary. Thus we

expect the firm’s demand for cash to rise as its own business rises. In the
aggregate, tve expect it to rise as national income rises.

Speculative balances: One other major reason for holding cash is in

order to speculate on the course of future events. The future is never certain,
so that any transaction that takes place over time is necessarily somewhat
speculative. If we think prices are now very low and will soon rise, the
tendency is buy now or to put
to off selling until prices we think
rise. If

prices are high now and will soon fall, the tendency is to sell now and to
postpone buying until prices have fallen. This applies to anything that is

bought and sold, including stocks and bonds. If the price of bonds is very
high (the rate of interest is very low) in relation to what people think is the
normal price, the tendency will be to sell bonds now and postpone intended
purchases until prices have come down. In such a situation, large quantities
of cash will be held in anticipation of a more favourable chance to purchase
stocks and bonds in the future. If, on the other hand, the price of bonds is
very low in relation to what is thought to be the normal price (the rate of
interest is high), the tendency wall be to buy bonds now and to postpone
sales until a more favourable price can be obtained. In this case, the ten-
dency will be
to hold as cash as possible and hold bonds instead.
little

Thus we would expect the holding of cash for speculative purposes to

vary with the rate of interest; the lower the rate of interest (the higher the
price of bonds), the more cash will the public wish to hold.

DETERMINANTS OF THE DEMAND FOR MONEY

We have discussed some of the main reasons for holding money. We now
define the demand for money as the total amount of money that all the

households and firms wish to hold. We may summarise the

in the economy
previous discussion by Listing our hypotheses about the main factors that
determine the demand for money:

1 The demandfar money depends on institutional arrangements. If, for example,

households were paid daily instead of weekly, then unavoidable balances
would be much less than they now are.
income,
2 The demandfor money depends on the level of income. The larger the
the larger the amount of money held in unavoidable balances. The
larger
security
the income, the larger the amount needed to provide a given level of
against unforeseen fluctuations in receipts and payments.
firms that can
3 The demand for money depends on the rate of interest. For
rate of interest (i.e., or
borrow all the money they require at the going
698 MONEY, BANKING AND THE PRICE LEVEL

firms that are not subject to credit rationing), the rate of interest measures
the opportunity cost of hofding cash TTie higher the rate of interest, the
higher the cost to these firms of holding cash and the less cash they will wish

to hold For all firms and households, the rate of interest influences their
decision whether or not to hold cash for speculative purposes The lower
the rate of interest, the less attractive bonds will seem to be and so the higher
the demand will be to hold cash instead of bonds
4 The demand for cash vemes with those factors, other than the rate of mlntsi,

that determine ike opportumty cost of holding cask If a firm is subject to credit
rationing (i cannot borrow all the cash it wishes to borrow at the
e ,
if it

going rate of interest), the opportunity cost of holding cash will not be
measured by the rate of interest It will be measured by the return to the
firm of another pound invested in the firm The opportunity cost will be

higher, the higher the internal rate of return and the more severe the credit
rationing to the firm If, for example, a tight monetary policy makes it very

hard to borrow money for investments, then firms that feel the effects of

this policy will find the opportunity cost of keeping cash holdings increased

They will be tempted to reduce their cash holdings, using the cash thus

Fig 50 1 The relation between the

quantity of money demanded and
the level of national income, the
rate efinterest and the opportunity
cost of holding money

Cash hordmgs Cash holdings

 THE DEMAND FOR MONEY 699

freed to fulfill at least some of their investment plans, which, before the tight
money policy, they were planning to fulfill by spending borrowed money.
A graphical expression of these hypotheses about the demand for cash
is given in Figure 50.1. The demand for cash to hold is shown ceteris paribus
as var)'ing directly with income, inversely with the rate of interest and
inversely with the opportunity cost of holding cash.
CHAPTER 51

THE DETERMINATION
OF THE PRICE LEVEL

In this chapter we discuss not relative prices, but the price level, the average
levelof all prices If all money prices were doubled, relative prices would

be unchanged but the absolute price level would be doubled A nsc m the
price level is often referred to as an tnjlatton and a fall in the price level as a
dejiatm What determines the absolute price level’ What causes it to

change’*
Price levels often change dramatically Table 51 1 shows the average rate

of change in the price level of twenty-one countries over the eight-year

period from 1954 to 1962 It also shows the most and least rapid nse m the
price level from one year to the next Probably the most striking thing about
the table is that it and variability of experience
indicates the great range
among no mechanism m the modern world to
countries Evidently, there is

keep price levels even roughly the same across countries

In Chapter 41, we pointed out that vanations m the circular flow of in-
come could result either because of changes in the quantity of goods bought
and sold, or because of changes m
the prices of these goods Up to this point,
we have assumed that there were unused supplies of all factors of production
and that variations m aggregate expenditure caused changes in quantiues
produced, not m pnees We now wish to take account of the observed fact
that fluctuations m aggregate demand often affect the level of prices, as well
as the level of output and employment

MONEY, AGGREGATE DEMAND AND PRICES

Verv early in the history of economics, changes in the price level were
linked to changes m
the quantity of money Thus the theory of the detcr-
t At thu point you should review your undenunding of the distinction between absolute

and by re reading pages 174-6

relative prices
 THE DETERMINATION OF THE PRICE LEVEL 701

Table 51.1

RATES OF CHANGE IN PRICE LEVELS OF TWENTY-ONE

SELECTED COUNTRIES, IN PERCENTAGES

Average Maximum Minimum

annual rate, rate of rate of
Country 1954-1962 increase increase

26-3 53-2 12-4

Brazil
Canada 1-3 3-1 0
29-3 78-9 0-7
Chile
China 8-1 13-8 0

Denmark 1-3 3-2 -1

Ecuador 0-5 5-1 -1
France 4-2 11-1 0

Germany 0-9 2 -1
3-1 14 -8-8
India
0-5 6-2 -2-9
Iran
Italy 0-3 3 -3
0 5 — 6*5
Japan
Mexico 4-4 14-2 1

0-4 3 2
Netherlands
3-2 0
New Zealand 1-3

Norway 1-5 4-2 -2

Sudan -M 9 — 14’5
0-2 7-2 -7-3
Syria
0-5
UK 2-1 3-9
0
0-9 3-2
US
1-5 4-2 1
Yugoslavia

o t e
mination of the price level was related to the theories
the supply of money. There is, however, an intermediate
n m
e rplatinn

for, or supp y
between money and prices. A change in demand ’
,

e
is thought to affect aggregate demand, and
a change in aggrega
contemporary econ
is thought to affect the price level. Past and .

about the relation between money

tains many controversies
we rst cons
These controversies are more easily studied if
go on to con
between aggregate demand and prices, and then
between money and aggregate demand.
702 MONEY, BANKING AND THE PRICE LEVEL

The Relation Between Aggregate Demand and the Price Level

The L-shaped Function
A simple theory that makes a stnct dichotomy between changes in output
and employment on the one hand, and changes in the level ofpnces on the
other,IS commonly used in elementary macro-economics below ‘full cm
ployment’, prices are assumed to be fixed, and fluctuations in aggregate
demand are assumed to cause only fluctuations in output and employ
ment at ‘full employment’, output and employment are assumed to be
fixed, and fluctuations in aggregate demand are assumed to cause only
fluctuations in the price level Let us consider each of these situations
In periods of less than full employment, excess supply exists in most
markets, and the simple competitive theory of price outlined in Part II pre

diets that each individual price, and hence the average level of all prices,
will be falling We now assume, however, that pnccs arc sticky in a down
ward direction When there arc unemployed resources, pnccs either do not
fall at all, or else they fall so slowly that, for all practical purposes, they can
be regarded as being constant over a period of several years The constancy,
or at least the extreme stickiness of pnees m
a downward direction, is
assumed to be caused by trade-union resistance to a reduction in wages, by
government price support policies, particularly those the agncultural m
sector, and by manufacturers’ resistance to a fall in prices As a result of the

downward inflexibility of prices, variations in aggregate demand below the

full employment level will cause variations in the level of output only Thus
the circular-flow theory of the previous chapters provides a theory of the
level of output and employment m the economy
Once the economy reaches full employment, further increases in output
become impossible in the short run A rise m aggregate demand cannot now
be met by a rise in output, so pnees nse instead The theory assumes no
stickiness of prices in an upward direction
This theory IS illustrated in Figure 51 I In this figure, the level of resource
utilisation on the
IS plotted X
axis and the rate of change of the pnee level
ISplotted on the Y axis Each point thus indicates the percentage utilisation
of resources and the accompanying amount of inflation or deflation ob-
served m the economy According to the theory, all observations should be
clustered in the shaded band drawn around the axes When there are
unemployed resources, the price level should be observed to be constant
A change in aggregate demand will change the percentage uuhsation of
resources, moving the economy, say, between points a and b without causing
any change m the price level Prices only nse when full employment exists
so that all observations of inflationary price changes should be clustered m
the narrow vertical band at or very near 100 per cent utilisation of resources
Fig 51.1 The between changes in the price level and the volume of
relation
employment when there is no policy conflict between full employment and
stable prices.

between changes in the price level and the volume of

Fig 51.2 The relation
employment an
a policy conflict between
full
employment when there is

stable prices.

such as that sho^vn by

What not allowed by the theory^ is a situation
is
is combine wi a su
point X in which a significant degree of inflation
stantial amount of unutilised resources. ,

pohey imp ic -
The ‘L-shaped’ relation of Figure 51.1 has two important
reversed. Prices nse v e
tions. The first is that inflation can never be
there is excess supp y
^

is excess demand, but they do not fall

when
^
in this circumstance that the price level
displays a ratchet effec : i

imphcaPon
only in one direction. The second and more important
and deflationary gaps on
the inflationary
1 This is the theory^ we used when we introduced
pages 647-51.
704 MONEY, DANKING AND THE PRICE LEVEL

L shaped rchtion is that there is no conflict between the two policies of full
emplo>ment on the one hand nnd maintaining a stable price le\el on the
other If there unemplo>mcnt m the economy, aggregate demand can be
is

raised until full employment occurs, without any consequent inflationary

pressures iwU result only if demand u increased in a iituationin
Inflation
M hich full employ ment already exists The economy can be kept m a posi
tion such as / in Figure 51 1 with full employment and stable prices If the
economy is at with unemployment, then aggregate demand needs only
ti

to be expanded to mo\e it to/, if the economy is at cisith an inflation under

\\ay then aggregate demand need merely be reduced until the inflation

stops and the economy returns to

I HE E \ iDFNCE Tlie theory of the L shaped curse ha* been subjected to
I great deal of testing in the last decade The testing of any interesting
ihrors almost always raises complicated technical problems It is fairly

« howeser that the simple dichotomy changes m output only at less

lear
than full employment and changes m pnccs only at full employ ment has ,

not stood up well to testing As a matter of obscnation, it appears that

changes in aggregate demand usually cause clianges both in output and in

jinces The closer the economy is to full employment, tlic more likely it is

that any change m demand will cause a pnee change and the less likely it is

that it will cause an employment change, but, no matter how high the level

ot resource use is it is alwayspossible to expand the rate ofproduction a little

hit so that a nsc m demand is always accompanictl by some rise m output

The Relation Between Aggregate Demand and the Price Level the
Phillips Curve

The actual observations for to he moTc in the curved

any economy seem
band 2 than in the L shaped band of Figure 51
illustrated in Figure 51 1

The relation between the level of employment of resources and changes m

the price level illustrated in Figure 51 2 shows pnccs falling slowly for large
amounts of unemployment in the economy for some smaller volume of ,

unused resources the price level remains steady, while the price level nses
when resource utilisation is at a hi^h level The higher the lev el of resource
utilisation the more rapid the nse in prices but it is always possible to
obtain a further increase in resource utilisation, and hence output at the m
cost of a more rapid rise m
pnees If, Ibr example, excess demand were
causing an inflation of de per cent per year, and the percentage of resources
used was Od per cent, then a further nse in excess demand would raise the
rate of inflation to
resource use to
say, ^ per cent but simultaneously raise the level of

Of per cent
According to the relation shown in Figure 51 2 the economy docs not
 THE DETERMINATION OF THE PRICE LEVEL 705

suddenly move from a situation of underemployment of resources and con-

stant prices to a situation of full (constant) employment with varying prices;
instead, the economy moves by degrees from one to the other. Any increase
in expenditure will lower unemployment to some extent, and so raise out-
put, but the lower the level of unemployment already achieved, the less
will any rise in aggregate demand affect unemployment and the more will
it affect prices.'

We conclude that the evidence supports the theory that variations in the
level of aggregate expenditure cause variations in both prices
and output;
the lower the level ofunemployment, the larger the change in prices and
the smaller the change in unemployment for a given change in the flow of
aggregate expenditure; the higher the level of unemployment, the more will
changes in expenditure cause output variations and the less will they cause
the price level to var^'.
Situations such as the one shown by points e and g in Figure 51.2 repre-
sent disequilibrium positions: if the inflation does eventually eliminate the
excess demand, the level of resource utilisation will fall back to that com-
patible with a stable price level. This does not affect the fact that, if the
central authorities arc
prepared to take steps to allow the inflation to con-
tinue, theycan achieve a level of unemployment lower than that which
would obtain if there were no inflation. Also, the more rapid the rate of
price inflation the authorities arc willing to take steps to maintain, the
lower the rate of unemployment they need to accept. Of course, to maintain
a level of unemployment below h per cent, the authorities must take steps
to ensure a continuing disequilibrium with continuing inflationary pressures
in the economy.
Note that the policy implications of
this relation differ from those of the
L-shaped relation of Figure 51.1: the control of inflation and the main-
tenance of the highest possible level of employment are now conflicting
policy objectives. If the economy is at point e, for example, with 100 per —
cent unemployment and with de per cent inflation per year, then the infla-
tion can be slowed down at the cost of increased unemployment; on the
other hand, the level of unemployment can be reduced, but only at the cost
of more inflation. We
shall have more to say about the problems raised by
this relation when we discuss macro-economic policy in Chapter 58.

the relation between money and aggregate

demand
So far we have discussed the relation between aggregate demand and the
price level. In this section we discuss the relation between money and
aggregate demand.
1 T^hc theory behind this is spelled out in a bit more detail in Appendix A to this Chapter.

23
706 MONEY, BANKING AND THE PRICE LEVEL

In Chapters 49 and 50, we developed theories of the supply of money and

the demand for money demand for money is equal to its supply,
If the
everyone has the quantity of money he wishes to hold, and a situation of
equilibrium exists Problems arise when we are not in equilibrium if the
demand for money is not equal to its supply, households and firms either
have more money than they wish to hold {supply exceeds demand) or less
money than they wish to hold (demand exceeds supply) in both cases they ,

can be expected to do something about it In the former case we would

expect them money for some purpose, in the latter case
to use the surplus
we would expect them to try to ma^e up their dclTcicncy by attempting to
obtain more money from one source or another

Quantity Theory and Keynesian Theory of Money a Preview

One important feature that distinguishes different theories of money is the

hypothesis made about what happens in a situation m which the demand
for money is not equal to the supply of money VVe would expect household!
and firms to do something, but different theones make different behavioural
assumptions about what (hey will do
We shall study two extreme hypotheses The first is that all the cffecu of
a difference between the demand for and the supply of money are mam
Tested solely m changes in aggregate demand for goods and services This
behavioural hypothesis is associated with the quantity theory of money It is

based on an assumption that, when firms and households have more

money than they wish to hold, they will spend the excess on currently
produced goods and services, while, when they have less money than they
wish to hold, they will try to build up their balances by reducing their
expenditure on goods and services to an amount less than their current
incomes, adding the difference to their cash holdings According to this
theory, disequilibrium between demand for and supply of money causes
changes m aggregate demand for currently produced goods and services
The other extreme hypothesis, which is associated with the Keynesian
theoryof money, is that a difference between the demand for and the supply
of money is manifested solely in the demand for and the supply of securities
According to this theory, when firms and households do not have sufficient
cash balances they attempt to build them up by selling bonds, and when
they have more cash than they need to hold, they attempt to invest the
surplus in bonds Thus, a difference between demand for money and its
supply will not have any direct effect on aggregate demand for goods and
services There may, however, be an indirect effect The attempt on the
part of households and firms either to buy or to sell bonds in large quantities
I That u stocks and bonds In this text we shall talk about bonds
 THE DETERMINATION OF THE PRICE LEVEL 707

of these bonds. A change in the price of bonds is, as we

will affect the price
have already seen, the same thing as a change in the rate of interest. Insofar
as investment decisions are affected by the rate of interest, there may be an
effect on aggregate demand. The links between money and aggregate
demand are less direct in this theory, and the quantitative effect on aggre-
gate demand of a difference between demand for and supply of money is
likely to be small, because depends on the fairly weak link between
it

changes in the rate of interest and changes in investment expenditure.

The significance of these two hypotheses is great. If one believes the
quantity hypothesis, one believes that changing the money supply will have
a strong and direct effect on aggregate expenditure. This makes monetary
policy a tool of control of great potential power. If one believes the Keynesian
hypothesis, one believes that changing the money supply will have only a
weak and indirect effect on aggregate expenditure and only on that part of
it that is composed of investment expenditure. This makes monetary policy

a much less potent tool for controlling the economy than its appears to be
to supporters of the quantity theory. This is probably the most important
difference between the quantity and the Keynesian theories, and it is what
the student really needs to understand. It is now necessary, however, to
elaborate on these two theories in order to develop the differences and in
order to consider the relevance of each theory to actual observadons.

The Quantity Theory

The two basic assumptions of the theory are ( 1 ) that the demand for money
is proportional to the level of transactions in the economy, and (2) that an
excess demand for, or supply of, money is reflected in aggregate demand for
currently produced goods and services.

The demand for and the supply of money: It is assumed that house-
fraction
holds and firms wash to hold cash balances equal to some constant
of the annual value of their transactions. We
indicate this proportion by
the symbol k. Thus the demand for money is merely k times the
money
value of transactions, and the demand curve for money is the one shown
in
then
Figure 50.1 (i). Assume that k is j^Tf transactions are ;^1,000 million,
the money value o
;(^100 million will be required as cash balances; if

transactions rises to £2,000 million, then £200 million wiU be deman e

for cash balances. The value of A depends, in this theoiy,

on institutional
Chapter 50, or
factors such as the pay period. In the examples given in
of wages re
example, transactions regarding the earning and spending
quired that households and firms hold balances equal to 52
of the 1 w^e
when the pay period and
was once a week, when it was once a ay.
3^
708 MONEY, BANKING AND THE PRICE LEVEL

these were the only transactions in the economy, k would be just less than
02 (2 per cent) when wages were paid weekly As it is, there are other
transactions, each type requiring the holding of some charactcrisuc proper
tion of cash balances For the whole economy, k is the average of all these,
and It expresses the over all need to hold balances as a fraction of the total
value of all transactions m the economy *

Next let us take the money value of transactions and split it into two
parts, a real part T’, which isthe number of transactions that occur over
some stated time period, say, one year, and a value part, P, which is the
average price at which these transactions take place We can now express

the demand for money in the following equation

A/j = kPT (1)

In this equation, Mi is the demand for money, PT \% the annual money

value of transactions, and k is the fraction of this value required to be held

m transaction balances It should be noted that we have now related the

amount of money held in balances. Mi, which is a stock, to the money value
of transactionswhich is a flow ^ The job is done by the humble little k,
which seems so insignificant but which actually is so powerful The hypo*
thesis that thecash balances one needs to hold are some fraction of the
annual value of transactions creates a link between the stock of money and
the flow of income
Next we assume that the supply of money is an exogenous constant deter
mined b> the central authorities To show this, we write

A/. = Af, (2)

where Af, is the supply of money and M

is some constant amount (measured
in pounds sterling) Equation moncydoes
(2) merely says that the supply of
3t depend on any other factors
m
the economy, it is simply what the
itral authorities want it to be
In equilibrium, it is necessary that the supply of money should equal the
demand for u Thus we can wntc as an equilibrium condition

Mi = A/, (3)

When (3) holds, households and firms will have just the amount of money
balances they require When (3) does not hold, they will have too much or
too little money, and we expect them to try to something about it do
In Figure 40 3 page 548 the total value than
£590 This n much more
1
of transactions is
the total value of national income because every time something changes hands its total value
part of total transactions but only thewAcr
eiMttfby the selling firm is part of national income
IS

2 One of the tnckieit problems in monetary theory stocks and flo'*'*

is to distinguish between
(see page 36) and to discover relations between stocks of money and assets on the one hand
and flows of expenditureon the other
 THE DETERMINATION OF THE PRICE LEVEL 709

Excess demand for money and aggregate demand for com-

modities: The second basic assumption of the quantity theory is the one
we have already mentioned: differences between demand for and supply
of money bring about changes in aggregate demand; if the supply of
money exceeds the demand, firms and households will try to spend their
excess money on goods and services, and this raises aggregate demand; if
the demand for money exceeds the existing supply, households and firms
\vill cut do%vn on their purchases of goods and services in order to build up

their balances of money. This will reduce aggregate demand.*

velocity and balances: The equilibrium condition in (3) above says

demand for money must equal the supply of money. If we substitute into
equation (3) our theories about what determines the demand for money,
equation (1), and the supply of money, equation (2), we obtain

kPT = M, (4)

which says that, in order that the demand for money should equal its supply
in equilibrium, the fixed stock of money {M) should be a fraction, k, of the
annual money value of transactions, PT.
In many treatments, equation (4) is not used. Instead of dealing with
(4), k is inverted and called the velocity of circulation. Let us see how this
is done. Take (4) and divide both sides by k to give equation (5)

PT-=\-M. (5)
k

Next, as a simple matter of terminology, let us use the symbol V to stand

for \jk. This gives us equation (6):

PT= VM.

It makes no difference whether we work with (4) or with (6).

have k on the left-hand side of the equation; in (6), we have \fk
the right-hand side. Although it is as simple as that, we have >

careful about the way in which we interpret V and k.

In order to see what is involved, consider a simple example. Assu

1 In the terminology of Part 7, an excess demand for money creates a withdra*

the income stream. Firms and households .save some of the money that they receive in
increase their cash holdings. An money creates an
excess supply of injection. When fn
households spend their extra money, they make an addition to the circular flow that i

result from their own receipts of income.

708 MONEY, BANKINO AND THE PRICE I EVEL

these were the only transactions m

the economy, k would be just less than
02 (2 per cent) when wages were paid weekly As it is, there are other
transactions, each pe requiring the holding of some characteristic proper*
t>

tion of cash balances For the whole economy, k is the average of all these,
and It expresses the over-all need to hold balances as a fraction of the total
‘
value of all transactions in the economy
Next let us take the money value of transactions and split it into two
parts, a real part, T, which is the number of transactions that occur over
some stated time period, say, one year, and a value part, P, which is the
average prtee at which these transactions take place We can now express
the demand for money in the following equation

Urf = kPT (1)

In this equation At^ is the demand for money, PT is the annua! money
value of transactions, and k is the fraction of this value required to be held
in transaction balances It should be noted that wc have now related the
amount of money held m balances, Mg, which is a slock, to the money value

of transactions which is a flow* Tlie job is done by the humble Imlc k,

which seems so insignificant but which actually is so powerful The hypo-
thesis that the cash balances one needs to hold are some fraction of the
annual value of transactions creates a link betw cen the stock of money and
the flow of income
Next we assume that the supply of money is an exogenous constant deter-
mined by the central authorities To show this, wc write

M, = 7
,\ . ( 2)

where M, is the supply of money and M

some constant amount (measured
is

in pounds sterling) Equation (2) merely says that the supply of money docs
not depend on any other factors in the economy, it is simply what the
central authorities want it to be
In equilibrium, it is necessary that the supply of money shoul,
demand for it Thus wc can wntc as an equilibrium condition
Afj =
) holds, households and firms will have unt ol
they require When (3) does not hold, too mu
it money, and wc expect them to try to dc it

Figure 40 3 page 546 (he total value of traiuactiom h more

il alue of national income because every time something ^
\ total v
of total transactions but only the ra/vrs^ir^by ihe selling fv I ince
One of the trickiest problems in moneury theory u to disting and
page 36) and to discoser relations between stocks of mone>
1 flows of expenditure on the other
 : :

THE DETERMINATION OF THE PRICE LEVEL 709

Excess demand for money and aggregate demand for com-

modities. The second basic assumption of the quantity theory is the
one
we have already mentioned differences between demand for and supply
:

of money bring about changes in aggregate demand; if the supply of

money exceeds the demand, firms and households will try to spend their
excess money on goods and services, and this raises aggregate demand;
if
the demand for money exceeds the existing supply, households and firms
will cut down on their purchases of goods and services in order to
build up
their balances of money. This will reduce aggregate demand.^

VELOCITY AND BALANCES: The equilibrium condition in (3) above says

demand for money must equal the supply of money. If we substitute into
equation (3) our theories about what determines the demand for money,
equation (1), and the supply of money, equation (2), we obtain

kPT ~ M, (4)

which says that, in order that the demand for money should equal its supply
in equilibrium, the fixed stock of money {M) should be a fraction, k, of the
annual money value of transactions, PT.
In many treatments, equation (4) is not used. Instead of dealing with
(4), k is inverted and called the velocity of circulation. Let us see how this
is done. Take (4) and divide both sides by k to give equation (5)

PT=\-M. (5)
k

Next, as a simple matter of terminology, let us use the symbol V to stand

for 1 jk. This gives us equation (6)

PT= VM. (6)

makes no difference whether we work with (4) or with (6). In (4), we

It
have k on the left-hand side of the equation; in (6), we have 1/A: (=V) on
the right-hand side. Although it is as simple as that, we have to be very
careful about the way in which we interpret V and k.
In order to see what is involved, consider a simple example. Assume that
ivithdrawa! from
I In the terminology of Part 7, an excess demand for money creates a
money that they receive in order to
the income stream. Firms and households save some of the

increase their cash holdings. An excess supply of money creates an injection. When firms and
that does not
households spend their extra money, they make an addition to the circular flow
result from their own receipts of income.
708 MONEY, BANKING AND THE PRICE LEVEL

these were the only transactions m the economy, k would be just less than

02 (2 per cent) when wages were paid weekly As it is, there are other

transactions, each type requiring the holding of some characteristic propor-

tion of cash balances For the whole economy, k is the average of all these,
and It expresses the over all need to hold balances as a fraction of the total
‘
value of all transactions in the economy
Next let us take the money value of transactions and split it into two

parts a real part T, which is the number of transactions that occur over
some slated time period, say, one year, and a value part, P, which is the
average price at which these transactions take place We can now express
the demand for money in the following equation
= kPT (1)

In this equation, Mi is the demand for money, PT is the annual money

value of transactions, and k ts the fraction of this value required to be held
m transaction balances It should be noted that we have now related the

amount of money held in balances, M 4 which

money value
1 $ a stock, to the
of transactions which is a flow The job is done by the humble little k,
*

which seems so insignificant but which actually is so powerful The hypo-

thesis that the cash balances one needs to hold are some fraction of the
annual value of transactions creates a link between the stock of money and
the flow of income
Next we assume that the supply of money is an exogenous constant deter-
mined by the central authorities To show this, we write
M, = M, ( 2)

where M, is the supply of money and M

1 $ some constant amount (measured

in pounds sterling) Equation (2) merely says that the supply of money does

not depend on any other factors m

the economy, it is simply what the
central authorities want it to be
In equilibrium, it is necessary that the supply of money should equal the
demand for it Thus we can wnte as an equihbnum condition
Mi = M, (3)

When (3) holds, households and firms will have just the amount of money
balances they require When (3) docs not hold, they vrtll have too much or
too little money, and we expect them to try to do something about it
1 rnFigure403 page 548 the local valoeof transactions is ,£590 This is much more than
the total value of national income because every time something changes hands its total value
IS part of total transactions but only the po/aea^i&dby the selling firm is part of national income

2 One of the trickiest problems in monetary theory is to distinguish between stocks and flows
(seepage 36) and to discover relauons between stocks of money and assets on the one hand
and flows of expenditure on the other
 THE DETERMINATION OF THE PRICE LEVEL 709

Excess demand for money and aggregate demand for com-

modities: The second basic assumption of the quantity' theory- is the one
we have already mentioned: differences between demand for and supply
of money bring about changes in aggregate demand; if the supply of
money exceeds the demand, firms and households wiU tiy to spend their
excess money on goods and services, and this raises aggregate demand; if
the demand for money exceeds the existing supply, households and firms
will cut dowTi on their purchases of goods and services in order to build up
their balances of money. This w'ili reduce aggregate demand.^

VELOCITY AND BALANCES: The equilibrium condition in (3) above says

demand for money must equal the supply of money. If we substitute into
equation (3) our theories about what determines the demand for money,
equation (1), and the supply of money, equation (2), we obtain

kPT = M. (4)

which says that, in order that tlie demand for money should equal its supply
in equilibrium, the fixed stock of money [M] should be a fraction, k, of the
annual money value of transactions, PT.
In many treatments, equarion (4) is not used. Instead of dealing with
(4), k is and called the velocity of circulation. Let us
inverted see how this

is done. Take (4) and divide both sides by k to giv'e equation (5)

PT=\-M. (5)
k

Next, as a simple matter of terminologvL let us use the svmbol V to stand

for Ijk. This gives us equation (6):

PT= VM. (6)

makes no difference whether we work with

It (4) or with (6). In (4), we
hav'e k on the left-hand side of the equation; in (6), vve have Ijk (=Vj on
the right-hand side. .'Although it is as simple as that, we have to be veiy
careful about the way in which we interpret I' and k.

In order to see what is involved, consider a simple example. Assume that

I In the terminologv- of Pan an excess demand for money creates a withdrawal from
7,

the income stream. Firms and households save some of the money tliai they
receh e in order to

holdings. excess supply of money creates an injection. When firms and

increase their cash .An
make an addition to the circular flotv that does not
households spend their extra money, thes'
result from their own receipts of income-
 710 MONEY, BANKING AND THE PRICE LEVEL

6 million transactions occur each year at an average pnce of £2 per trans

action making an annual value of transactions of £12 million Assume that
1$
(s 0833) so that desired money holdings are of the annual value
of transactions and that the total quantity of money in existence is £1
million This economy is in monetary equilibrium, since the demand for
money is equal to its supply, as is shown by substituting the following values
into equation (4)

Af = £1 milhon
P=£2
T= 6 million
k = 0833

we now replace k by V, we obtain y = 12, substituting mto equation

If
(6),we also find, of course, that the economy is in equilibrium We may
now interpret V as measunng the average number of times that a unit of
money changes hands m order to effect the total value of transactions In
our present economy, the typical unit of money must have changed hands
12 times m order that £12 million worth of trade could be done with a
money stock of £l million Thus we give the name velocity op circc-
^

LATioN to V It tells us the average number of times each unit of money

must have changed hands in order that a given annual volume of business
could be accomplished

Disequilibrium behaviour Now consider what happens to any

economy not in equilibrium Assume that, in the above example, the money
supply IS increased to £2 milhon Firms and households now have more
money than they wish to hold and, according to the assumptions of the
theory, they try to spend the extra on goods and services If we make the
simplifying assumption that ail the resources of the economy are fully
employed no more can be produced, then all that can happen is that
so that
pnees will rise^ Transactions wiH remain constant at 6 million and prices
will rise As long as the price level remains below
£4, firms and households
will be trying to spend some of their balances, which means that they will
be spending in excess of their incomes This continues the upward pressure
on pnees Once the pnce level has doubled (so that in our numerical ex
ample the average pnce of all transactions is £4), all the money will be
required for transactions balances, the demand for money will equal its

'/l should help considerably to re read the first few pages of Chapter 40 at this point
It

2 t\e assume that is that we are at the full-employment point on the L-shaped curve of
Figure 51 1
 :

THE DETERMINATION OF THE PRICE LEVEL 711

supply, and the economy will again be in equilibrium. This can be seen by
substituting the new set of values into equations (4) and (6)

M = £2 million
P = £^
T = 6 million
k = -0833 = -^) {

V= 12.

The predictions of the q,uantity theory: So far we have made three

basic assumptions: first, that the supply of money is exogenously deter-
mined second,
;
that the desired ratio of cash holdings to transactions, k, is a
constant; and, third, that disequilibrium situations with respect to demand
for and supply of money cause disequilibria in the market for goods and
services. If we make one further assumption, we can develop a simple
quantity theory. This assumption is that the economy is always fully
employed so that T remains constant at the level of transactions appro-
priate to full employment. ^ Thus changes in aggregate demand affect only
prices.
These assumptions now give rise to the most famous prediction of the
quantity theory: changes in the quantity of money will induce exactly
proportional changes in prices. To derive this prediction, take equation (4),
which expresses the condition for monetary equilibrium.

kPT = M, (4)

and divide both sides hy kT to obtain

M
‘-w (7)

Since k and T are constant, we see that, in equilibrium, M and P must vary
in proportion to each other.^
Thus any given percentage change in M
will, according to this theory,

cause an equal percentage change in P. Furthermore, as long as Af is

determined exogenously by, say, the supply of gold (under the gold
standard), the direction of causation is clear: variations in cause the M
variations in P, not vice versa.
The predictions of the quantity theory have been applied to two some-
what different sets of circumstances: long-term changes in the price level.

For one justification of this assumption, see Chapter 56, pages 799 800.

2 One can write P = aM, where a — •

 712 MONEY, BANKING AND THE PRICE LEVEL

where the problem is to predict changes from one decade to the next over
a span of anything from fifty to several hundred years, and short-term
changes in the price level, where the problem is to predict changes from
year to year (or even from month to month) over a span of, say, five or ten
years Each of these applicaUons raises its own characteristic problems,

which we now consider in turn

The (QUANTITY THEORY APPLIED TO LONG-TERM CHANCES IN THE PRICE

LEVEL The long-term application of the quantity theory is based on a
number of behavioural assumptions (1) Tis determined by the economic
growth of the economy, changes in T due to growth are large relative to
those caused by temporary lapses from full employment, thus T can be
taken to be exogenously growing at the long-term growth rate, which will
be around 2 or 3 per cent per year for the typical Western society, (2)
although V can vary from year to year, it is a trend free variable (i e we ,

hypothesise that its average expected value in the 1950s is the same as it
was in the 1850s),* (3) variations in the money supply were (until very
recent times) exogenously determined because the money supply was linked
to the gold suppl) by the requirement of convertibility Thus, T can be
determined, V can be predicted (with an ascertainable margin of error) to
be constant, and P will be observed to vary directly with Therefore, our M
theory states that

MV = PT,
where Fis a constant, A/ is the exogenously determined stock of money, and
variations in T arc determined by the economic growth of the economy
There is considerable favourable evidence linking long-term trend
changes in the pnee level to long-term changes m the quantity of money
Four of the most dramatic cases may be mentioned During the period
when the Spanish were importing gold and silver from the New World, a
major increase in the quantity of money occurred followed by a major nse
in prices The influx of gold caused a great nse spending without any m
corresponding increase m
output, therefore, prices rose The first tvvo-thirds
of the jiineteenth century was a period of rapid economic growth in Europe
and America, but it was a period in which the money supply did not expand'
was nsing only slightly Thus T was
rapidl) because the world’s gold supply
growing faster than Af, and the theory predicts that a downward pressure
should have been exerted on pnees Such a downward pressure is clearly
observable in the price levels of most of the world’s major countries Later

t This formulaiion is more restrictive than is necessary ^^e do expect V to change in the
long run because of various msticutiona] shifts such as changes in the pay period According to
the theory however, thesechangeswitlbeoccasional discrete exogenous and easily identifiable
 THE DETERMINATION OF THE PRICE LEVEL 713

in the nineteenth century', gold strikes in America, Canada, and South

Africa brought a more rapid increase in the money supply, and the price
level responded, as the theor>' predicted it would. During the First and
Second World Wars, the governments of several countries found themselves
unwilling to raise by taxes the money they ^vere committed to spending.
Large budget deficits occurred, which they financed by borrowing from the
central bank, which in turn created the requisite amounts of new money.
Very rapid money supply ensued, and these were imme-
increases in the
diately followedby the effects predicted by the quantity theory. Many such
periods have been documented, and they provide impressive evidence of the
close link between major changes in the quantity of money and major
changes in the price level.

The quantity theory applied to short-ter.m changes in the price

level: The major problem in using the quantity theory to predict the
short-term beha\'iour of the price level is that the actual obser\'ed velocity
of circulation of money does vary considerably over short periods of time.
We must be very careful to distinguish here between the actual and the
desired velocity of circulation. The V that we have been speaking of so far
has been the desired V based on the desired ratio of cash holdings to trans-
actions (desired k). For any economy whatsoever, there mil be a velocity that relates

the observed money stock to the observed value of transactions. If we say that F* is

the observed velocity of circulation and that k* (

= 1/F*) is the observed
ratio of money in the economy to the value of transaction, we can write

k*PT = M, (4*)

and PT = V^. (6*)

The identity signused here because these relations are necessarily true
is

they always hold by virtue of the way in which k* and F* have been defined.
For example, if we observe an economy in which F7’=;^l,000 million and
A?=;^100 million, then we say that observed F* = 10 or observ^ed k* = -\.
If, economy,
for this M
increases to million and nothing else happens,
we merely say that F* has fallen to 5 or k* has risen to '2, which is the same
thing. Expressions (4*) and (6*) do not constitute a theory of money and
prices; they are true by definition and do not allow us to predict anything
about the economy.
To have a theory of prices we need a theory about desired F (or k). This
is what the simple quantity theory provides by hypothesising that desired k
is a constant, and this gives rise to the prediction that in equilibrium a
change in the quantity of money wdll cause a proportional change in the

price level. The observed variability of F* from year to year refutes this

23
714 MONEY, BANKING AND THE PRICE LEVEL

Simple quantity theory We

cannot predict the relation between changes
in M and changes in P because we cannot assume
that V* will remain con-

stant from one year to the next, even though its average value may be
constant from one decade to the next
There is, however, another possibility We may accept the short-run
variability of V* and develop a theory that predicts how it will vary by
linking desired k {and hence desired V) to other observable quantities in
the economy As long as we have a theory that predicts how actual V* will
change, VtC can predict changes in P on the basis of changes in M, since we
know what to expect V* to do Seen in this light, the simple theory that V
IS constant is just a special case of the more general class of quantity theories,
all of which attempt to link changes in M to changes in P and all of which
require, in order to do this, a theory of the behaviour of V
Some economists have tried recently to restore the quantity theory on
these lines by developing a theory of how V changes They have hypo
thcsized that V changes m a predictable fashion because it is a function of
certain observable economic variables ' According to these economists, the
quantity theory will be useful in predicting the short-run behaviour of prices
because short run changes in V will themselves prove to be predictable
Evidence is still being gathered on this sophisticated attempt to restore the
quantity theory as Ike general theory of the price level, but it is still too early
to say how successful the attempt will prove to be
We may summarise our discussion of the quantity theory as follows
(1) The theory that states that long-run movements in the price level can be
explained by changes in the quantity of money has stood up to testing
(2) The theory that states that short-run movements m the price level can
be associated with changes in the quantity of money because V is reasonably
stable has not stood up to testing (3) The theory that states that short-run
changes in the price level can be associated with changes in and M pre-
dictable changes in V is at present the subject of debate and testing

The Keynesian Theory of Money

The Keynesian diverges in two cntical ways from the
theory of money
quantity theory The Keynesian theory accepts that the demand for money
depends partly on the value of transactions, but it adds that it also depends
on the rate of interest (see Figure 50 !(**))
We have already mentioned the second basic assumption of the Keynesian
theory concerninghow firms and households behave when the demand for
money does not equal supply In the quanuty theory, firms and households
1 Professor Milton Friedman of the Univenity of CSiicago is the champion of this attempt
 THE DETERMINATION OF THE PRICE LEVEL 715

alter their expenditure on goods and services; in the Keynesian theory, they
alter their behaviour in the market for monetary assets. If the demand for
money exceeds its supply, people try to sell bonds, thus adding to their stocks
of cash. If the supply of money exceeds its demand, people try to buy bonds
with their excess stocks of cash. It is, of course, quite possible for one indi-
vidual to adjust his money holdings by dealings in the bond market. But all
individuals cannot do this simultaneously unless the total supphes of money
and bonds vary. If they do not vary, we get a situation in the bond market
analogous to the one that under the quantity theory when conditions
arises
of full employment in the goods market prevail. If the quantity of bonds is
pretty well fixed, then variations in. the demand for bonds can only affect
their price. In the quantity theory, it is the price level of goods and services
that is affected when the demand for money does not equal its supply; in the
Keynesian theory, it is the price level of bonds.
Consider this point in more detail. If a single firm is short of cash balances,
it can sell some of its bonds and immediately replenish its stocks of cash. On

the other hand, if the firm has excess stocks of cash, it can invest these forth-
with by buying bonds on the open market. If everyone tries to do this

simultaneously, however, it be possible unless the stocks of money

will not
or bonds change in size. If the stocks of money and bonds are fixed in size,
then general attempts to add to or subtract from bond holdings will only
succeed in altering their price. Assume, for example, that all firms and
households are short of cash. They all try to sell bonds in order to add to
their cash holdings. This causes the price of bonds to fall. A fall in the price
of bonds is the same thing (as we have already seen) as a rise in the rate of
interest. ‘ As the rate of interest rises, people will try to economise on cash
holdings; they will also tend to reduce speculative balances of cash, since
bonds now seem like very good investments.^ Eventually, the rate will rise
high enough so that people will no longer be trying to add to their cash
balances. The demand for money will again equal supply. There will no
longer be an excess supply of bonds, so the interest rate will stop changing.
The net effect of the original excess demand for money will have been an
increase in the rate of interest. Aggregate demand will be affected only in so far as
consumption or investment expenditure is affected by the change in the interest rate.

Now consider a case in which people have too much cash. They decide
to spend the excess on buying bonds ; if everyone tries to do this, however,
they will force the price of bonds up (i.e., force the interest rate down).
When this happens, people itre prepared to hold more cash, both because
the opportunity cost of doing so is reduced and because bonds now look like

1 If you are at all uncertain about this very important proposition you must re-read pages
446-9 now.
2 See Chapter 50, page 697.
716 MONEY BANKING AND THE PRICE LEVEL
a bad buy, so that people will hold large speculative hoards of cash m
expectation of more favourable bond prices the future m
Thus according to the Keynesian theory, a monetary policy that expands
the money supply does not lead to any direct increase m aggregate demand,
as does in the quantity theory, it leads only to a fall m the interest rate
It

untileveryone is prepared to hold the expanded supply of money Any effect

on aggregate demand occurs insofar as investment or consumption expen-
diture responds to the change in interest rates

Et ALUATiON In the quantity theory, a shortage of money causes firms and

households to try to replenish cash stocks by not spending all the money
they receive Thus an excess demand for cash causes aggregate demand for
goods and services to fall In the Keynesian theory, an excess demand for

cash causes firms and households to try to replenish their cash balances by
selling bonds This causes the rate of interest to nse (the price of bonds to
fall) and affects demand for goods and services only insofar as demand for
consumption and investment goods is sensitive to changes m the rate of
interest
The determinants of the demand for money and the effects on the
economy of monetary disequilibrium have both been studied intensively
Many of the problems are beyond an elementary treatment of the subject
There does however, seem to be ample evidence that the demand for

money is influenced both by the value of transactions (which is directly

and by the rate of interest On the effects of
related to the level of income)
monetary disequilibrium, the studies do not seem to support cither of the
extreme views To some extent, both the goods and the bond markets seem
to be affected when firms and households have cither too little or too much
cash If there is an excess supply of cash in the system, there does appear to

be an increase both in the demand for bonds and m the demand for goods
and services, and if there is an excess demand, there appears to be a reduc-
tion m the demands both for goods and services and for bonds The con
tinuance of the present debate about the effectiveness of monetary policy as
a tool of short-run control of the economy reflects the fact that, although
neither extreme position can be held, we are not yet sure where the balance
lies 6etwcen them fs tfie 6uiX m''ifie impactof an increase in the quantity
of money feltm the market for goods and services or in the market for bonds,
and how fast is each effect felt’ Here we must wait and see, and here we
must recognise that the results of a current debate on practical economic
policy turn on future academic work on how the economy behaves m
response to changes m the monetary sector
APPENDIX A TO CHAPTER 51

MICRO-ECONOMIC IMPLICA-
TIONS OF VARIOUS
AGGREGATE RELATIONS
BETWEEN EXPENDITURE AND
PRICES

One of the basic postulates of economics aggregate demand is so low that there
is that decisions are made by firms and is excess supply in all markets. In this

households and that any observed or case, no individual price would be

hypothesised behaviour must be cap- rising and if some prices were falling
able of being traced back to decisions then the whole price level would also
made at this micro-economic level. be falling. Thus we see that the assump-
(See page 540.) In this appendix, we tion that the price level does not fall
shall briefly consider what behaviour implies that price does not fall in any
of individual markets is implied by the individual market when there is excess
relations we have been considering. supply. The second implication con-
The theory of the L-shaped relation has cerns the possibility of the economy
implications about behaviour at the existing at a point such as f (where
micro-economic level that we need to there is a kink in the curve) where full

notice. To study the first relation, turn employment combined with a stable
is

back to Figure 51.1. Concentrate on price level. Full employment implies

the band along the x axis, which indi-
cates that a zero rate of change of prices
is associated with anything less than
full employment of resources. This indi-
cates that the price level does not fall
that there is no excess supply in any
markets; a constant price level (com-
bined with the previous implication
that prices never fall) implies that there
is no excess demand in any market.

even when the percentage of resource
utilisation falls to a very low level. The
micro implication of this is that excess
supply does not cause price to fall in
any single market of the economy. To
see this, consider a situation in which
Thus each and every market in the
economy must be in equilibrium when
the economy is at the point of full em-
ployment without inflation indicated
by/ inthe figure. To see this, imagine
the economy starting at a point of
718 MONEY, BANKING AND THE PRICE LEVEL
heavy unemployment with excess
supply m all markets Now, consider
an increase m
aggregate demand that
causes a nghtward shift m
the demand
curve meach individual market The
theory we are considering implies that
expansion of demand occurs in such a
way that excess supply decreases
equally in each market, for, if it did
not then the pncc level would begin to
rise before full employment had been
reached Indeed, it must nse as soon as
excess demand occurs m
some market
Clearly, the micro-economic relations
for the L shaped curve are very special
ones
There are two micro-economic impli-
cations that follow from the relation
shown in Figure 51 2 First, nouce that
the relation implies that pnces do fall
in individual markets when there is
excess supply The macro economic
relation states that the price level does
fall when aggregate demand, and hence
the degree of resource utilisation, is
very low If the average level of pnces
is to fall, then at least some individual
prices must be falling Some price
ngidities are possible, it is not neces-
sary that all prices should fall when
there is excess supply, but it is neccs
sary that some should
The second implication concerns the
relation among individual markets when
the percentage utilisation of resources
is high To see what is involved, con-
sider an economy with such a low level
of aggregate demand that there is
excess supply in each and every mar-
ket Now consider an expansion in
aggregate demand that raises the over-
all percentage utilisation of resources
The rise m
aggregate demand will
mean a rightward shift m
the demand
curve m each individual market These
shifts reduce excess supply and raise
output and employment m
each part of
the economy As aggregate demand
goes on increasing, excess demand will
develop msome market and price will
begin to rise mthat market This will
happen while there is still excess supply
m other markets As aggregate demand
goes on increasing, more and more
markets will begin to develop excess
demand In everyday language, we
might say that bottlenecks and short-
ages begin to develop m
some parts of
the economy Eventually, pnces will be
nsing in enough markets to offset the
effects of price reductions in other mar-
kets, so that the average level of pnces
will begin to nse As long as excess
supply exists m any market, the Its cl
of resource utilisation can be increased
(i e ,
the level of unemployment
lowered) by raising aggregate demand
But the more markets there are m
which excess demand already exists,
the more any further nse m demand
mil serve merely to increase excess de-
mands in these markets (and so speed
up the rate of inflation) and the less it
mil serve to reduce excess supplies in
other markets (and so increase the
level of resource use) Thus, the higher
the level of a^regate demand, the
greater the effect on pnee and the less
the effect on employment of yet further
increases in demand We see, therefore,
that the micro economic implication of
the relation of Figure 51 2 is that mar-
kets are not m a state of perfect equi-
Jibnum relative to each other To
make the theory work, it is necessary
that markets should be m
a state such
that excess supplies could exist in some
markets while excess demands exist in
others
Should we be surprised by this^ No,
not as long as the economy is subject to
the kinds of changes that necessarily
 MICRO-ECONOMIC IMPLICATIONS 719

accompany economic growth. As pro- markets to be in equilibrium simul-

ductivity grows, some supplies expand
faster than other supplies, and as real
incomes grow, some demands expand
faster than other demands. Except in
the most unlikely of circumstances,
would expect these changes
about a of resources.
reallocation
Since such changes do not happen
instantaneously, we would expect some
markets to be exhibiting excess sup-
plies while other markets exhibit
excess demands.* Because demand and
supply curves are shifting continuously,
due to the reasons mentioned above,
and because adjustments take time to
accomplish, we would never expect all
I This paint is elaborated in Chapter 56. It
was also discussed at some length
the long-term problems of agriculture
Chapter 1 1.
taneously. If the theory' of price is
correct, we do expect that prices -will
work as a mechanism to direct resources
to where demand is greatest so that
we there is a continual movement in the
to bring direction of equilibrium, but since the
equilibrium values are themselves al-
ways changing, we never expect equi-
librium to exist simultaneously in all
markets. Thus we should not really be
surprised that the empirical evidence at
the macro level refutes the idea of the
L-shaped curve shovvm in Figure 51.1
and supports the idea of the smooth
cur\'e shown in Figure 51.2. Indeed,
the micro implications of the L-shaped
curve are so special that it would upset
in relation to
many of our ideas of micro relations in
in the economy if the L-shaped curve
were observed at the macro level.
appendix B to chapter 51

COST PUSH VERSUS DEMAND

PULL: A CASE STUDY^

1 HIS appendix is devoted to a con of prices in the United Kingdom has

sideration of the modem controversy
on the causes of inflation This is an
issue on which there is still much dis
agreement amongst economists, and we
cannot present final conclusions on the
subject Instead, the appendix repre*
sents an attempt to display a current
controversy in its and to
present state
give some idea how one might go about
settling such a controversy
If we wish to test two theories of in-
flation there are a number of things
which we must do first, we must lay
out the competing theories, we must
then examine the theories to sec if they
Yield different predictions, and, finally,
we must seek to test these predictions
by seeing if they do provide adequate
explanations of the price level pbenom
ena to which they arc supposed to refer
The Phenomena to be Explained
We may now ask what is the data that
requires explanation ^ The general level
undergone continuous and sometimes
quite violent changes over the past one
hundred years In only 1 1 years out of
the last 97 was the annua! fate of
change of prices less than a half of one
percent, m 65 years the rate of change
exceeded one per cent, while in 17
years it was greater than 5 per cent In
the 1920 s and early 1930’s the price
level fell, and at least a cursory study
of the data docs not suggest that it was
any less flexible downwards than it
was in the nineteenth century Since
1935, the pnce level has shown a per
sistent tendency to rise and a positive
rale of change of prices has been ob
served every year since the Second
World War It will also be noted that
quite regular patterns can be observed
in pnce level changes, the movements
are rfeiTnitefy not ranefom
A Preliminary Survey
If we, tolerably intelligent students of

1 This appendix is an abridged version of a economics, set out in search of a theory

public lecture delivered at the University of of the price level, we might commence
Liverpool in May 1959 It deals with some ad our search by surveying the financial
vanced topics and can be omitted if desired prras and journals If we did look there,
 GOST PUSH VERSUS DEMAND PULL 721

we would find a large number of plaus- If intuitive plausibility is our only

ible explanations, each one backed by criterion on the basis of which to judge
some facts which gave the appearance theories, we must stop at this point in a
of being evidence supporting the hypo- state of confusion. If we hold different
thesis. In one place we would read that theories, there is nothing we can do but
it is all the fault of the wicked trade appeal to each other’s intuitions, or to
unions. As evidence we might be told what, in our opinion, every reasonable
that in industry X
an increase in the man would surely accept. But, as is
wage rate was very soon followed by an well known, this method of settling
increase in prices. In another journal argument is most unlikely to produce
we would find that it is all the fault of general agreement. If we still find our
the nasty government. Here the writer intuitions differing, there is nothing
will point to government deficits and more to be done at a rational level
other government ‘distortions’ of the except to agree to differ. The argument
economy such as price supports and is,however, unlikely to end here. We
subsidies. Yet another journal would may write polemics directed at each
tell us that price-level changes are all other’s views. We will imply that our
the result of changes in the general opponent is either foolish or else down-
level of demand, and abetween
relation right dishonest or, if we are really adept
cyclical fluctuations in demand and at this sort of argument, we will
changes in the price level will be poin- manage to imply that he is both simul-
ted to as evidence. Finally, if one feels taneously.
inclined to round out the picture by It is the job of the professional econo-
reading the press of yet another shade mist to try to decide between opposing
of the political spectrum, one will read theories and, more fundamentally, to
that inflation is all the fault of the dirty ask how one should set about deciding
monopoly capitalists. Here high and between such theories. There are two
rising profits will be pointed to as evi- points to caution us which we can learn
dence and, in the more sophisticated from our general survey of the press.
organs, it will be argued that the fact First, almost all the explanations offered
that the share of profits in the national are uni-causal ones. Perhaps there is so
income rises during periods of inflation much conflict because each person is
and boom makes nonsense of the ‘cost- looking for the cause of inflation, while
push-by-the-unions’ theory. there are in fact several causes. This is
It is little wonder that, having com- certainly one way of accounting for the
pleted his perusal of this literature, our fact that evidence can apparently be
student finds himself in a state of adduced in favour of several different
complete confusion. There is a large theories which are in flat contradiction
number of theories, some of them over- of each other. A second more general
lapping with others, some completely point is that plausible explanation plus
contradicting others, and, most con- some supporting facts which appear to

fusing of evidence is adduced to

all, constitute evidence is not sufficient.

support each theory so that, especially

in the hands of a skilful expositor, each The world of experience is com-
theory in turn is made to seem over- plex enough so that some facts can
whelmingly plausible. be found which are consistent with
722 MONEY, NKINO AND THE PRICE LEVEL

just about any plausible theory market is first observed in a state of dis-

one cares to dream up. equilibrium with the price at Op Price

now nscs until it comes to rest at Op'
Such an approach is not good enough We may now ask what caused this
and we must set about judging between pnce nsc, we may ask whether higher
these theories in a more systematic
fashion First, we must state the theories
as explicitly as possibleThen we must
examine the theonea to find out where
they differ If, for example, one theory
says that when you find X, pnces will
rise, while another theory says that
when you find X, pnces will fall (or
else that X is not related to pnce changes
so that pnces may either nsc or fall),
then the theones have been made to
produce contradictory predictions
about the world Finally, we must
check these predictions against actual
empirical observations In short, we
make the theories produce testable
hypotheses and we then proceed to test
them

An Analytical Distinction

Before addressing ourselves to these

two theories, there is one simple but
very important analytical distinction
which It IS necessary to make Consider
a single market illustrated in Figure
51B l(i) in a state of equilibrium at
price Op Now assume that the price
rises to rest at price Op We
and comes
may now why this pnce change
ask
occurred, we may ask whether it was pnces were offend by the unsatisfied
brought about by a demand shift, by a buyers or higher pnces were asked for
supply shift, or by some extra-market by the more than satisfied sellers Stan-
disturbance which caused prices to dard pnce theory predicts that, when
change, although neither the demand ever positive excess demand exists,
curve nor the supply curve shifted The pnce will nse, it says little or nothing
pnce change is, of course, consistent about who actually raises the price In
with a shift of demand to IX, of D some markets the suppliers may be
supply to S'S'y or some combination of passive so that the price rises solely as a
these two shifts Now consider the mar- result of the competitive bidding of the
ket illustrated m Figure 51B l(n) The buyers, m other markets buyers may be
 COST PUSH VERSOS DEMAND PULL 723 >

passive and suppliers may raise prices most commonly entertained is that the
as they see that they can sell their rise in factor prices comes because of
existing quantities at a higher price union power; the unions have, and
than is at present ruling. In other exercise, the power to raise wages inde-
markets the price changes may be pendently of the state of excess demand.
brought about by both the ‘pushing’ of Cost-push theory says more than that
the sellers and the ‘pulling’ of the prices can sometimes rise because costs
buyers. It is most important for us to rise (which has never been denied by
realise that we are here asking two anyone) it says that there can occur
;

quite different questions when we ask: increases in the level of money wages
why did the price rise ? first in the case which cannot be accounted for by the
of market situation and, second in
(i) existence of excess demand somewhere
the case of market situation (ii). In the in the economy. It is the particular
first case we are asking: what caused cost-push-through-union-power theory
the disequilibrium ? In the second case on which we vv'ish to concentrate.
we are asking: given that a disequi- The terms cost-push and demand-
librium already existed, who actually pull suggest a distinction based on
raised prices towards the equilibrium? cur\'e shifts, and such a distinction is
The failure to distinguish between these satisfactory' if we are considering goods
two quite different questions appears to markets. Demand-pull sees changes in
have been the cause of much of the the demand for goods as the main cause
confusion in the controversies over of price-level changes, while cost-push
various theories of inflation. sees prices changing when exogenous
changes in factor prices are passed on

The Competing Theories

in the goods markets —
through shifts
in the supply curves for goods. Thus
Now let us consider the two theories of the manifestation of a successful cost
inflation. Looking simply at the causal push will be a supply-curve shift in the
sequences involved, the cost-push markets for goods. If, however, we
theory' says that it is increases in factor conduct our analysis of the factor mar-
prices that cause increases in final goods ket, then the distinction should be
prices, and that these changes in factor between market-force and extra-
prices can occur independently of the state of market-force theories, the former ex-
excess demand',the demand-pull hypo- plaining factor-price changes by the
thesis reverses this order of causation state of excess demand in the factor
and says that it is increases in the de- market, and the latter explaining
mand for final goods (consumers’ goods factor-price changes in terms of forces
or investment goods) that cause in- which are independent of the state of excess
creases in their prices, these price demand. The market-force theory' (de-
increases cause a rise in the demand for mand pull) looks to changes in excess
factors of production which, in turn, demand for factors, which are induced by

cause an increase in factor prices ; thus, changes in the goods market, as the cause
it is increases in the level of goods prices of fkctor-price changes. The extra-
which cause increases in the level of market-force theory (cost-push) says •

factor prices. Cost-push could have a that factor prices can be raised even if
number of possible variants. The one for example, demand equals supply ai
724 MONEY, BANKING AND THE PRICE LEVEL
existing prices These factor pncc This hypothesis is refuted b> a most

changes will cause income changes casual inspection of a time series for
which will cause changes demand m the rate of change of the price level
for goods which will, m
turn, induce
Second version The second, and
changes in the demand for factors, so slightly less crude, hypothesis is that
that the higher factor prices will, after the pushing is related to union strength
some lapse of time, be paid voluntarily To make this testable, we would need a
without necessarily causing any measurable index of union strength
changes in employment This is the (e g percentage of labour force union-
general difference m
causal sequences ised,
,

or the size of union funds) Much

belween the two theories We must work could be done on this hypothesis,
now take each theory in turn, state it but It seems to the present writer that
more precisely, and attempt to test it all acceptable indices of union strength
would show a secular increase over the
Some Cost push Theories Tested last hundred years All of these, then,
would produce the prediction of an
The first thing we must remember is
ever faster rate of increase in wage rates
that the rate of change of prices has and the price level Such predictions
varied considerably from year to year
are clearly refuted by the data referred
If we are to maintain a co$C*push
to m
the beginning of this appendix
^

theory, that u is exogenous changes in

factor prices, particularly in wage rates, Third version The third of the pos-
that cause these price-level changes, sible cost-push hypotheses is that the
then we must have a theory of vartcitons pushing IS related to changes in the cost
in the strength of must
the cost-push We of living (and, since movements in the
ha\ c a functional relation between price level of goods consumed by
pushing and something else If we do workers conform pretty closely to
not, we will be able to say after the movements in the general pncc level,
event that there was a great deal of that puling varies with the pncelevel)
pushing one year because prices rose a This vanant gives rise to the famous
great deal and that there was very little cost push spiral prices are determined
pushing mother year because prices rose
t I have tert this passage as a was written in
only alittle But this is only name-calhng
1960 in order to illustrate the power of positive
and we will be none the wiser for this economics Mr B Hines has taken up the
change in nomenclature What then challenge of this paragraph and has refuted it
are the possible hypotheses about the empirically He has produced a measure of union
strength of the cost-push^ strength which does not have the properties I
jtuessed it must have and has shown that his index

First VERSION The first
cost-push hypothesis is that unions al-
ways push upwards on wages with a
uniform pressure This theory produces
the testable prediction that wages, and
therefore the general price level, will
show a steady upward trend (with ran-
dom variations around that trend)
and simplest of union strength correlates very well with the
rate of change of money wages See B Hines
Unions and the Change in Mone) Wages'
Review of EcMomic Sludtes 1962 Whatever is the
fina l outcome of the discussion on the meaning
of Mr Hines discovery there can be no doubt
that he has discovered a most interesting relation
that requires explanation and that he has clearly
refuted the contention in the present text
 COST PUSH VERSUS DEMAND PULL
by wages (e.g., price = cost + standard
mark-up) and wages are determined
by prices (a 1 per cent increase in prices
causes a 1 per cent increase in wages).
Now the world can be in neutral equi-
librium at any price level but, as soon
as some exogenous factor causes prices
to rise, there will develop an explosive
situation in which wages chase prices
and prices chase wages in a never-end-
ing spiral. Assume that an increase in
wages causes an equal increase in prices,
then, if we are to have a cost-push
spiral, an increase in prices must cause
an equal increase in wages. If this is
true, then a 10 per cent exogenous in-
crease in prices will cause a 10 per cent
increase in wages which will cause a
further 10 per cent increase in prices,
causing a 10 per cent increase in wages
and ad infinitum. This is probably
etc.,
the most seriously entertained cost-
push theory.^
One approach to testing the cost-
push spiral is to attempt to measure
directly the relationship between a
change in prices and a change in wages.
The first attempt to measure this rela-
tionship was made by two economists
at the National Insritute of Economic
and Social Research. These economists,
C.Dow and L. Dicks-Mireaux, con-
cluded from a fairly detailed and so-
phisticated study of the post-war data
of the United Kingdom that the coef-
ficient relating wages to prices could
not be more than -5, which means that
a 1 per cent rise in prices causes only a
\ per cent rise in wages.^ In another
1 Clearly, then, the theorj' requires either that
the economy isnot up against a monetary con-
straint or that new money is created by the central
authorities.
2. L. Dicks-Mireaux and C.Dow, ‘The De-
terminants of Wage Inflation in the
Proceedings of the Royal Statistical Society, 1959.
725
Study the present writer estimated the
coefficient between the years 1923 to
date (omitting the period of the Sec-
ond World War) to be -67.^ Both of
these direct estimates are significantly
than one, and both show that there
le.ss
is no factual evidence of the existence
of an explosive wage-cost spiral in the
United Kingdom during this period."^
Taking the Dow and Dicks-Mireaux
figure, an exogenous rise in prices
would set off a wage-price spiral, but it
would be of the rapidly damped sort
familiar in multiplier theory. The
wage-price multiplier in the Dow-
Dicks-Mireaux case is, of course, two,
so that the feedback from prices to
wages and back to prices again would
double any exogenous price change
that occurred. Thus an exogenous rise
in prices of 10 per cent would lead to a
5 per cent rise in wages; if prices then
roseby the full 5 per cent wages would
rise by 2j per cent and if all this were
passed on in higher prices then wages
would rise by a further 1^ per cent and
so on. Recently Mr Dicks-Mireaux has
made a more sophisticated estimate of
this relationship.^ The results are
3 R.G.Lipsey, ‘The Relation between the
Rate of Change of Money Wage Rates and the
Level of Unemployment in the United Kingdom,
1862-1957: A Further Analysis’, Economica
(February 1960). There are reasons for believing
the figure of -67 to be a considerable over-
estimate of the ‘true’ value.
4 Both of these are maximum estimates; they
assume that all observations of prices and wages
rising togetherwere caused by wages chasing
prices and not prices rising because
wages had
risen.
5 two-equation model was used and a two-
A
stage, least-squares estimation
procedure em-
ployed. This technique allowed for the fact that
not only do prices affect wages, but wages
affect
U.K.’,
prices.
i 726 MONEY, BANKING AND THE PRICE LEVEL
broadly similar to the other studies A test of this theory amongst individual
already mentioned
‘
industries gave no support for this
So far all that we know is that any th«)ry ^ There isno evidence that wages
j
exogenous change in the price level nsc particularly fast in years when busi-
^
will be magnified by this feedback ness IS particularly profitable and vice
effect of pnces on wages If we want to versa This theory appears to be re
go further, we have two possibilities futed for the United Kingdom
( either we take the exogenous changes
in prices as random or assume them to Fifth version The final venion of
be functionally related to some other the cost-push theory that we need con
variable in the system If we adopt the sidcr IS that pushing is related to the
first approach we get the testable hypo- level of business activity On this
thesis that pnce-level changes are ran- theory, unions push hard when busi
dom This hypothesis is clearly refuted ness activity is high and less hard, or
by the observation of systematic varia- not at all, when business activity is low
tions m
the price level If we adopt the Thus we have the hypothesis that
second approach we are right back JW=^((A), where dlF is the annual
where
1 we started We are without a change in wage rates, where A is some
theory of the price level until we have index of the level of activity, and where
a theory of what we have called the f IS such that the two are positively re
‘exogenous changes’ Given that there lated One very good index of the level
ISa systematic movement in the price of activity IS the percentage of the
levelwe are no closer to explaining it labour force unemployed (U) Since
The measurement of the coefficient re- considerable empirical work has been
lating wages to prices does nothing to done on relations between wages,
provide us with a theory of infiation as prices and unemployment, we will use
long as the coefficient is less than one, this index of activity The theory there
and It does refute the theory of the fore produces the testable hypothesis
explosive cost-push spiral We
will postpone consider-
ing tests of this hypothesis until we have
Fourth VERSION A fourth version of investigated the demand pull theory
the cost push theory is that the strength
of the cost-push is related to business The Demand-pull Theory Tested
profits According to this theory, labour The demand-pull theory looks to fluc-
unions push strongly upwards on wages tuations in the level of demand for both
when business profits are high and less consumers’ and capital goods as the
strongly when business profits are low causal force operating on prices and
Jaclor ensLs * Urxoand goes lyj, prices
Note that the value ot this coeffiaenl does
tot imply that real wages have fallen over ihe 2 S«e RG Lipsey and M D Steuer The
last ten years money wages have nsen greatly Relation Between Profits and Wages m the
or reasons other than this price level effect The United Kingdom ,
Ecanomua May 1961
/alueofthis pnee level coefficient merely means 3 This IS not name calling nor is it a dead
hat a fall in national living standards due say to end theory for we do have fairly well worked
i rise in the prices of imports is not wholly ont theoncs of fluctuations in the level of de
tvoided by wage earners by increases in thesr mand for capital and consumption goods See
money wages for example Chapters 41 and 45
 COST PUSH VERSUS DEMAND PULL 727

are bid up, production becomes more could be associated with changes in un-
profitable, demand for factors of pro- employment (and hence demand for
duction goes up, and so their prices are labour). My own more detailed study
bid up. This theory produces the hypo- of the same period was suggested by the
thesis that variations- in prices will be work of Phillips. I found that 80 per
associated wdth variations in the level cent of the variance^ in money wage
of demand. This theory' could be tested rates over the period1861-1913 could
by considering the markets for factors be explained by variations in the level
of production or the markets for final of unemployment. In the period after
goods. We shall concentrate here on the First World War, the association
the factor market in order to make our between wage rate changes and de-
results comparable to the earlier tests mand changes was somewhat less
of cost-push and in order to be able to marked, with only 60 per cent of the
use some existing empirical work. The variance in wage rates being explained
theory implies that fluctuations in fac- by variations in demand. The evidence
tor prices including, in particular, then is that there is a fairly close asso-
labour prices, are associated wfith fluc- ciation between changes in the level of
tuations in the demand for factors. demand and changes in wage rates,
Since there is a strong positive correla- but that this association issomewhat
tion beUveen changes in demand for less close in the twentieth century than
labour and changes in the percentage itwas in the nineteenth."^ This evidence
of the working force unemployed (
U) - means that, if we know the state of
an increase in demand causing a fall in excess demand for labour, we will know
U - the hypothesis is that AW={{U)} approximately what is going to happen
Now we may consider the tests of to wages.
Three works may be
this hypothesis. There are a number of points to be
mentioned here: that of Dow and
Dicks-Mireaux, Phillips, and Lipsey. 3 For those unfamiliar with statistics ‘vari-

Dow and Dicks-Mireaux, in the de- ance’ may be read approximately as ‘variations
tailed study of post- 1945 data for the in’. To be precise, the variance is the square of

United Kingdom already mentioned, the standard deviation (cr)

found that a high proportion of the

variations in money wages was associ- a" = 1 £ {AW,-AW)\
n.-i
ated wth variations in the level of
excess demand for labour. Phillips, in where AiV, is the rate of change of wages in each
his study of United Kingdom data for of « years, and AH' is the mean rate of change of
a period of almost one hundred years,^ wages over all of the n years.
concluded that a very high proportion 4 The reduced association seems to be the
result of an increased sensitivity of wage rates to
of the variations in money wage rates
price changes. The two variables, unemployment
1 The correlation will now be negative because and prices, explain about 90 per cent of the vari-
U falls as demand rises. Thus A fV is expected to ance in wages over the whole period, but whereas
be large when U is small. in the nineteenth centurj’ unemployment ac-
counts for almost all of it, in the twentieth
2 A. W. Phillips, ‘The Relation Between the
Rate of Change of Money Wage Rates and the century this variable’s explanatory power is

Level of Unemployment in the United Kingdom, reduced somewhat, while that of price changes
1861-1957’, Economica (November 1958). has definitely increased.
728 MONEY, BANKING AND THE PRICE I EVEL

made about this demand pull hypo- stitute control of the labour
ejective

thesis First %ve must note that the test- market’ This question is critical if we
able prediction following from this are considering government interven-
theory IS the same as that from the hfth tion, possibly m
the form of an incomes
version of the cost-push theory given policy, to prevent wages from rising in
above The two theories m fact both spite of excess demand This is a ques-
stand or fall on the same of obser- set tion about the dynamic behaviour of
vations It IS true that, in the past, the labour market and is related to the
some people have argued that the em- discussion of Figure 51B I(ii) rather
pirical observation of the relation than SIB l(i) The question of how to
‘proved’ cost-push, while control the wage rises turns partly on
others argued that the same observa- the question of who it is, the unions or
tion ‘proved’demand-pull The answer the employers, who actually raise wages
is produce the
that, if the t\so theories when there is excess demand Impor-
same hypothesis, we cannot discrimin- tant though this question is, however,
ate between them by testing this hypo- we we do
shall cause great confusion if
thesis, we can discnmmate between not keep completely distinct from the
it

them only jf the two theories can be question ^Vhat is the cause of the
made to produce other predictions which excess demand, a supply shift, a de-
contradict each other *
mand shift or some exogenous change
Given that both theories produced m price ’ If it is true that, when there is
the same hypothesis, A\V—i{V), why excess demand for labour, both sides
is It that the observation of this relation exert pressure on wages, then wages
might cause controversy’ Both sides will rise unless both sides are restrained
must accept the existence of excess de- If, however, only one side is active and
mand m the factor markets associated the other is totally passive, then re-
with an increase in the wage for labour, straining one side would be sufficient to
and the larger is the excess demand the stop the wage rise even if there were
faster is the rate of change of wages excess demand
The question now arises Is excess de- The picture wc have arrived at is
mand a sufficient condition for wage that the major part of the pnee and
rises’ Clearly no one would argue that wage nscs in the United Kingdom can-
it was have been
sufficient since there not be accounted for by the exercise of
times, particularly during wars, when union power independent of the state
governments ha\e effectively controlled of the market Wages rise mainly as a
factor markets, preventing wages from result of excess demand If, however,
rising in spite of substantial excess de- wc WTshed to break this sequence by
mand for labour This observation restraining one side or the other in the
raises the question What sort of wage bargain, we would be successful
government intervention would con- only if one side is active in raising
wages m
the face of excess demand We
1 Of course, part of the trouble stems fitiin
could now do another study of the
the belief that observations can prom rather than dynamic behaviour of the labour mar-
roetely /ail to itfuit theones Boih theories arc ket with a view to seeing how it could
merely not refuted by the empirical observation be controlled, but this would be quite
AW=({U) distmet from the question we ha\ e
 COST PUSH VERSUS DEMAND PULL
been asking about tbe market con-
which are associated with wage
ditions
rises.
The realisation that questions asso-
ciated with Figure 51B.l(ii) may
have been at issue gives us a clue
that enables us to assess another piece
of evidence often quoted by both sides,
but more often by the ‘it’s-all-the-fault-
of-the-unions-cost-push-school’. This
the argument from temporal sequence.
We will be told, for example, that in
industry X wage rates were increased
substantially when the new contract
was signed, and that a very short time
later the price of the product was
raised. (It may
or may not be added
that, in the annual report, the chair-
man mentioned increased wage costs as
the reason for increasing prices.) But
can weaccept this temporal sequence
— increases in wages followed by in-
creases in prices —
as evidence of the
causal sequence —
increases in wages
cause increases in prices? In an infla-
tion there will generally be excess de-
mand both for consumers’ goods and
for factors of production. Now all that
stadc price theory predicts is that prices
will tend to rise in both the goods
markets and the factor markets. No-
thing is predicted about the relative
speed of price changes in the two sets
of markets. Also, if the price changes
are not continuous but take place at
discrete intervals of time, as will be the
case in union-dominated labour mar-
kets and in markets for manufactured
goods, the theory predicts nothing at
all about the sequence in which these
price changes will occur. Clearly, if
both unions and management are con-
cerned about public relations and do
not wish to appear to be the cause of price
increases, it will be in the interests of
each to try to have their price increase
729
follow the price increase of the other
side. In this case the unions whose con-
tracts come up for renewal at a definite
time are at a very strong disadvantage.
The obvious thing for the employer to
do is to time his price rise to follow, not
to precede, the rise in the factor price.
This would be advisable both because
he then knows what his costs will be,
is rather than having to guess at them,
and because it tends to shift the blame
for the price rise on to the union. What
is not mentioned in this sort of argu-
ment is that the existence of excess
demand in the goods market might
have caused a rise of prices in that
market whatever had happened in the
factor market. We must conclude,
therefore, that when both the goods
and the factor markets are in disequi-
librium, any observations of the time
sequence of price changes in the two
markets must be interpreted with
extreme caution.
Conclusion
We may now summarise our conclu-
sions. Two naive variants of the cost-
push hypothesis are probably refuted.
The third version is the wage-price
spiral. There seems to be no evidence
consistent with the strong version of
this theory that wages and prices chase
each other in a never-ending spiral.
The evidence is, however, consistent
with a weaker version of this theory,
that unions do have some power to
raise wages in response to price rises,
independent of the state of excess de-
mand, but not sufficient power to raise them
by the full amount of any price rise. Finally,
we must note that this weak version of
the theory does not provide us with a
theory to account for systematic varia-
tions in prices. All it says is that the
 —
730 MONEY, BANKING AND THE PRICE LEVEL
existence of exogenous (and random) rather than over the question of how
disturbances to the price level will the disequilibrium was brought about
cause the wage rate to move in ways The tentative picture which emerges
that cannot be predicted by variations from this analysis is as follows Varia
m the level of demand alone Thus de tions in wage rates are associated signi
mand will not correlate perfectly with ficantly with variations in the level of
wage rates This prediction is m fact excess demand A theory of systematic
borne out The fourth version, that the variations m the demand for goods
cost push IS associated with business (eg, a multiplier-accelerator theory)
profits, is refuted The fifth version of will thus also provide a theory of syste-
cost-push IS found to be formally iden- matic changes m the level of costs and
tical with the demand-pull theory, and prices The pnee level, however,
the evidence is found to be consistent changn for other reasons such as exo
with the weak version of this theory genous changes in import prices Wage
that a significant proportion (but by rates m this century are sensitive to
no means all) of the variations in wage these movements m the pnee level and
rates can be accounted for by variations nsc to a wage-price spiral,
this will give
in the level of excess demand We may but only to one which will be rapidly
also conclude that much of the argu- damped so that it multiplies the exo-
ment has probably been over the ques- genous price change by a factor roughly
tion of who actually changes wages between two and three but will not set
when demand is not equal to supply off a never-ending cost-push spiral
 PART 9

THE INTERNATIONAL
ECONOMY
CHAPTER 52

THE GAINS FROM TRADE

In Chapter 46 we considered the effects of trade on the circular flow of

income. We were then emphasising the effects of trade on aggregate expen-
diture and regarding imports as a withdra^val from total domestic
%ve -were
expenditure and exports as an addition to total expenditure. \\^e now wish
to enquire into the problems of international trade in a more fundamental
way. The sequence will be as follows: in this chapter we ask if there are
gains from international trade and if so, what is the source of this gain
in the Appendix we ask if free international trade will lead to a pattern of
imports and exports that will allow the potential gains from trade to be
realised. In Chapter 53 we enquire into exchange rates, asking what they
are, what they do and why they var>'. In Chapter 54 we study reasons that
have been advanced for interfering with the free flow of trade that emerges
from the operation of the unhindered price system. In Chapter 55 we give
a brief account of some of the most important experiences in the world of
international trade.

THE VALUE OF FOREIGN TRADE

The foundations of modern economics were laid by men intimately con-
cerned with the problems of foreign trade. T\vo of the earliest schools of
European economists, the Physiocrats and the Mercantilists, took strong
and divergent stands on the value of foreign trade. Economists in eighteenth-
century France were responsible for major reforms in the government
machinery for regulating both internal and foreign trade. The great
eighteenth-century English philosopher and economist, David Hume, who
was one of the first to work out the modem theory' of the price system as a
control mechanism, developed his theory' mainly in terms of prices in
foreign trade.* Adam Smith and David Ricardo, the two English econo-

1 There is very little in Chapters 5-1 1 that would have been strange to Hume. With great

theoretical insight, he perceived how a price system worked and stated its basic principles.
734 THE INTERNATIONAL ECONOMY

mists who developed to its full height the classical theory of the functioning
of the economy, were greatly concerned with problems of trade Smith,
writing m
1776, developed a nnging attack on government intervention in
foreign trade and was personally responsible for many reforms the branch m
of the British civil service concerned with the control of trade Ricardo,
wnting m 1817, developed the basic theory of the gams from trade that is
still maintained today The repeal of the Com Laws, and the transformation
of Bntain in the mid nineteenth century from a country of high tanffs to
one of complete free trade, was to some extent the result of agitation by the
economists whose theories of the gains from trade led naturally to a con-
demnation of all tanffs
In the heyday of free trade between the repeal of the Com Laws and
1930, It was accepted by most economists that free trade was the best
policy for all governments at all times (assuming that their policy goal was
to have living standards as high as possible) Newly developing countries
who were late arrivals in the industrial field were never so certain about
this Classical advice Germany under Bismarck accepted the advice of the
economist Frederick List who advocated tanffs as a means of economic
development Of course it is possible to ai^e that German economic
development would have been even faster if she had adopted a free trade
policy, but there is no doubt that by the eve of the First World War, high-
tariff Germany was overhauling free-trade Britain as the world s leading
industrial producer Dunng the 1930’s, under the impact of the most
disastrous depression of recorded history most countries of the world adopted
high tanffs and even Britain abandoned her 100-year old policy of free
trade After the Second World War there was a reaction against the high
tanff policies of the 1930's, the most significant move back towards free
trade being in Europe where the six countnes of the European Economic
Community adopted a policy that was intended to remove all tanffs on
trade between themselves well before the end of the century
What IS all the argument about and who is right - the free traden or the
protectionists'^

THE GAINS FROM RESOURCE RE ALLOCATION A SIMPLE

EXAMPLE
If there were no trade, everyone would have to produce what he required
for himself, with trade, vanous groups of people specialise m
the production
of different commodities Trade therefore results in a different allocation of
resources than would occur if there were no trade, and, if we are to search
for the source of the gams from trade, we must look to the gams from
resource re-allocation that result from it
 THE GAINS FROM TRADE 735

Let us first consider a simple example relating to resource allocation

within one country. Assume that you are the economic dictator of a country
and that you wish to increase the production of X. Assuming that there is
full employment, you will have to produce less of something else if you are

going to produce more of X. Let us say that you have decided to reduce the
production of good Y in order to free the resources necessary to produce
more of good X.
Now assume that there are two distinct groups of labourers producing Y
at present, and that the two groups have different capabilities. Your
problem is to decide from which of the two groups to take workers in order
to raise the production of X
by some stated amount, say, 16,000 units per
year. Let us imagine two situations in the first situation you would prob-
;

ably have no trouble in solving the problem.

Table 52.1

SITUATION I

Existing annual production Potential annual production

of Y per person of X per person

Group A 1,000 1,600
Group B 500 2,000

Obviously you would move eight workers from Group B, getting your
16,000Z at a loss (opportunity cost) of 4,000 F. The opportunity cost of LF
when Group-B workers are moved is clearly \Y. If you had been so silly as
to move Group-A workers, you would have had to transfer 10 workers,
giving a loss of 10,0007. When Group-A workers are moved, the oppor-
tunity cost of a unit of Y f of a unit of 7 ( = 10,000/16,000). There is a
is

clear loss of an extra 6,0007 if Group-A workers rather than Group-B

workers are transferred.
Now consider a second situation.

Table 52.2
SITUATION II

Existing annual Potential annual

production of 7 production of X
per person per person

Group A 1,000 1,600

Group B 1,500 2,000

Stop and think from which group you would now move workers in ordc
7. Nex
to produce an extra 16,000Y with the least possible loss of
736 THE INTERNATIONAI. ECONOMY

yourself from that group and not from the other than you select
why it is

your workers The student who answers these questions before reading on
will have gone a long way towards, discovering from himself the idea of
COMPARATIVE ADVANTAGE
The most obvious thing to do would seem to be to transfer workers from
Group B, because they are more effiaent at producing X than are Group-A
workers But you must also consider the efficiency of each of these groups at
producing Y What matters is not just how efficient is each group at pro-
ducing X, but how much Y you lose for every extra unit of you get with X
each group The student should now calculate the loss in output of }, per
unit of X
produced, for each group of workers, he should then ask himself
again from which group should I move workers^ Clearly, it is Group-A,
not Group B workers who should be moved, with Group-A workers the loss

of Y per unit of X gamed than with Group-B workers Let us check

is less

this calculation If we transfer Group-A workers we must take 10 workers

to get our 16,000 extra X, and the loss is thus 10,0007 If we take Group-B
workers we only need to transfer 8 workers to get our 16,000^, but by so
doing we lose 12,0007 Thus there is a clear extra loss of 2,0007 moving m
Group-B workers rather than Group-A workers When Group-A workers
are transferred, f of a unit of 7 is lost for every extra unit of produced, X
when Group B workers are transferred, ofa unit of 7 is lost for every extra
J
unit of X produced
Now the moral of this story is that it does not matter that Group-B
workers are absolutely more X than are Group-A
efficient at producing
workers What matters is two groups in
the comparative efficiencies of the
both because
lines what we need to compare is the amount of the loss of one
commodity, per unit gamed of the other, for each group of workers In
other words we need to know the opportunity cost of a unit of in terms of 7 X
for each group and we naturally choose the group with the lowest oppor
tunity cost
The conclusion that Group-A rather than Group B workers should be
- . . , . ,
annual production of 7
transferred depends on the fact that the ratio ; -r-r, is less
annual production of X
for Group A than for Group B annual production of both
If the and 7 is X
changed m
the same proportion, there is no change these ratios The m
student should check for himself that multiplying by the same amount the
annual outputs of both X and 7 for Group-A workers does nothing to affect
the conclusion that it is Group-A workers that should be transferred rather
than Group-B workers
Now let us consider,
using the data of Table 52 2, if there is any gain from
trade between our two groups, A and B Assume that the Group-B workers
 THE GAINS FROM TRADE 737

get and say: ‘This association with the inefficient Group-A

together
characters is down our standard of living; after all we can
only dragging
produce both X and Y more efficiently than they can. Let us split awa)' and
form our own economically independent state with our own dictator; then
those poor Group-A workers can wallow in tlieir own poverty without
dragging us down with them.’
two groups form two independent self-sufficient countries there is
If die

no doubt that the people in B will have a higher standard of living than
those in A. People in B can produce a higher per capita output of both X and
Y than can those in A. Now let us suppose that we have our two indepen-
dent countries,A and B, both producing the X and Y needed to satisfy their
own requirements. Is it not possible to do better? A glance at Table 52.2
will show that it is possible. Country B sacrifices
f unit of Y for every X it
produces, while A only sacrifices of a unit. Now do the division the other
way and find the amount of X given up for every unit of Y produced in :

country A
f while in
it is B it is only f A". Thus B has a lower opportunity
cost of producing Y while A has a lower opportunity cost of producing X*
Now what would happen ifB produced less X and more 7 while A produced
less 7 and more A”?
We assume that, in country B, two labourers are moved from the pro-
duction of X to
that of 7; the gains and losses resulting from this re-alloca-
tion of resources are recorded below. Wc
also assume, that in country A,
three labourers are moved from the production of 7 to that of the gains X ;

and losses are also recorded in the table below. There is, in this case, a
clear gain of SOOX per year. (Remember that we have not changed the
total volume of employment in A or B, we have merely re-allocated it

between the production of the two commodities.)

Table 52.3
CHANGES IN PRODUCTION CAUSED BY A
RE-ALLOCATION OF RESOURCES
Change in annual output
Y X
3 labourers in A are moved from 7 to X —3,000 -t- 4,800
2 labourers in B are moved from X to 7 -1-3,000 — 4,000

NET GAIN 0 +800

The example of Table 52.3 was constructed so that there was only a gain in
X but it is always possible to move resources so that there is an increase in

.1 Question: Is it ever possible for B to have a lower opportunity cost of producing both X
and Y than does A?

24
738 THE INTERNATIONAL ECONOMY

the production of both X and Y To check on this the student is left to

calculate for himself, the gams and A are moved from
losses if 8 labourers m
producing Y to producing X while, at the same
time 6 labourers in B are
moved from producing X
to producing Y The changes in output can be
calculated from Table 52 2 and should be recorded a table similar to m
Table 52 3 This is a most important calculation and no one should proceed
without completing it

We now see that the inhabitants of B were wrong to believe that they
could do better by being self-sufficient than by trading with A, if B pro-
duces more Y and A produces more X, the total X and Y
production of both
can be increased Thus there is more to go round for everyone However,
since A is producing more X and B more Y than when they were self-
sufficient, It will be necessary, so that consumers m each country can have
the X and Y they desire, to trade A’s T*for B’s Y It is necessary, that is, to
engage m international trade
This IS the basis of the gams from trade different opportunity costs m
the two countries It matters not one bit whether or not one country is

absolutely more efficient than the other m all lines (e g ,

Table 52 2), or
only in some lines (e g ,
Table 52 I) The gams from trade depend only on
there being different relative costs — hence the term comparative costs
Let us pursue this a step further by expressing the data m Table 52 2 in a
slightly different form

Table 52 4

OPPORTUNITY COST RATIOS

(Derived from Table 52 2)

Amount oj y (hat must Amount of X that must

be giien up to get one unit of X be given up to get one unit of Y

Country A iy iix
Country B lY liX

Now It IS as simple as this if B produces one more unit of Y the loss is if

A produces one less unit of Y, the gam is lyA', Y production is unchanged and
X production has increased
We may now state these propositions more formally

Definition A country has a comparative advantage vis-a-vis a second

country in the production of the commodity which it has a lower m
opportunity cost than the other country
 THE GAINS FROM TRADE 739

Proposition: If two countries are both producing some of two com-

modities, then the total output of both commodities can always be increased
if each country transfers resources into the production of the commodity in

which it has a comparative advantage.

Comparative advantage must not be confused with absolute advant-
age: we say that a country has an absolute advantage in the production of
some commodity if, with a given amount of resources, it can produce more
of that commodity than can the other country using the same quantity of
resources. Thus in Table 52.1, A has an absolute advantage in 7, while B
has an absolute advantage in A; in Table 52.2, B has an absolute advantage
over A in both X and 7. That comparative advantage is independent
of absolute advantage can be seen by multiplying A’s figures for X
and 7 production in Table 52.2 by the same amount. No such multiplica-
tion will affect the opportunity costs expressed in Table 52.3. If, for
example, A’s figures are multiplied by 10, giving 10,0007 and 16,OOOZ as
the annual outputs of a labourer in these two industries, then A becomes
ver)' much more efficient in both X and 7 production than B, but the
opportunity cost of X in terms of 7 or of 7 in terms of X is unchanged. The
figures in Table 52.4 and the argument immediately below it remain quite
unchanged.*
The only time when it is not possible to get more production by re-
allocating resources when both countries are producing both commodities,
is when the opportunity cost is the same in both countries. Assume, for
example, the figures in the following table. Now if A produces one more 7,
then 2X is sacrificed, while if B produces one less 7, then 2X is gained,
making a net gain of zero for both X and 7.

1 Text-books of economics often quote numerical examples in which the total quantity of
resources is stated, and divided in some arbitrary way, between the production of and that X
of Y, when A and B are self-sufficient. They then compare this output with the output resulting
from a situation in which A produces only one good and B only the other. It is then usually
found more of one good and less of the other than in the
that, in the latter situation, there is

former situation, and it to go through some valuation process to establish

becomes necessary
that specialisation raises the value of production. Such examples tend to obscure the funda-
mental issue: Specialisation according to comparative advantage makes it possible
to produce more of all goods; if both countries are producing some of both com-
modities, it is always possible to re-allocate resources so as to get more of both
goods.

So far we have only given an example. A simple algebraic proof follotvs In Country
: A the

opportunity cost of 1 T is aX, while in B it is bX. Assume that a>b. It follows that the oppor-

tunity cost of 1 AT is 1/a in A and 1/i inB and that \la<Xlb. Thus A has a comparative ad-
vantage in producing X and B in Y. Now if B produces one more Y she sacrifices bX, while if A
produces one less Y she gains aX. As a result Y production is unchanged and A production
changes by aX — bX~X{a — b") which is positive since a>b. The increase in X depends on the
magnitude of the differences between the two countries in the opportunity cost of producing X.
740 THE INTERNATIONAL ECONOMY
Tadle 52 5
SITUATION III

Amount of Y that must be Amount of X that must be

guen up to get a unit of X gtien up to get a unit of Y

Country A ^1 2X
Country B 2X
The importance of comparati\e rather than absolute costs can be seen by
considering the cost data which could have given rise to the opportunity
Table 52 5 If in country A a year’s la^ur could produce 200^ and
costs in
1001, would give nsc to the costs shown in the table In country B a
this

year’s labourmight produce 20^ and 101 in which case A would have an
,

absolute advantage over B in both commodities, but no comparative advan-

tage On the other hand, a year’s labour m
B might produce 2,OOOX and
I OOOF, in which case B would have an absolute advantage over A in both
lines of production, but there no comparative advantage We can
is still

multiply the output of a year’s labour in A and Y by the same amount

without aH'ectmg the data in Table 52 5
We have shown that there can be a genera! increase in production if the
dictators of groups, now countries, A and B get together and decide that
each country shall produce more of the commodity in which it has a com-
parative advantage How they then decide to divide up this gam will
probably be a political problem, and will depend on the relative bargaining
strength of the two dictators
It must be noted that, m all the examples used so far, we have implicitly

assumed constant costs of production, that the output per hour’s work by each
group in each industry was independent of the total amount being produced
The gams from specialisation according to comparative advantage come as
long as opportunity costs differ, they do not require that costs should change
when specialisation occurs In the case of economics of scale there is an
additional source of gam If output perman rises when more is produced
then there will be extra gain when each country produces one commodity
and satisfies its demand for the otherthrough trade Trade thus may
increase world production by allowing
for the exploitation of previously
uncxploited economies of scale In the simple theory of international trade
we shall continue to assume that costs are constant, that a year’s labour
produces a certain output of A” or F independently of how much labour is

already devoted to the production of either of these commodities

APPENDIX TO CHAPTER 52

DOES THE PRICE SYSTEM LEAD

TO A PATTERN OF TRADE WHICH
ACCORDS WITH THE BALANCE
OF COMPARATIVE ADVANTAGE?

In the body of Chapter 52 we showed 52, Table 4. The figures in the second
that potential gains from trade arose column are of course just the reciprocal
wherever there existed comparative of the corresponding figures in the first
advantages between different countries column. (Any student who is not
in the production of various com- absolutely certain why this is so must
modities. We next need to ask if the go back and re-read the previous
price system would lead countries to chapter.)
goods in
specialise in the production of
Table 52A.1
w'hich they had comparative advantage
and import those in which they had a Opporlunity Opportunity

comparative disadvantage. In other cost of X in cost of Y in

words can we expect the unhampered terms of Y terms of X

price system to realise the potential Countr)' A fT IfZ
gains from trade?' Country^ B |F HX
We have alreadt' shown that countiy' A
has a comparative advantage in the
EXCHANGE RATES IN A SIMPLE production of X
and countiy' B has a
TWO-COUNTRY, TWO-COM- comparative advantage in the pro-
MODITY WORLD duction of Y.
We can answer this question by con-
sidering the simplest possible case with THE RELATION BETWEEN
two countries and two commodities. OPPORTUNITY COSTS AND
Let us repeat, since we are already PRICES
familiar with it, the example of Chapter
Now, in order to see what wall happen

1 This Appendix is difficult and it can be in a market society, we must discover

omitted if you are prepared to accept that the what wall be the prices of the two com-
answer to this question is yes. modities in each of the two countries.
742 THE INTERNATIONAL ECONOMY
It IS convenient to start by assunung
that the tv,o countries do not allow
trade between each other so that each
country must produce those quantities
of A’ and Y that it requires for its own
consumption
The proposition which we shall prove
below IS as follows under competitive
conditions it will be true for any one
country that
Opportunity cost of X in terms of Y
marginal money cost of production
of a:
marginal money cost of production
of y two units of y which were foregone
money pnce of X
money pnce of Y
The opportunity cost of X m terms of
y is the number of units of I that must
be sacrificed in order to produce one
more unit of X (by moving resources
out of y production into X production)
This IS a physical measure Now, under
competitive conditions this will be
equal to the ratio of the (money) costs
of producing one more unit of each
commodity Why’ Under competitive
conditions any factor of production
must be paid the same in both the
and the Y industries (otherwise the
factor would move from the low-paid
industry to the high paid one) Thus,
if It costs twice as much to produce a
unit of ^ as It does to produce a unit of
\ 'ifeiWi Vv
, Twice iVit -qaaifcrty
of factors to produce a unit of ^
as a
unit of 1 In this case, if sufficient
factors are transferred from Y in order
to produce one more unit of X, two
units of y will be lost (The reader
should pause here and think out care-
fully for himself the argument of die
last two sentences Thus, in general.
)
the opportunity cost ratio will conform
to the ratio of the marginal costs
Under competitive conditions the ratio
of the money prices of the two com-
modities will be the same as the ratio
of the marginal costs If this is not the
case then firms will transfer resources
from one line to the other and thereby
increase their profits Assume, for ex
ample, that the marginal cost of X is
twice that of F, while the pnce of A n
three times that of Y Now, by pro
ducmg two less Y and one more X the
firm’s outlay will be unchanged but its
revenue will be increased because the
one unit of X
sells for more than the
Under these conditions profit-maxi-
mising firms will move resources out of
y and into X
The pnce of {Px) will X
&ll as Its production increases and the
pnce of y (Py) will nse as its produc-
tion diminishes This movement will
continue until px is only twice Py^ in
which case there will be no gain in
producing more X
and less 1 The
student should now repeat for himself
the Similar argument for the case in
which Px is less than twice Py (say
px~Pt) It now follows m general that,
under competitive conditions, the equi-
X librium pnce ratio will conform to the
opportunity cost ratio
Thus, as soon as wc know the tech
meal conditions of production, then we
also know the relative pnees
'
Wc do
not, however, know the absolute price
iso -wt i^iiaVi ifraA
be anything we wish and later enquire
^
Vv
mto the consequences of changes in the
absolute level of prices Let us say that
1 The reason why pnce depends only on costs
(i e supply) and not on both demand and supply
as It did earlier is that costs are assumed to be
constant so that supply curves will be perfectly
elastic
 BALANCE OF COMPARATIVE ADVANTAGE 743

the price of X in country A is S2 while Now, in these

circumstances, B’s X
inB it is ,^1. Now it follows from what which sells for £\
domestically can be
we have just said that we know the sold for SI in A. Also B’s Y can be sold
price of jT in A and B. These prices are forSli in A. On the other hand A’s X
listed in Table 52A.2. must sell in B for £2 and A’s 7 must
sell in B for £'i 4r. In other words, at the
Table 52A.2 exchange rate Sl=£^l, B can under-
MONEY PRICES OF X AND Y sell A in both X
and 7; B will thus
IN COUNTRIES A AND B export X and 7 to A but will buy
nothing in return. This means that on
X Y
the foreign exchange market everyone
Country A S2 S3.20
will be wishing to buy pounds in order
Country B £\\ to purchase B’s goods which are cheaper
than A’s, but no one will ivdsh to buy
A comparison of Tables 35.1 and 35.2 dollars in order to purchase A’s goods
wU show that in each country the which are more expensive than B’s.
price ratios and the opportunity cost
This means that everyone will want
ratios are the same. (We have now pounds but no one
to trade dollars for
given quite a long chain of reasoning.
will ivish to trade pounds for dollars.
Although each link may be simple Clearly this is not a stable situation:
enough the whole chain may seem the poundis undervalued and the
difficult at the first attempt. The stu-
dollar overvalued. The price of the
is
dent who has any difficulty in under- pound in terms of dollars will rise. Let
standing the relationship between
us say that the price of pounds rises
Tables 52A.1 and 52A.2 should go back until the exchange rate becomes £\ =
and re-read this section from the
S2, the resulting prices are listed in
beginning.)
Table 52A.3.

THE DETERMINATION OF THE Table 52A.3

EXCHANGE RATE^ MONEY PRICES OF GOODS
EXPORTED FROM ONE
We now know the prices that will exist COUNTRY TO THE OTHER
in countries A and B when they do not WHEN THE RATE OF
engage in trade. Now let us assume EXCHANGE IS = S2 ;,(;i

that trade opened between these two

is

countries and that the cost of transport-

X 7
,

ing X and Y between them is small Sale price in A of

the commodity im-
enough so that it can be ignored.
ported from B S2.00 S2.66
Before we know which way trade will
Sale price in B of
flow we must know the rate of exchange
the commodity im-
between the currencies of the two
countries. Let us start by saying that
ported from A £\ \2s.

one pound exchanges for one dollar. Now' it is a matter of indifference

1 If you are uncertain what an exchange rate whether X is brought from A or B.
isyou should read the first two pages of Chapter Clearly this is the cheapest possible rate
"
53 at this time. of exchange for the pound. As long
744 THE INTERNATIONAL ECONOMY
the pound cxchtnges for anythmg
than S2 then both A and }' will be
cheaper when bought from B than from
A, trade t'lll Boo onI> from R to A,
ever) one will wish to exchange dollars
for pounds but no one will wish to do
the opposite, and so the price will
change At the exchange rate £\ =
S2, B will sell } to A but A could sell A
to B As soon as the exchange rate rises
abo\e;{, 1 =S2 then \ s A will definitely
be cheaper than B*s, while B‘s } will
still be cheaper than A's The student
should himself work out whit the
prices will be if the exchange rate is
£1 = S2 25
Now assume that the sahie of the
pound goes as high as £\ =S2 10 ixt
us see what the prices of imported
goods would be at this exchange rate
Country B's A would sell in A for $2 10
while her 1 would sell for S3 20(=jf
at S2 40 to the pound) Country A's A
ivould sell for = S2 at S2 40 to the
( must he
pound) while her } would sell for C\\
=S3 20 at $2 40 to the pound) Nosv,
at this exchange rate, A will sell \ to R
while B could just sell } to A At an>
exchange rate higher than ifl =S2 40,
A will be able to undersell B in both A'
and I all trade would flow from A to
, will sell A' to
B, eseryone would want dollars and no
one would want pounds, and the ex-
change rate would have to change.
There are a number of things to
notice about this example It was noser
possible to pick an exchange rate at
w'nic'ii A could sc'fl '} to'B w'liile at the
same time B could sell \ to A At all
exchange rates at which trade flowed in
both directions A sold A’ to B and B sold
I to A Trade was always in the direc-
tion of comparatue advantage 4Vcsaw
in the last chapter that A had a com-
parative advantage m
the production
of A and that B had a comparatnc
less advantage in the production of Y, ite
hare now seen that on out assumption the
pme system ensures that A produces and
exports Xuhtle B produces and exports }
\Vc should also note, when construct-
ing other examples, that the limiting
exchange rates arc easily calculated
Tiiid tlic rate at v»hich the commodity
in whicli A has a comparative ad-
vantage will have (he same pnee as that
commodity has when produced in B,
then find tlie exchange rate at which
the commodity in which B has a com-
parative advantage will have the same
pnee ns that commodity has when it is
produced in A These two exchange
rates will be the limiting rates between
winch the actual rate must he If the
rate should be outside these limits one
country will be able to undersell the
other in both rommodities and the ex-
I change rales will have to change In
the present example the exchange rate
heivvt en the limiu/^I = S2 and
;C1=S2 10
We shall
novv consider some changes
in our basic example and sec what
efTecis they would have We start by
assuming the et^uihbnum exchange
rate to be 4l = S2 20 in which case A
B for while B will sell
r to A for S2 93 i
THE EFFECTS OF A DOMESTIC
INFLATION m ONE COUNTBY
.\5vumc that there is an inflation in
country B w'hitc Oic price'lcvcl remains
constant in A Let us say, for example,
that all money pnees in B double Now
A' will cost ^
and 1’ ;£23 when pro
duced in B This means lliat B's export
good, }, will now have to be sold in A
for S5 86^ in order that the domestic
producers can obtain ^25 Country
B’s Twill be priced out oi the market,
 BALANCE OF COMPARATIVE ADVANTAGE
being unable to compete with Y pro-
duced in A. Thus the inflation in-
creases the price of B’s goods until she
is no longer able to export at all. Now
evei^’one will wish to purchase dollars
and no one will want pounds. The
value of the pound will fall. Consider,
before reading on, what would happen
if the e.xchange rate fell to ^(^1 = 51.10.
This halving of the value of the pound
will halve B’s export prices and, since
they were doubled by the inflation, they
will now return to their original pre-
inflation level. On the other hand,
imports from A will have to be doubled
in price in order that A’s producers
receive the same number of dollars per
unit as originally. Since the inflation
doubled B’s domestic prices and the de-
valuation doubles her import prices,
the relative attractiveness of imports
and domestically produced goods will
be unchanged. We now draw a most
important conclusion:
An inflation of X
per cent in one
country (with no corresponding
change in other countries) com-
bined with an X per cent deprecia-
tion of that country’s exchange
rate leaves the relative price of all
goods unchanged, both at home
and abroad.
This a most important proposition
is
which helps to relate the equilibrium
exchange rate to the price level. A
change in the price level in one country'
affects the competitiveness of that
country’s products. If the price level
rises then the country’s exports rise in
price while imports become relatively
cheaper because their price will be un-
changed while domestic prices will have
risen. These price movements will tend
to encourage imports and discourage
24*
745
exports. If the exchange rate of the
countr)'’s currency is then depreciated
by the same proportion as her price
level has risen, the original pre-inflation
competitive situation will be re-estab-
lished. The prices of all export goods
will fall to their original levels in terms
of foreign currency the prices of all im-
;
ported goods ivill rise in terms of
domestic currency in exactly the same
proportion as did the prices of all
domestically produced commodities.
THE EFFECT OF A FALL IN
LABOUR COSTS IN ONE
COUNTRY
Now let us return to the original situa-
tion inwhich the prices of domestically
produced goods are those shown in
Table 52A.2 while the exchange rate is
£\ =S2.20. Country A will be export-
ing Xto B while B will be exporting Y
to A. Now assume that country A
suddenly becomes a ‘cheap labour
countr)'’. Let us assume that the prices
of all factors of production are halved
in country A and that, as a result, the
prices of all goods are halved in that
country. The new set of prices is given
inTable 52A.4.
At the original exchange rate oi£l =
S2.20 A will be able to sell both X and
7 at a lower price than can B. (If you
convert B’s prices into dollars at the
going exchange rate both will be higher
than A’s, or, if you convert A’s prices
into pounds, both will be lower than
B’s.) In fact this change affects the
relative competitiveness of A and B in
just the same wayas did B’s inflation
the inflation period priced B out of the
market beeause it doubled B’s prices,
this change prices B out of the market
because it halves A’s prices; as far as
relative competitiveness is concerned.
746 THE INTERMATIOHAE ECONOMY
doubling B’s paces with A’s paces con- devaluation also halves the dollar price
stant JS exactly the same as halving A’s of imports from B Thus the relative
paces with B’s prices constant. As a attractiveness of goods produced in A
consequence of this change everyone and goods imported from B is left as it
will try to buy dollan in order to obtain was onginaliy.
the cheap A goods and no one will wish We have now derived some predic-
to buy pounds Thus the pound will fall tions of great importance
m value against the dollar Let us say
that the rate falls from its old level of
Neither the absolute level of prices
;{;i = S2 20 to the level of;(;i = Sl 10 nor that of factor costs (wages,
If this happens the original eompeUlire stlua- interest, etc.) is important in
tion iLill be exactly restored The fall in determining the direction of trade.
The equilibrium exchange rate will
Table 52A 4 always adjust so that trade flows
PRICES OF GOODS in both directions whatever the
PRODUCED DOMESTICALLY level of money prices and costs.
X Y Which commodities are imported
Country A SI SI 60 and which exported will depend on
Country B ^([1 relative prices and relative costs
only. Under these conditions trade
domestic prices m
A plus the fall the m will always follow the pattern set
value of the pound leaves the pound
pace of A’s exports at their oagmal
by comparative advantage. The
free price system thus causes
level Thus as far as D is concerned all
import prices are back where they were countries to specialise in the pro-
originally On the other hand, the duction of commodities in which
dollar prices of goods produced in A they have a comparative advan-
are halved by the cut in costs, but the tage.
 ;

CHAPTER 53

EXCHANGE RATES

We begin this chapter with a word of warning. In common speech, and

indeed in what follows here, convenient to speak of nations as trading
it is

various commodities. This convenient anthropomorphic form of expression

should not mislead the student into thinking that all, or even the majority
of, decisions about trade are actually made by governments. In most coun-
tries, governments do play some role in foreign trade; but, in market
economies, most of the decisions are made by individual households and
firms. Firms may think they see an opportunity of selling goods abroad, and
they may arrange to have these goods exported other firms
;
may think they
see an opportunity of selling foreign goods in the home market and arrange
to have these goods imported. If households find the goods attractive and
purchase them, the venture will be successful; if they do not, then the goods
will remain unsold and will no longer be imported. Governments may, of
course, try to influence this process; they may put subsidies on exports,
seeking to encourage foreign sales of domestically produced goods by making
their prices more attractive; they may put on imports, seeking to
tarifis

discourage domestic sales of foreign-produced goods by making their prices

less attractive. But in free-market economies, foreign trade, just like domestic
trade, occurs as a result of independent decisions made by firms and house-
holds, decisions that are coordinated — more or less effectively — by the
price system.
The major complication in foreign trade is that different countries use
different currencies. The currency of one country, while generally accept-
able within the bounds of that one country, will not be acceptable to the
firmsand households of another country. If, for example, an importer in
India wishes to purchase British goods, he cannot pay for them in rupees
he has to obtain sterling first. In general, trade between nations can occur only
if it is possible to exchange the currency that of another. We must
of one nation for
now ask: what is the mechanism for exchanging currencies and what are
the consequences of the need to do so ?
748 THE INTERNATIONAL ECONOMY

WHAT IS AN EXCHANGE RATE?

The exchange rate between two currencies is nothing more than the rate
at which these two currencies exchange for each other, it is thus the price
of one currency in terms of another If, for example, the exchange rate

between Britishpounds sterling and US dollars is ;^I = S2 82, then one

pound will exchange for two dollars and eighty-two cents or one dollar will
exchange for 7j \d Thus, if a holder of sterling gives up ;^I, he will receive
$2 82 in exchange, whereas if a holder of dollars gives up SI, he will get
Is \d in return
Exchange rates arise because the (paper) currency of one country is not
an acceptable medium of exchange in another country They arise because
It IS necessary in the course of international trade, to trade the currency of
one country for that of another If a UK
manufacturer sells goods abroad,
he wishes to receive sterling in exchange rupees, yen, or dollars arc of
little use to him, for he cannot pay hts workers, or purchase matcnals, or
pay dividends to his shareholders in these currencies, for all of these pur-
poses the only acceptable currency is sterling If a manufacturer sells goods
if he knows that
abroad, he will accept payment in other currencies only
he can exchange them for sterling
The American importer of British goods requires pounds in order to
purchase these goods from the British manufacturer When he buys these
pounds he will offer dollars in exchange for them Thus the American
importer is a dmander of pounds and a supplier of dollars The Bntish im-
porter, on the other hand, requires dollars in order to purchase goods from
the American manufacturer When he buys dollars he offers pounds in
exchange He is, therefore, a demander of dollars and a supplier of pounds
This gives us all the conditions that wc need for the determination of a price
Some people are trying to trade pounds for dollars, and others arc trying to
trade dollars for pounds

THE BALANCE OF PAYMENTS

The Balance of Desired Payments
If, at the current rate of exchange, the demand for dollars exceeds the
supply, It follows that holders of sterling arc trying to make more payments
m dollars than holders of dollars wish to make m
sterling In other words,
desired payments between the two countries are not in balance If the total
volume of payments that holders of sterling wish to make to America is
equal to the total volume of payments that holders of dollars wish to make
to Britain, the demand for dollars will equal the supplv, and the demand
for sterling will equal its supply Desired payments between the two
 EXCHANGE RATES 749

countries will be equal. Finally, if the amount that holders of dollars wish to
pay to Britain exceeds the amount that holders of sterling wish to pay to
America, then the demand for sterling will exceed the supply, and the
demand for dollars mil be less than the supply. Again, desired payments
between the two countries will not be in balance. Thus, to say that there is
disequilibrium in the foreign-exchange market (i.e., demand for foreign
currency does not equal supply) is the same thing as saying that the desired
payments between the two countries are not equal.

The Balance of Actual Payments

In order to know what is happening to the course of international trade,
governments keep track of the actual payments between countries. The
record of such payments is called the balance of payments. Although it is quite
want to purchase more dollars in exchange
possible for holders of sterling to
for pounds than holders of dollars want to sell in exchange for pounds, it is
not possible for sterling holders actually to buy more dollars than someone
sells. Every dollar that is bought must be sold by someone, and every dollar

that is sold must be bought by someone. Since the dollars actually bought
must be equal to the dollars actually sold, the payments actually made be-
tween countries must be in balance, even though desired payments may
not be.
The which payments are
balance-of-payments accounts record the reasons for
made. Thus, we can tell what volume of payments was made by foreigners
to British citizens for the purchase of British goods, the use of British
services — shipping, insurance, etc. - the lending of money to British house-
holds, firms or governments, or the investment of money in the ownership
of firms in Britain. The accounts also tell what volume of payments was
made by Britons to foreigners for the purchase of foreign goods, the use of
foreign services, the lending of money to foreign households, firms, or
governments, or the investment of money in the ownership of firms located
abroad. Any item that gives rise to a purchase of foreign currency is re-

corded as a debit item on the accounts, and any item that gives rise to a sale
of foreign currency is recorded as a credit item.
Although the total number of pounds bought on the foreign-exchange
market must equal the total number sold, this is not true if we look at
purchases and sales for a particular purpose. It is quite possible, for
example, that a larger number of pounds was sold for the purpose of
obtaining foreign currency to import, say, foreign cars than was bought for
the purpose of buyingUK cars for export to other countries. In such a case,
we would say that the UK had a balance-of-payments deficit on the car
account’, by which we would mean that the value of the UK imports of
750 THE INTERNATIONAL ECONOMY

cars exceeded the \ alueof its exports of cars, or, in other words, the number
of pounds sold because of car imports exceeded the number of pounds
bought because of car exports For most general purposes, we are not
interested in the balance of payments for single commodities but only for
larger classes of transactions
The most important division in the balancc-of-payments accounts is
between current account and capital account The balance of payments on
current account includes all payments made because of current purchases
of goods and services There is no automatic reason why current-account
payments should balance It is quite possible for more pounds to be sold in
order to purchase our imports than were bought m order to purchase our
exports If so, the pounds must have come from somewhere, and the excess
of sales over purchases on current account must be exactly matched by an
excess of purchases over sales on the capital account
The capital account records transactions for everything other than what
is recorded in the current account The main items are capital transfers and
sales from or purchases of stocks of gold and foreign exchange When a UK
citizen wishes to invest abroad, he must obtain the currency of the relevant
foreign country, he must sell pounds and purchase foreign currency This
IS recorded as a deficit item in the balance of payments, since the transaction
uses pounds
We may think of the balance-of-payments accounts in a variety of ways
We have alread> seen that
Sales of foreign currency and Purchases of foreign currency and
purchases of domestic currency s sales of domestic currency
If we look instead at the transactions that lead to these exchanges of
currency, we obtain
The value of all exports of goods and The value ofall imports of goods and
services plus all capital imports s services plus all capital exports

If we now divide these transactions into two major divisions, we obtain

Credits Debits
CuTTent account
A Exports of goods and services C Imports of goods and services

Capital account
B Imports of capital D Elxports of capital

The fact that A+B is necessarily equal to C+D shows that a deficit on one
account must be matched by an exactly opposite surplus on the other
account
Now assume that in a given year the value of UK imports exceeds the
 EXCHANGE RATES 751

value of UK exports, considering all current-account transactions. The

foreign currency necessary to finance the imports that were in excess of
exports had to come from somewhere, and, clearly, it could not come from
people selfing foreign currency for pounds sterling in order to buy UK
goods and services. The money must have been lent by someone or else
provided out of the government’s reserves of gold and foreign exchange. If
foreigners are investing funds in the UK, they will be selling foreign
currency and buying sterling in order to be able to buy stocks and bonds
issued UK firms. Such foreign lending can provide the foreign exchange
by
necessary to allow the UK to have an excess of imports over exports. The
other possibility that the UK central authorities have reduced their
is

holding of foreign currency or gold by selling some to persons wishing to

purchase foreign goods and accepting dollars in return.

A deficit on current account must be matched by a

surplus on capital account, which means either
borrowing from abroad or reducing the foreign
exchange and gold held by the domestic central
authorities.

What about a surplus on current account? This means that the value of
exports exceeds the value of imports. This means that foreigners will not
have been able to obtain all the sterling they needed in order to buy
UK goods by making purchases from persons who were eager to supply
sterling in return for foreign currency in order to buy foreign goods. The
excess of exports over imports could only have been paid for if foreigners
obtained sterling from other sources. Again there are two main possibilities
sterling may be provided by UK
investors wishing to obtain foreign
currency so that they can buy foreign stocks and bonds. In this case the
excess of exports over imports is balanced by UK
loans abroad. The other
possibility is that foreign governments reduced their holdings of
may have
sterling by selling them to persons wishing to buy British goods and accept-
ing their own domestic currency in exchange.

A surplus on current account must be matched by a

deficiton capital account, which means either lending
abroad or running down the reserves of gold and
foreign exchange held by the foreign central authori-
ties.

The make-up of the current account: The current account is usually

subdivided into the trade in visibles and invisibles. Visibles refer to goods,

i.e., to all those things, such as cars, wood pulp, aluminium, coffee and
iron ore, that we can see and touch when they cross international borders.
752 THE INTERNATIONAL ECONOMY

Invisibles refer to services, to all those things we cannot see or touch, such
as insurance and freight When a US firm insures a shipment of
haulage
goods consigned to Australia with Lloyds of London, the firm consumes a
British export just as surely as if it purchased and used a British-made car
or sent its president on holiday to Scotland Payment for the insurance
services and for the car and the holiday must be made m
pounds - and thus
each IS a US import and a British export

The make up of the capital account The capital account records all

foreign capital movements If UK

investors invest abroad, they must supply
sterling and demand foreign currency A Bntish investment abroad thus
contributes toward a foreign-exchange deficit because it uses foreign cur-
rency An investment in the UK
from a foreign source contributes to a
surplus on capita! account
Capital movements can be divided into long-term capital movements,
short term mov ements, and changes in the exchange reserves Long term
capital movements represent genuine international investment Allowing
for nsk and other such factors, investors will seek to invest where the return
IS highest Just as capital moves from industry to industry within one
country in search of its most productive uses, so capital moves from country
to country m search of the highest rates of return Such capital movements
mean that the households and firms of one country arc investing m the
industry of another country, and they show up as long term capital mo\e-
ments in the payment accounts
Short term capital holdings anse in many ways The mere fact of inter-
national trade forces traders to hold money balances Traders’ receipts and
expenditures are not perfcctlv synchronised, and they necessanly hold
‘
balances because they must be able to pay their bills when they fall due
It usually does not matter w here in the world such funds are held The funds

can thus be moved from one country to another m response to small changes
in incentives or because of real or imaginary fcan of all sorts The large
quantity of these funds is a potential source of international instability, be-
cause a sudden rush of short term capital out of one currency into another
can cause violent shifts m demands and supplies for foreign currency
The hna! ei’cment account is efianges in goi’rf anJ tfrmgn-
in the capital
exchange reserves held by central authonties Central authorities of most
countnes hold supplies of gold and foreign exchange in order that they may
intervene m
the foreign exchange market for a wide variety of purposes
Gold (being universally acceptable at a fixed price) is immediately transfer-
able into any foreign currency that may be required If a country has a
payments deficit on all other counts (i e , if it uses more foreign currency
1 These are the unavoidable balance* already discussed on page 695
 EXCHANGE RATES 753

than it must be made up by an equivalent reduction

obtains), this deficit
in its and foreign exchange.
reserves of gold
When we add up all the uses to which foreign currency is put and all
the sources from which it came, these two amounts are necessarily equal,
since the foreign currency that goes to any use must have come from some-
where. The over-all accounts on all international payments are thus neces-
sarily in balance. When we speak of a balance-of-payments deficit, we
mean on some part of the accounts. Usually the term refers to
the balance
the balance excluding changes in the reserves held by the central authorities.
A balance-of-payments deficit thus means that the reserves of the central
authorities were being run down by exactly the amount of the deficit.

A THEORY OF EXCHANGE RATES

We have seen that the exchange rate is the price at which one currency
trades for another. We must now ask how this price is set and what are the
factors that cause it to change.
The theory we are about to develop applies to all international trade
and to all exchange rates, but for the sake of expositional simplicity we
shall refer to two countries, Britain and America, and to the rate be-
tween their two currencies, pounds and dollars. Since one currency is
traded for another on the foreign exchange market, it follows that a demand
for dollars implies a supply of pounds while an ofier (supply) of dollars
implies a need (demand) for pounds. If, at an exchange rate of ^^1 =$3, a
British importer demands S6 he must be offering -£ 2 and if an American ,

importer offers $6 he must be demanding 2 For this reason we can deal

£ .

either with the demand for and the supply of dollars, or with the demand
for and the supply of pounds sterling; we do not need to consider both. We
shall conduct the argument in terms of dollars.

The demand for dollars: The demand for dollars arises because holders
of sterling wish to make payments in dollars; the demand for dollars thus

arisesfrom imports of American goods into Britain and from the movement
of investment funds from Britain to the US.

The supply of dollars: Dollars are offered in exchange for sterling

because holders of dollars wish to make payments in sterling. The supply of
dollars arises therefore on account of British exports to the United States
and the movement of investment funds from the United States to Britain.

Price changes caused by exchange-rate changes: A British manu-

facturer is concerned to receive a certain payment for his goods in
pounds
 EXCHANGE RATES 755

price of British exports. "We can therefore restate our previous conclusion
in the following manner:

A depreciation of the pound will increase or decrease

the dollar value of British exports, and hence the
supply of dollars, according as the American demand
for these exports is elastic or inelastic.

Now let us consider the demand for dollars. The demand for dollars in
exchange for pounds arises on account of the need pay
to obtain dollars to
forAmerican goods imported into Britain. What can ive say about ivhat
happens to the volume of this demand for dollars as the pound price of
American goods varies ? Let us assume that the dollar price of these Ameri-
can goods remains unchanged but that their sterling price changes in
consequence of a change in the exchange rate. (For example, an American
good ivith a constant price of SI w'ill have a sterling price of 65 Qd w'hen the
exchange rate is S3=^l, a price of lOsif the exchange rate falls to S2—£l,
and a price of 5s if the exchange rate rises to S4 = ;^l.) Now, if the pound
price of American goods falls in consequence of a fall in the exchange value
of the dollar, more wall be bought and British purchasers will spend more
or less sterling on these goods according as their elasticity of demand is

greater or less than unity; but, since the quantity of goods bought rises
and their dollar priceunchanged, they will necessarily spend more dollars on
is

them. On the other hand, if the sterling price of American goods rises as a
consequence of a fall in the value of the pound less goods wall be bought
and, although the sterling expenditure on them will go up or down depend-
ing on the elasticity of demand, the dollar expenditure on these goods
necessarily falls since we are assuming that dollar prices are unaffected by the
devaluation.* We have now reached the followdng important conclusion:
A devaluation of the poxmd will necessarily decrease
the dollar value of British imports and hence de-
crease the demand for dollars ; the more elastic is the
British demand for American goods, the more will the
demand for dollars fall with any given devaluation.
The argument leading up not an easy one;
to the last three conclusions is

the student should now go back and re-read this section and, when he gets
to this point a second time he should put the book dowoi and make sure that

he can work out the argument for himself.

1 This is only a very elementary theory and we thus use the simplifying assumption that
products.
exports can be increased or decreased svithout affecting the domestic prices of these
The theory' which allows for variations in domestic prices is much more coraple.K than the
present one.
756 THE INTERNATIONAL ECONOMY

We ma) now introduce a diagram that will be useful for analysing

exchange rate problems in terms of the theory just spelled out In Figure
53 1 we measure the sterling pnee of a dollar on the vertical axis (The
student will be more familiar with the exchange rate expressed the other
way round, as the dollar price of one pound sterling and he should, for
practice, convert a few of the pnccs in Figure 53 1 into the more familiar
mode of expression W hat, for example, is the dollar price of one pound if
the sterling price of one dollar is 6j^)
The demand cur\ e for dollars is always downward sloping This indicates

Ftg 53 1 The determination of the equilibrium exchange rate

under competitive conditions

that as the sterling price of dollars falls, a larger quantit> of US goods will
be sold in Britain and, assuming the dollar price of US goods to be un-
changed, a larger quantity of dollars will be needed to purchase these goods
(When considering the demand curve remember that the figures on the
vertical axis tell us the number of shillings for which an American good
costing SI must be sold m the UK at each exchange rate ) The elasticity
of the supply curve of dollars depends on the elasticity oF the US demand
for British goods If the US demand is of unit elasticity then the same
number of dollars will be spent on British goods, irrespective of price In
this case the supply of dollars will be the same whateser the exchange rate
and the S curve in Figure 53 1 will be of zero elasticity If the US demand
for Bntish goods is elastic, the supply curve of dollars will slope upwards to
the right (e g , in Figure 53 1), while if the US demand for British goods
IS inelastic, the supply curve of dollars will slope upwards to the hft (e g ,
5”; m
 EXCHANGE RATES 757

the figure). Before reading on, the student should be sure that he can ex-
why this is
plain so.

We may now one or two problems in terms of this

briefly consider
analysis; the student should himself stand ready to apply the analysis to
other problems. We will now explicitly assume that the US demand for
British goods has an elasticity greater than unity. The demand and supply
curv^es will thus have the shape shown in Figure 53.2. Later we shall drop

this assumption.

Fig 53.2 The effects of shifts in the demand for dollars on the equilibrium
exchange rate.

The determination of the e<2,uilibrium exchange rate; Assume

that the current price of dollars too low, say 5s to the dollar. At this
is

exchange rate the demand for dollars exceeds the supply. In other words
desired payments are not in balance, for desired payments to the US
exceed desired payments from the US to Britain. Dollars will be in scarce
supply; some people who require dollars to make payments to America
758 THE INTERNATIONAL ECONOMY

Will be umbic to obtainthem and the pncc of dollars will be bid up The
\ilue of the dollar vis ^ the pound will appreciate or, what H the same
its

thing the value of the pouml vu itts Uic dollar will dcpccciale j\s the
price of dollars rises the sterling pncc of Amencan exports to Britain rises
and the demand for US dollars to bu) these poods falls off On the other
hand as the dollar price of Rniwli exports falls a larger quantity will be
sold and on the assumption that the US demand for Bntish goods is elastic,
the supply of US dollan will nsc Thus this rise m the price of the dollar
reduces the quantity demanded and increases tlie quantity supplied W here
the two curves intersect demand equals supply and the exchange rate is m
equilibrium In Figure 53 2 the erjuihbnum exchange rate is at S = 7/ \d 1

which IS to the nearest cent £1 = 5262 Now let us see what happens if the
pnee of dollars is too high In this case the demand for dollars will fill short

of the supply the dollar will be m

excess supply so that some people who
wish to convert dollars into pounds will be unable to do so Tlie pnee of
dollars will fall fewer dollan will be supplied more wall be demanded and
an equilibrium will be rc established

A CHANOF IN TASTES I ct US comidcf for cxamplc the cfTect of a change

wherebt the Brtiish prefertnee for American go^s increases so that there
ISa rightward sliift in the British demand curve for Amencan goods This
means that at each sterling price more US goods will be demanded than
previously Tims at each excliangc rate more US dollars vmII be demanded

(in order to pay for tliesc goods) Tlius the demand curve for dollan wall

shift to the right say to D m Figure 53 2 At the original exchange rate of

7r per dollar demand exceeds supply Tlie pnee of dollars will nse until
a new equilibrium is readied in Figure 53 2 this new equilibnum is at 8f
to the dollar (i c £1 = S2 50) We thus conclude that an increased prefer
ence on the part of British consumen Ibr Amencan goods will lead to a n<c
in the equilibrium value of the dollar (a fall in the value of the pound) The

student should work out for himself the effect of an increased preference of
American consumers for British goods

A FALL IN THE DOMESTIC SUPPLY PRICE OF EXPORTS AsSUmC forCXamplC

that the domestic price of US goods falls This means that at any given
exchange rate the sterling pncc of US goods will fall W hat will this do to
the exchange rate"^ Can you work out the answer before reading on If the
dollar pnee of US goods falls by A per cent then at each exchange rate the
sterling pncc of these goods will also fall by \ per cent If the Bntish demand
for US goods IS elastic the quantity demanded wilJ increase by more than
X per cent and if it is inelastic, the quantity demanded will increase by less
than \ per cent Thus the British will demand more or Jess dollars in order
lo purchase US goods according as their demand for these goods is elastic
 EXCHANGE RATES 759

or inelastic. Thus a fall in the dollar price of US exports will shift the de-
mand for dollars to the right if the British demand for US goods is elastic,
to the left if inelastic and leave the demand for dollars unchanged if the
demand for the goods is of unit elasticity. We conclude therefore that the
sterling price of dollars will rise, fall or remain the same according as the
UK demand for US goods is elastic, inelastic or of unit elasticity.

A CHANGE IN THE PRICE LEVEL OF ONE COUNTRY: We may Consider, by

way of example, the case of an inflation in the UK. This means that the
sterling price of British goods will rise. British goods will become more
expensive in the US and, assuming the US elasticity of demand for British
goods to exceed unity, the supply of dollars will diminish. American exports
to Britain willhave an unchanged sterling price, while the price of British
goods sold at home will have increased. Thus US goods will be more
attractive compared to British goods (because they have become relatively
cheaper) and more of them will be bought in Britain. Thus, at any given
exchange rate, the demand for dollars will be increased. The demand cun^e
for dollars shifts to the right ^vhile the supply curve shifts to the left so that
the equilibrium price of dollars must rise.' We conclude therefore that a
British infladon leads to a depreciation in the equilibrium value of the
pound (an appreciation in the value of the dollar).

An EQ,UAL PERCENTAGE CHANGE IN THE PRICE LEVEL IN BOTH COUNTRIES:

Let US consider by way of example a 10 per cent inflation in both the US
and the UK. In this case, the sterling prices of British goods and the dollar
prices of US goods both rise At any given exchange rate,
by 10 per cent.
therefore, the dollar prices of British goods and the sterling price of Ameri-
can goods will also rise by 10 per cent. Thus the relative prices of imports
and domestically produced goods will be unchanged in both countries.
There is now no reason to expect any change in either countr)'’s demand for
imports at the original exchange rate, so that the inflations in the two
countries leave the equilibrium exchange rate unchanged. This argument
can easily be extended to establish the following important conclusion

If the price levelof one country is rising faster (falling

slower) than is that of another country, then the
equilibrium value of its currency will be falling
relative to that of the second cormtry.

shifts to
1 If the US demand for British goods is inelastic then the supply curve of dollars
the right. It can be shown, however, that the supply curve shift in this case must always be
less

not
than the demand curve shift so that the conclusion about the equilibrium exchange rate is
introduction to the theory of
upset. The demonstration is too difficult to be included in this
exchange rates.
760 THE INTERNATIONAL ECONOMY

Fixed and fluctuating exchange rates In the analysis so far we

have been concerned with the effect of various changes on the equilibrium
rale of exchange If exchange rales arc left free to be determined by the
forces ofdemand and supply, this analysis provides predictions about what
happens to actual exchange rates This would have been the case, for ex-
ample, with the rate of exchange between the Canadian and American
dollars because the value of the Canadian dollar was allowed to fluctuate'
on the free market according to the demand for and the supply of such
dollars Such an exchange rate is called a fluctuating exchange rate
The rate of exchange between the British pound and the US dollar, on the
other hand, is fixed, within very narrow limits, by government decree In
this case fluctuations in the demand for and supply of dollars vts-i-vts pounds
cannot affect, except within very narrow limits, the actual exchange rate,

but they do change the magnitude of the excess demand for, or supply of,

dollars Let us consider this mailer further

Assume, for example, that the exchange rate m Figure 53 2 is fixed by
decree at 5s per dollar (it £ls*S4) At this rate of exchange demand for

dollars is Oa while the supply is an excess demand for dollars,

Ob There is

a potential ‘dollar gap’ of ba dollars If the government does not intervene,

some demanders will be able loobuin dollan and others will not, a black
market is likely to develop with the available dollars commanding a clan-
destine price far in excess of 5j per dollar In order to prevent these con-
sequences, the government may
and ration the available supply of
step in
dollars All persons wishing to convert dollars intopounds are forced to do
so through the government authority, and the available supply of dollars is
then rationed out according to pnonties established by the government
Assume, for example, that the exchange rate is fixed at 5r to the pound
in Figure 53 2 and that the demand for dollars is curtailed by rationing to
Ob dollars Now anything that shifts the demand curve DD will change the
equilibrium exchange rate and will change the potential dollar gap (le ,

desired purchases of dollars miniu actual supply of dollars), but will leave
the actual dollar earnings and actual dollar purchases unchanged A nse m
demand io D D will raise the potential dollar gap from ba to be, while a
in demand to D''D’
fall lowers it to bd, in both cases actual dollar earnings
and hence maximum dollar purchases remain unchanged at Ob Now
consider a supply curve shift Again the equilibrium exchange rate changes,
again the potential dollar gap changes, but now actual dollar earnings and
hence possible dollar expenditure changes Thus all of the previous analysis
of equilibrium exchange rates also provides predictions for the case of fixed
exchange rates
A long debate has raged amongst economists about the relative ments of
1Until May 1962
 EXCHANGE RATES 761

fixed versus fluctuating exchange rates. By and large, the supporters of fixed
rates believe that the stability of such rates is conducive to trade. The
advocates of fluctuating rates believe that these rates would not fluctuate
enough to upset trade, that speculative capital movements would reduce
and that the simplest way to maintain
short-term fluctuations in the rate,
international equilibrium allow relative prices of the exports of all
is to
countries to change as conditions of demand, supply and price change -
these relative price changes of internationally traded goods can be effected
by changes in the exchange rates. No general conclusion has been agreed,
and the whole debate has been more notable, with a few major exceptions.

for the passions involved than for the objectivity used, in assessing empirical
evidence. We cannot here go into the issues involved, except for the problem
considered below, but we might note in passing that, contrary' to the con-
tentions of some of the protagonists, neither free rior fluctuating rates seem
to provide the royal road to disaster, for both policies have been successfully
maintained at one time or another by several countries.

The possibility of unstable exchange rates; One of the issues

involved in the exchange rate controversy has been the problem of the
effects of devaluation on the balance of payments. Assume that the American

demand for British goods is supply curve of dollars will now be

inelastic; the
backward sloping as is S, in Figure 53.1. If the supply curve is actually
flatter than the demand curve then a perverse situation arises in which the
exchange rate is unstable. Such a case is shown in Figure 53.3. Assume, for
example, that the price of dollars is fixed at Is to the dollar, and at that
price there is a ‘dollar problem’ because the demand for dollars exceeds
762 THE INTERNATIONAL ECONOMY

supply Now, if the price of dollars is allowed to rise (the pound is depre
ciated), as will happen on a free market, the balance of payments problem
IS actually worsened The rise m the price of US exports to Britain does
cause a fall in the demand for dollars, but the fall in the price of British
exports to America causes an even greater reduction in the earnings of
dollars because the US demand for British goods is very inelastic Clearly,
if the world is like this, the obvious policy of devaluing the pound in order
to cure a persistent ‘dollar problem* would not be successful Instead the
pound should be appreciated in value’
How likely is this perverse case^ It can be shown that if the domestic
prices of export goods do not change as demand changes (supply curves of
exports have an infinite elasticity), the perverse case anses if the sum of the
US elasticity of demand for British goods plus the British elasticity of

demand for US goods is less than one If this is the case, then a devaluation
will worsen the balance of payments (i e increase the size of the gap be-
tween the demand for and supply of dollars) If the two elasticities sum
to more than unity then a devaluation will improve the balance of
payments
Clearly, it is important to know what the relevant elasticities are If one
believed that they were very low, one would not favour free exchange rates
nor would one advocate a policy of devaluation to improve the balance of
payments In this case, howe\ er, one might advocate a policy of inflation to
cure a balance of payments problem' The reader should now be able to
show for himself that the following two pieces of advice are self-contra-
dictory We should not try to cure our balance of payments problem
through devaluation because elasticities arc too low, we should control infla-
we shall price ourselves out of international
tion rigidly, for otherwise
markets This problem of the effect of devaluation on the balance of pay-
ments has led to many attempts to measure international trade elasticities
The resulting measures seemed to show lower elasticities than most econo
mists had expected In general, empirical studies of international demand
show rather small price effects For some lime the balance of evidence
seemed strongly on the side of the ‘elasticity pessimists’ who argued that
elasticities were too low for one to have real confidence in the policy of

devaluation as a cure for balance of payments deficits Later the statistical

theorists entered the controversy and general theoretical reasons were ad-
vanced to show why the measures that had been made tended to understate
The present state of the debate is rather unsatisfactory
the actual elasticities
The position of the extreme pessimists has been shaken, but completely
convincing evidence that elasticities are high enough to make a moderate
devaluation a practicable cure for balance of payments deficits has not yet
been advanced In fact the degree of uncertainty is such that different
 EXCHANGE RATES 763

economists exercising their own judgment about the facts still come to
widely differing conclusions.

EXCHANGE RATES AND NATIONAL PRIDE

It is an interesting sociological and psychological observation that the value

of exchange rates often becomes an important symbol of national pride.

The economist does not seek to explain this phenomenon but he can wonder
at There can be no doubt that there
it. are circumstances when the rise in
would be taken as a good sign. There are
the value of a countr)'’s currency
other circumstances in which such a change would be symptomatic of
domestic circumstances, about which it would be unusual to be proud. A
major domestic depression and deflation could, for example, easily lead to
a rise in the externa! value of a country’s currency. On the other hand,
major technical innovations which reduced domestic costs and prices might
lead to a fall in the exchange rate if foreign demands were inelastic. It is not
immediately obvious that this should cause a loss of national prestige. A
large inflow of foreign capital, leading to a transfer abroad of the control
of one’s industries, would cause an appreciation of the exchange rate. It is

not obvious that such a move is a cause for national congratulation in all

circumstances. It would by
certainly be a peculiar position (frequently held
Canadians during the 1950’s) to point with great pride to a high exchange
rate and to point with grave concern to the inflow of foreign capital which
was the cause of the high rate. In fact exchange rates can appreciate or
depreciate for many different reasons and to take the price of one’s currency
symbol of national pride is to commit oneself in advance
per se as a to being
proud of a great rag bag of varied events.
 I

APPENDIX TO CHAPTER 53

MORE ABOUT EXCHANGE RATES

In appendix we consider briefly
this and sells dollars At the price of 7x2<fto

three problems (1) exchange reserves the dollar, they offer to sell dollars,
and exchange rates under a fixed rate for permitted purposes, in unlimited
of exchange, (2) short term capital amounts, at the price of Is per dollar,
movements, and (3) Jong term capital they enter the market and buy dollan
movements m unlimited amounts If purchases
equal sales, the authority’s reserves of
dollars will be unchanged, if purchases
MANAGING FIXED EXCHANGE
RATES
Figure 53 Al shows frce-markcl de
mand and supply curves for dollan
Assume that the British Government
fixes the exchange rate between the
limits of Is and Is 2d to the dollar and
restricts demand, through an exchange
controlsystem of rationing and pro-
hibitions, to the curve DD Certain
goods will be allowed to be imported
without limit, others will be subject to
a quota restnction setting the maxi-
mum amount of imports, and other
goods will be prohibited from being
imported altogether By these means, Fig 53 A The stabilisation of the exchange
the central authorities keep the demand rate through the intervention of the central
for dollars lower than it otherwise authorities
would be Even this restricted demand
will, however, be' subject to seasonal, do not equal sales, these reserves will
cyclical and other fluctuations change
Having restricted demand by the I If the demand curve cuts the
means outlined above, the authorities supply curve in the range 7j to Is 2d,
then control the exchange rate the m then the authorities need not touch
following fashion all transactions go their exchange reserves The amount of
through the central agency, which buys dollars being supplied for pounds will
 ;

MORE ABOUT EXCHANGE RATES 765

be equal to the amount of dollars being level.In other words, they can devalue
demanded in exchange for pounds. the pound. Or, they can try to shift the
2 If the demand curve shifts to cuiA'es so that the intersection is in the

D"D", then the authorities must sell band 7s to 7s 2d. They can take further
dollars to the extent of rs in order to steps to restrict demand for dollars;
prevent the price of dollars from rising they can impose import quotas and
above 7s 2d. These dollars must be re- foreign-travel restrictions, or they can
moved from the exchange reserv'cs. increase the supply of dollars by en-
3 If the demand cur\'e shifts to couraging exports.
the central authorities must buy The actual balance-of-payments sur-
of tu
dollars to the extent and add them pluses, or deficits, reported in the press
exchange reserves in order to pre-
to its are the quantities we have just indi-
vent the price of dollars from falling cated: the diflference between current
below 7s. expenditure of dollars and current
If the authorities have restricted de- earnings of them, i.e., the changes in
mand sufficiently so that on the average the reserves held by' the central authori-
the demand and supply curves inter- ties. This must be sharply distinguished
sect in the range 7s to 7s 2d, then the from the potential balance-of-pay-
exchange reserves wU
be relatively ments surplus or deficit, which is the
stable with the authorities buying difference between the demand for and
when the demand is abnormally
dollars the supply of dollars as they would be if
high and selling them \s'hen the de- everyone were free to buy and sell
mand is abnormally low. In the former dollars without government restriction.
case, the reserv'es will rise, and in the
latter case, they will fall; but over a
long time their average level will be CAPITAL MOVEMENTS ON A
stable. FREE MARKET
If the central authorities have guessed
What about the exchange of currencies
wrongly, the exchange reserves will rise
for the purpose of making loans abroad
or fall more or less continuously. Say
rather than for the purpose of pur-
that the average level of demand is
chasing foreign goods ? Assume that in-
D"D" with fluctuations on either side
vestors inAmerica wish to loan money
of this level. Then the average drain on
in Britain, possibly to British firms. The
exchange reser\'es will be rs per period
British require pounds, and the
will
sometimes it will be more and some-
Americans will, therefore, have to pur-
times less, and occasionally when de-
chase pounds on the foreign-exchange
mand is extremely low, reserves will be
market.^ Such a transaction entails a
added to. This situation cannot continue
rise in the demand for pounds, ^vhich,
indefinitely. Eventually, if nothing is a free market, wall bid up their
in
done, reserves wiU zero and the price. A transfer of funds from country
fall to
controlled price have to be
will
A to country B will tend to appreciate
abandoned. The authorities have two
jB’s currency and depreciate A’s.
possibilities; they can change the con-
trolled price so that the band of 1 This is, what happens. The actual
in effect,
permissible prices straddles the inter- transfer will be accomplished through one of
section of the curves at their normal se\'eral different institutional channels.
766 THE INTERNATIONAL ECONOMY
be a demand on the part of these in-
Movements of Short-term Capital
vestors to turn pounds into dolian The
A temporary deficit in the balance of government can now sell tu dollars pet
payments may be alleviated by attract- penod to these investors
ing short-term capital into the country Thus short-term deficits m the
Assume that, in Figure 33 AI, the de- balance of payments can be covered by
mand for dollars is D^D’ so that the attracting short-term capital into a
British suffer a deficit in their balance of country, but this policy will only be
payments of rs per period Now assume successful ifan equivalent short term
that this high demand is only a surplus develops so that the capital can
temporary one, but that the central again be transferred out of the country
authorities wish neither to put on addi-
tional restrictions nor to let reserves run
down by rs What else can they do^ Movements of Long term Capital
They can raise the rate of interest they
arc willing to pay for short-term loans Long-term capital is capital lent for
(loans of a few days’ to a few months’ long penods of time, anything from a
duration) and attract short-term capi few to twenty or more years is common
tal There is a great deal of money Such capital is used by governments
available to traders time
for short and firms for long-range investment
periods These traders do not need the projects that normally increase, cither
money now, but they will require it directly or indirectly, the productive
shortly If someone is prepared to pay capacity of the economy Often people
for the use of this money for a short in wealthy countries where savings are
period, then this is better from their high will be prepared to invest funds m
point of view than leaving it idle ff the less well-developed countries where
central authorities raise the rate of savings are not high enough to finance
interest for short term loans, people a high rate of investment Money by
holding dollars will wish to obtain itself does not build or produce things,
pounds m
order to lend them out at the only productive resources can do this
high Bntish rates Thus the supply of When a saver lends his money to an
dollars shifts to the right If it cuts the entrepreneur in his own country, the
demand curve at 2 , then the inflow of saver is electing not to exercise his
short-term cash, rs, just covers the claim to the output of the country’s
on current account, and the ex-
deficit resources and is transferring this claim
change reser\es arc not run dosvn to the entrepreneur When the latter
Provided the government guessed conectly spends the money on investment, he is
thst fiVe «feiTranif was" adnarmaHj AigA, ifrrectrflg- the pmdeseCne senvees the
the policy will work If the demand country toward the manufacture of
now falls to an abnormally low level, capital goods and away from the manu-
say D“D", the government can buy tu facture of the goods they would have
dollars and add them to its exchange made if the saver had spent all his in-
reserves It can then lower short term come Thus, lending means transfer-
rates of interest so that people who have ring the claim to the productive services
lent money m
Britain would now prefer of the country from the lender’s use to
to lend It in, say. New York There will that of the borrower International lend
 MORE ABOUT EXCHANGE RATES 767

ing has the same effect. If a person in in this case, from B to A. In general,
country B lends money to an entre- lending from B to A entails a flow of
preneur in countr)' A, then the claim to goods from B to A (either directly, or
the output of country’ 5’s resources is via third countries, if we consider trade
transferred to a citizen of country' A. between many countries). The com-
All of the ensuing financial transactions are mon-sense reason is that borrowing
merely means of making this claim effective. represents nothing more than a transfer
Assume that an individual in B saves of claims to the output of productive
pounds and wishes to buy' the bonds resources from the lender to the
sold by a firm in A. Then, in effect, it borrower. The income of an individual
will be necessary' for the saver in B to in A gives him a claim on the output of
purchase dollars to purchase the bonds his own country, if he lends to an indivi-
in country A. This means that someone dual in country B, he is giving the
else must have sold dollars and taken borrow’er his claim on A’r output.
pounds in exchange. The simplest case It follows that international lending
arises when the firm in A wishes to buy and borrow’ing affects the demands for,
machinery from B. The lender in B and supplies of, currency in the foreign-
buys dollars and then uses them to buy exchange market. In terms of the
bonds in A. The firm in A then uses the analysis in Chapter 53, lending from
dollars to buy pounds and the pounds B to A entails a rise in the demand for
to buy machines. In the exchange dollars. This w'ill, among other things,
market, there is a rise in the demand for raise the price of dollars. We must con-
dollars arising from the saver in B and clude, therefore, that a transfer of
a rise in the demand for pounds (supply capital from one country' to another
of dollars) on the part of the firm in A. tends to appreciate the exchange value
Looking at the ‘real’ flows of produc- of the currency of the borrowing
tion rather than the money flows, we country and to depreciate the exchange
see that the lending from B to A entails value of the currency of the lending
a flow of B’s production, the machines country.
CHAPTER 54

TARIFFSAND THE GAINS

FROM TRADE

In Chapter 52, v'C showed how the classical theory of the gains from trade
points to specialisation according to comparauve advantage as the source
of the gams from trade, with the possibility of exploiting economies of scale
as an added source In this chapter, we first consider the theory of the
gams from trade as a positive hypothesis about the real world, and then go
oti to consider the case for interfering with free trade through tariffs
We have demonstrated that where opportunity costs differ betv^een
countries some degree of specialisation with some consequent amount of
trade will raise world standards of living There is abundant evidence to
show that such cost differences do occur and that potential gains from trade
do exist Today, no one seriously advocates complete self-sufficiency, but
some people do advocate increasing or diminishing the quantity of trade
that we now have This, as we shall see, is a more difficult issue to settle than
the issue of whether we should have any trade at all
It has sometimes been held that it is impossible for trade between any
two parties to be to the mutual advantage of both According to this view,
one trading partner must always reap his gam at the expense of the other
The principle of comparative costs, which shows that it is possible for both
parties to gam from trade, even if one of them is more efficient than the
other m of production, completely refutes the exploilation doctmt of
all lines

trade Seen in this light, comparative costs is to be viewed as a possibility

theorem It shows that, if opportunity-cost ratios differ in two countries,

specialisation and the accompanying trade make it possible to produce more
of all commodities, and thus make it possible for both parties to get more
goods as a result of trade than they could get in its absence Thus the answer
to the question Ts it possible for trade to be mutually advantageous is an

emphatic Yes, the answer to the question ‘Is trade m fact mutually
advantageous ’’ is quite another matter
 TARIFFS AND THE GAINS FROM TRADE 769

FREE TRADE VERSUS PROHIBITIVE TARIFFS

The theory of the gains from trade through the exploitation of differences
in comparative costs may be looked at not only as a possibility theorem,

but as a positive hypothesis about the real world.^ As a positive hypothesis

about the real world, the theory of comparative costs predicts that, in the
real world, there tvill be gains from trade in the sense of increased world

production, and that no country will lose from trade in the sense of having
less to consume than it could have if it were self-sufficient. This general
theors' has not been extensively tested, mainly because it has long been
believed to be self-evident. If asked to support it with empirical evidence,
most economists would probably point to the widely differing cost con-
most dramatic being those associated with
ditions in certain countries, the

climate. Itwould undoubtedly be possible, by using greenhouses, to grow

oranges, cotton and a whole host of imported raw materials and foodstuffs
in Norway or to grow coffee in the United States. But the cost in terms of

other commodities foregone would be prodigious, because these artificial

means of production require lavish inputs of factors of production. It would
likewise be possible for a tropical country' currently producing foodstuffs to
set up industries to produce all the types of manufactured products that it

consumes. The cost, in terms of resources used, for a small country' without
natural advantages in industrial production, could be very' heavy. It thus
appears that there a large gain to both countries in having specialisation
is

and trade. The output and consumption of both sets of countries would
real
be very much lower if each had to produce domestically all the goods that
it consumed.

Thus, almostall economists would agree, the most casual observ'ation

revealssuch major cost differences among countries that no one could doubt
that there are gains from trade. Careful empirical measurement might put
an actual numerical value on it, but it is inconceivable that it could refute
the general hypothesis that production and
and consumption in the w'orld,
in each major trading country', is higher with trade than it would be if all
countries were forced to be utterly self-sufficient.

FREE TRADE VERSUS THE LEVEL OF TARIFFS EXISTING

today
It quite a jump
is
from the proposition that ‘Some trade is better (because
it increases production) than no trade’ to the proposition that ‘A bit more
1 The theory is were a general law that cannot be challenged on any
often presented as if it
grounds. It unassailable only in the sense that it undoubtedly follows logically from certain
is

postulates. But in this sense, any theory to which the rules of logic have been correctly applied

25
770 THE INTERNATIONAL ECONOMY

trade than at present is better than a bit less trade’ Yet most
we have
arguments about commercial policy invoKc the latter sort of proposition
not the former Most actual policy disagreements concern the relative ments
of free trade versus controlled trade with tanffs on the order of, say, 10, 20
or 30 per cent Such tariffs uould not cut out imports of bananas, coffee,
sugar, diamonds, iron ore, or any of the commodities in whose production
we would be really incflicient Yet these arc just the commodities that
defenders of free trade use as examples when the doctrine of the gams from
trade is challenged If we accept the hypothesis that some trade is better
than no trade, we are not necessarily committed to accepting the hypothesis
that free trade is better than controlled trade with, say, 20 per cent tanffs
Let us, then, compare a position of free trade vsith one of, say, 20 percent
ad laloftm tariffs on all imports This is a simplified vcnion of the sort of

argument that really does take place over commercial policy, for tariffs
are seldom advocated to protect industries that arc violently inefficient
compared to foreign industries, they arc usually advocated to protect
industries that can very nearly compete, but not quite
Let us predict, as comparative cost theory does, that there will be gains
when we move from a position of, say, 20 per cent tariffs to one of free
trade If we have any doubt as to the empirical correctness of this pre-
diction we must look beyond the examples of bananas and sugar, for 20 per
cent tariffs will still allow trade in these goods In fact, there can now be
no adequate substitute for a careful empincal test of this hypothesis But, in
the absence of such a test, can we gam any idea of the possibility that the
hypothesis is false’ We can indeed gel some idea of what is involved by
changing some of the assumptions of the theory m wa)^ that seem relevant
and seeing what then happens to the prediction about the gams from trade
Let us take a few examples
The theory is baised on an assumption of competition that ensures that
relative prices will reflect real opportunity costs If the degree of competition
differs between different mdustnes, relative prices may not reflect compara-
tive costs Under these circumstances, free trade may force countries to
specialise ingoods in which they have a comparative disadvantage In such
circumstancTs,. Jradf wsB reduce wiwld The ccajclusion that there
are gams from trade is thus sensitive to variations in the assumption about
the relation between relative pnees and relative costs
The theory is based on the assumption that the existence of trade
also
does not affect domestic production possibilities The removal of a 20 per
cent tariff might create trade with a neighbouring country, but labour
w unassailable Like any other inCeresUng theory it u open to testing and to possible refutation
After alt any one of a half-dozen postulates on which it is based might prove to be empirically
false
 TARIFFS AND THE GAINS FROM TRADE 771

unions might oppose this policy and express their discontent in strikes and
slow-downs that would permanently lower output per man-hour. Under
such circumstances, trade might reduce the total output of the two
countries.*
The theory is based on the assumption of full employment. Most people
would agree that free trade would not be ivorth having if its price were
massive unemployment. But what if free trade led only to a slightly higher
average level of unemployment than a situation of 20 per cent tariffs? If it
did, then everything ivould depend on the magnitude of the gains resulting
from resource reallocation under free trade. If, for example, free trade led
to a reallocation of resources that was -5 per cent more efficient than one
resulting from 20 per cent tariffs, but led simultaneously to an average level
of unemployment 1 per cent higher than the one occurring with 20 per cent
tariffs, then free trade would bring losses rather than gains. We conclude,
therefore, that the assumption that free trade will bring gains depends
critically upon the assumptions that productivity per man-hour is not
lowered and the level of unemployment is not raised by the removal of
tariffs. At present, we have very little empirical evidence relating to either

of these conditions. It is critical, therefore, that ^ve know the order of

magnitude of the potential gains from free trade if these gains are small,
;

they could be \viped out by a small real-world deviation from one of the
many assumptions on which the theory is based.
The argument in the previous paragraphs does not refute the hypothesis
that there are gains from free trade as compared with, say, 20 per cent
tariffs. Such an hypothesis can only be refuted by measurement and testing.

In the absence of careful tests, such arguments as we have presented can

cast doubt on the idea that the hypothesis is self-evidently true. The
argument shows that the hypothesis is not self-evident; that its truth
depends on a large number of circumstances about which we can by no
means be certain ivithout careful measurement.
We must conclude that the proposition, ‘Some trade is better than no
trade’ might be accepted as self-evident on the basis of very casual observa-
tion, but that the proposition, ‘Free trade is better than some trade (with,
say, 20 percent tariffs) is by no means self-evident and need not be accepted
’

in the absence of careful empirical tests.

THE COMMON MARKET: HOW LARGE ARE THE GAINS?

Shortly after the Second World War, the movement toward a European
Common Market began. The six Common Market countries, France,

1 Thereno point in saying this does not upset the theory on the grounds
is
that, if workers

could be made to work as hard after trade as they did before, all would be well. We knote
772 THE INTERNATIONAL ECONOMY

Germany, Italy Holland, Belgium and Luxembourg were to remove all

tariffs on trade with each other As controversy developed about the

advantages and disadvantages of such a union and about the cost to other
countnes of staying out or the gains from coming in, economists set out to
try to measure the gains from such a union There have been three notable
empirical measurements Professor P J Verdoorn estimated the gam to the
SIX European Common Market countnes to be something on the order of

^ of 1 per cent of their national incomes' Professor Harry Johnson esu

mated the maximum cost to Britain of staying out of the Common Market to
be equal to approximately 1 per cent of her national income Professor
W Welmesfelder estimated the gain to Germany from major tariff reduc-

tions in the years 1956-7 to be on the order of less than I per cent of
‘
German national income
These measures came as a surprise to those convinced free traders who
believed the loss from existing tariffs to be large
There arc, however, good
general reasons for believing that the orders of magnitude obtained in these
three studies are the right ones Typical European lanffs on manufactured
goods are in the order of 20 per cent This means that industries from 1 to

20 per cent less efficient than foreign competitors will be protected by these
tariffs If the costs of different industries are spread out evenly, then some
tariff protected indusmes would be 20 per cent less efficient than foreign
competitors, but others would be only 1 per cent less efficient, and their
average inefficiency would be in the order of half the tariff rate, which is
10 per cent less efficient than foreign competitors ^ Typically, not much
more than 10 per cent of a country s resources would be devoted to pro
ducing behind tariff walls This means that 10 per cent of a country s
resources would be producing 10 per cent less efficiently than if there were
no tariffs, which makes a reduction in national income of something in the
order of 1 per cent

already that if all the cond tions assumed by the theory are fulfilled there will be a gam from
trade What we want to know if we are to give policy advice in the real world is Are the

conditions in fact fulfilled’

These studies are not suitable for the elementary student For the best account of
1

Veriioorn s wofK seeT ^citovs'icy ^aimnicTiuerf aniVffstcm European

Unwin 1958) See also H G Johnson The Gains from Freer Trade An Estimate Manchester
School March 1958 and W
Welmesfelder The Short Run Effects of the Lowenng of Import
Duties in Germany Economic Jountal March 1960 There are reasons to believe that
Welmesfelder s figure is a large overesumatc
2 If in the absence of foreign competition the sheltered industnes become slack and
inefficient their costs may all come to settle at 20 per cent higher than foreign costs This
implies that foreign competition is more effective for encouraging efficiency than is dotn«uc
competition and it makes the losses from moderate Unffs higher than they appeared in the
calculation above
 TARIFFS AND THE GAINS FROM TRADE 773

^Ve must conclude that, on the best available etddence, the maximum
gains from the extra trade resulting from removing 20 per cent tariffs is in
the order of 1 per cent of national income. When one considers the violent
feelings and passionate controversies over the differences between a policy'
of free trade and one of 10, 20 or 30 per cent tariffs, it is understandable why

many people were surprised at this figure when it was produced. This figure
is itself important, because one would view the use of tariffs for ‘non-

economic’ reasons in quite a different way if the cost of these tariffs were a
once-and-for-all reduction of 1 per cent in the national income rather than,
say, a 10 per cent reduction.* There is, indeed, a world of difference be-
tween merely establishing the direction of change and in actually measuring
its magnitude.^

THE CASE FOR TARIFFS

^Ve shall now consider some of the common arguments used in favour of
tariff protection.

Argument 1 : 'Mutually Advantageous Trade is Impossible'

Since we have already shown

this proposition to be uTong, it is not sur-

prising that the arguments based on it all contain crude fallacies. Consider,
for example, the argument that runs: ‘If I buy a foreign good, I have the
good and the foreigner has the money, whereas if I buy the same good
locally, I have the good and our country has the money too therefore we are ;

better off.’ The reader is left to provade his oivm retort to this argument by
recalling the discussion in Chapter 53.
Probably the most subtle argument that comes under this heading asserts
that it is impossible for a rich, high-wage countr)' to trade profitably ivdth a

poor, low-wage countr)'. It is argued that the low-wage country will under-
sell the high-wage one: that unemployment will ensue; and that the
standard of living of the high-wage country will be sacrificed. ^Ve considered
this argument earlier and saw why it was misleading. You should re-read
pages 745—6, if you cannot refute the argument.

1 Some economists give the impression that the losses from tariffs would be large when they
say that it may be necessary’ to become partially self-sufficient in certain lines of activity
at great cost because of fear of future wars. The attempt to imply that the cost will be great is

unwarranted in the light of present evidence.

2 This not to say that the gains from European union are necessarily small; it does say
is

from a reallocation of resources according to comparative advantage are

that the gains resulting
small. For a discussion of the other possible sources of gain, see R. G. Lipsey, ‘The Theory of
Customs Unions: A
General Survey.’ Economic Journal, September 1960.
774 THE INTERNATIONAL ECONOMY

Argument 2 Living Standards will be Higher with Tariffs than with

Free Trade

Most of the common arguments in this group are concerned with single
‘ One argument concerns the effects of trade on
countnes employment The
classical argument for the gams from trade assumes full employment If it
were true that a free trade country had a higher average level of unemploy
ment (perhaps because of wider cyclical movements in unemployment) than
a country levying tariffs, then it is possible that the loss in production
through unused resources would more than offset the gam through a more
efficient allocation of resources No detailed studies have been made of the
effect of trade on employment over any long period of time, m the absence
of any evidence one way or the other, no verdict can be rendered on this
possibility
Another argument says that if a country produces a significant portion
of the world output of some commodity, it will be able to exploit its

monopoly by mterfenng with the free flow of trade By buying less

position
from abroad and selling less as a result, world prices will be affected and the
country concerned can appropriate for itself a larger share of total world
production than it would obtain if all prices were set on competitive
markets If other countries follow a fairly passive policy, then one country
may be able to reap quite substantial monopoly gams K, however, several
countnes all try to do the same thing, then a battle of move and counter
move may ensue until, at the end, everyone may be worse off than they were
under free trade
Probably the most important argument under this heading is the one
relating to economics of scale which is usually referred to as the infant
industry case for tanffs If an industry has laige economies of scale, then costs
and must be high when the industry is small, but they will fall as the
prices
industry grows In such an industry, the country first m the field has a
tremendous advantage over latecomers A newly developing country may
find that its industnes are unable to compete in the early stages of their
development against established foreign rivals A tanff may protect these
industries from foreign competition while they grow up Once they are
large enough, they wiH be able to produce as cheaply as can foreign nvals
and thus be able to stand on their own feet without tariff support
This has always been recognised as a theoretically sound pro tariff argu
ment Most standard textbooks, however, wind up their discussion of it with
some statement such as the following ‘In practice, these industries never
admit to growing up and, even when they are full grown adults, they cling
1 It IS not difficult however to invent ciFcumstances that apply to the world although
whether or not these circumstances are found in the real world is of course an empir cal
 TARIFFS AND THE GAINS FROM TRADE 775

to their tariff protection j the infant-industry argument, although valid

theoretically, is thus to be rejected in practice’.
But is it true that such tariffs are seldom removed ? This is certainly part
of the folklore of tariffs, but it does not seem to have been established by
any careful study of the facts. Notable illustrative examples spring to min d,
of course, but in order to determine if it is true, we would have to compare
the tariff levels existing in the infant industries with the general level of
tariffsruling in the countr\' both when the infants were really infants and
again after they had grown up. Only then could we say wth any confidence
that there was no tendency to remove the tariffs after the infants matured.^
Second, it is not at all clear, even if this alleged fact is true, that this is a
sufficient reason for avoiding such tariffs. If the economies of scale are
realised, then the real costs of production are reduced and resources are
freed for other uses. Whether or not the tariff remains, a cost saving has been
effected. The existence of the tariff may protect the grown-up industry
from foreign competition and allow it to charge a higher price than it
othenvise could. Thus, if the tariff is not removed, all factors in the industry
may earn more than they othenvise would. In other words, the continuation
of the tariff may redistribute income in favour of factors employed in the
protected industr)' to the cost of the rest of the country. It is quite possible,
however, if there are sufficiently large economies of scale, that everj'one in
the country wll have a higher income than if the industry had never been
protected in the first place. Other cases are also possible, but the point is

that, just because the tariff is not removed, it does not necessarily follow
that there is a loss to the countiy^ as compared with a situation in which the
tariff was not put on at all.

Argument 3: Tariffs are Required to Achieve and Encourage Goals

Other than the Maximisation of Output'

Gains from trade, in the classical analysis, stem from increases in pro-
duction. It is quite possible for someone to accept the classical prediction
about output and yet rationally to oppose free trade because of his concern
with policy objectives other than production and consumption.
Comparative costs might dictate, for example, that a country should
specialise in the production of a few primary products. The central
authorities might decide, hoivever, that there are distinct social advantages
to having a more diverse economy, one that would give citizens a wider

I This kind of comparison is needed because it is not sufficient just to show that the

industries keep some tariff protection even after they grow up instead, we need to show that
;

thev' keep higher protection than we would have expected them to obtain if they could not
have availed themselves of the infant-industry' argument.
776 THE INTERNATIONAL ECONOMY

range of occupations m uhich to develop their talents The social and

psychological gams from having a diverse economy might more than com
pcnsate, the authorities decide, for a reduction in living sundards to a
le\cl, say, 5 per cent below what they could be with free trade

Another possibility concerns fluctuations m the prices of certain primary

goods Comparative advantage might dictate that a country should
specialise in the production of one or two primary commodities that are
subject to w'lde price fluctuations This would mean that the incomes of the
producers would also be subject to wide fluctuations Because problems of
a widely fluctuating national income can be serious, even though the
average level of income over a long period is high, the central authontics
may decide to sacrifice some income in order to reduce such fluctuations *
They could encourage the expansion of several stable industries that arc
protected by tariffs (They realise also that specialisation m one or two
products leaves a country highly vulnerable to shifts m demand due to
changes in tastes or technological innovations that make some materials
expendable
Many Canadians are passionately concerned with maintaining a
separate nation with traditions that differ from those of the United States
Many Canadians believe that the tanff helps them to do this,
of these
and they are fully prepared to accept a 5 or 10 per cent cut m living
standards m order to maintain this independence As many Canadian
economists have argued, Canadians may be mistaken in believing that the
tanff helps them m
preserving independence from the United Slates The
mam point, however, is that there is nothing irrational in their being
willing to accept substantial costs in order to obtain objects other than the
maximising of living standards There arc many policy goals other than
maximising national income Although most people would agree that,
celerts panbus, they prefer more income to less, the economist cannot pro-

nounce as irrational anyone who chooses to sacrifice some income in order

to achieve other goals The job of the positive economist is to point out what
the actual cost m income might be
Although one can think of many cases, particularly with the older
countries of the British Commonwealth, such as Canada, Australia and New
Zealand, where a tanff policy was pursued after a rational assessment of the
approximate cost, one cannot help feeling that, as often as not, high tanff
policies are pursued for rather flimsy objectives of national prestige with \ cry

little idea of the actual costs involved One of the reasons why many
economists about free trade is probably that they react
feel so strongly
against the arguments and motives of politiaans advocating high tanffs

1 Ju« a» many firms dccid« lo diversify ibeir outputs so as not to have all ihcir eggs in one
basket
 TARIFFS AND THE GAINS FROM TRADE 777

Many of the anti-free-trade arguments contain crude fallacies, and many

of the motives for advocating tariffs are, to say the least, suspect. The fact
that manyof the arguments used by tariff advocates are incorrect does
nothing to prove the correctness of the free-trade arguments. (We must
never forget that the failure of an argument for some case does nothing to
prove the contrary case.)
Our main concern in this chapter has been not to argue a case for or
against free trade, but to investigate what can be said about trade and
tariffson the basis of positive economics. As in all other realms, positive
economics investigates the consequences of certain actions: it cannot say
which goals one ought to pursue. Whether or not free trade is better than
moderate tariffs, depends on the policy goals that one is trying to attain, the
magnitude of the gains under a free-trade policy as compared to the gains
under a tariff policy, and on the extent to which the policy adopted
prohibits the attainment of goals other than the maximising of consumption.

25*
CHAPTER 55

INTERNATIONAL ECONOMIC
EXPERIENCE

The world into which your grandfathers were bom differed radically from
the world we know today It was a world with a large volume of inter-
national trade largely unobstructed by tariffs, it was a world m which the
BntishNavy policed the seas and guaranteed the free flow of world trade,
U was a world in which the universal acceptance of the gold standard
guaranteed the complete convertibility of one currency into another and,
hence, the ability of merchants to buy and sell where they wanted, un-
impeded by governmental intervention It was a world largely without
passports, a world m which people could cross many international frontiers
as easily as they now pass from one Bntish county to another It was a world
m which there was considerable free international movement of goods and
factors It was a world we are now inclined to regard as nothing more than
the economist’s mental construct That world is gone forever
Today, international trade is restneted and controlled by tariffs and
quotas Today, the sea is patrolled mainly by the American navy, but more
between the East and the West than to
in relation to the political battle
ensuring that everyone conforms to the rules of the game of free trade
Today, the gold standard is known to most people only as something that
existed in the ‘good old days’ Today, most currencies are not freely con-
one into another Today, most governments operate a scheme of
vertible
exchange controls that restricts the ability of merchants to buy and sell
anywhere they want, even if they arc prepared to pay all existing taxes and
tariffs What happened, and does it matter that it did happen ^

ACHIEVING EQUILIBRIUM IN THE BALANCE OF PAYMENTS

Before we consider what happened and why, we need to know how equi-
libnum m international payments established We analysed this problem
is

m Chapter 53 here, we need only draw ti^eihcr the relevant points

,

Establishing equilibrium in international payments in a frec-market

 INTERNATIONAL ECONOMIC EXPERIENCE 779

world requires that changes be made in the relative prices of goods pro-
duced by different countries. A country with a payments deficit must in-
crease its and reduce its imports. These changes are accomplished
exports
if the goods produced by the deficit country fall in price so that imports

become relatively expensive and exports become relatively cheap on the

world market. On the other hand, a country with a payments surplus must
reduce its exports and increase its imports. These changes are accomplished
if the goods produced by the surplus country' rise in price so that imports
become relatively cheap and exports become relatively expensive on the
international' market.
Basically, there are tw'O ways in which these relative price changes can
be effected: exchange rates can be varied, or domestic price levels can be
varied. A deficit country must either allow its exchange rate to depreciate
or its domestic price level to fall; a surplus country must either allow its
exchange rate to appreciate or its domestic price level to rise.
We studied fluctuating rates in detail in Chapter 53. Under a system of
fluctuating rates, changes in the relative prices of each country’s imports
and exports are accomplished by varying the exchange rates; there is no
need to change domestic price levels. If exchange rates are determined by
the forces of demand and supply on a free market, then surplus countries
\vill automatically have appreciating exchange rates and deficit countries

wll automatically have depreciating ones.

Under the gold standard, exchange rates were fixed, and changes in the
relative prices of a country’s imports and exports were accomplished by
changing the domestic price levels. If, under the gold standard, a countr)-
had a persistent tendency for imports to exceed exports, it was necessary
that the country’s price level be reduced. A fall in the domestic price level
w'ould mean a fall in the prices of all goods produced in that country, both

those for domestic consumption and those for export. A fall in the domestic
price level would thus mean that imports, whose prices \vould not have
changed, would become relatively less attractive than domestically pro-
duced goods, while exports, whose prices would have fallen, would become
relatively more attractive in foreign markets. These changes would tend to
encourage exports and discourage imports and so restore equilibrium in the
balance of payments. The gold standard required price-level adjustments
and it also provided an automatic mechanism for accomplishing them. We
do not need to go into the details here, but the essential links in the mechan-
ism were: (1) a deficit country would lose gold; (2) the loss of gold would
cause a contraction in the quantity of money in the country; and (3) a fall
in the quantity of money would cause a fall in the price level.

1 This last link between the quantity of money and

the price level is provided b> the

quantity theory, which was described in detail in Chapter 51.

780 THE INTERNATIONAL ECONOMY

THE GOLD STANDARD

Although the detailed workings of the gold standard are now only of hu
toncal interest, a few features provide important contrasts to the present
system
The gold standard was not designed Like the price system, it just hap-
pened It arose out of the general acceptance of gold as the commodity to
be used as money * In most countries, paper currency was freely convertible
into gold at a fixed rate, the currency was litdc more than receipts for gold,
and it circulated because paper was more convenient to use than metal
Rates of exchange between the standard units of currency of vanous
countries were fixed by their values in terms of the standard unit, gold In
1914 the US dollar was convertible into 053 standard ounces of gold,
while the British pound sterling was convertible into 257 standard ounces
This meant that the pound was worth 4 86 times as much as the dollar in
terms of gold, thus making one pound worth S4 86 US ^ There was nothing
good or bad nor was there any cause for national pndc or shame m the fact
that the pound sterling was worth more than the dollar The British had
simply chosen to make their basic unit, the pound, contain a larger amount
of gold than was contained in the American basic unit, the dollar If the
British had made their basic monetary unit the shilling, the American basic
unit would have been worth more than the Bniish one, since the shilling
was worth just over 24/
As long as alt countnes were on the gold standard, a person any one m
country could be sure of being able to make payments to a person in any
other country If he was unable to buy or sell the relevant currencies on the
foreign-exchange market, he need only convert the currency he held into
gold and then ship the gold
The century between the end of the Napoleonic Wars and the beginmng
of the Pint World War was the heyday of the gold standard dunng this ,

relatively trouble-free penod, the automatic mechanism seemed to work

1 The use of gold was not exclusive Some countnes used a silver standard and factions
sometimes developed because of this In the United Stales bimetallism (the official use of both

guib'anu’suVer as money/ daa'many exponenu m Oie iVtter daiy‘oi''ole inneiVemlVcemti'?'

2 In practice the exchange rate did fluctuate mthm narrow limits set by the cost of shipping
gold If It cost 2/ to ship S4 86 worth of gold from New York to London it would be worth
buying pounds in New York as long as (heir pnee did not rue above S8 At 54 to

pound an American who wished to pay a bill in London would still buy pounds At 54 89 it
would pay him to convert his dollan into g<4d (at S4 86 per pound) and ship the gold (at 2e
per pound s worth) making the total cost of paying bu biU $4 88 per pound rather than $4 89
he purchased sterling Thus no one would buy sterling at a price in excess of S4 88 A similar
if

argument for the British trader wishing to obtain US funds establishes that the rate cannot fall
below $4 84 Tbe values at which it paid b> ship gold were known as gold points
 INTERNATIONAL ECONOMIC EXPERIENCE 781

well. Subsequent research has shown, however, that the gold standard
worked well during the period mainly because it was not called on to do
much work. Trade flowed between nations in large and rapidly expanding
volume, and it is probable that existing price levels were never far from the
equilibrium ones. No major trading country' found itself with a serious and
persistent balance-of-payments deficit and so no major country' was called
upon to restore equilibrium through a large change in its domestic price
level.

In the 1920’s and 1930’s, the gold standard was called on to do a major
job. It failed, and it was abandoned. How did this come about? During
the First World War, most belligerent countries had suspended convertibility
of currenc)' (i.e., they went off the gold standard). Most European countries
suffered major inflations, but the degree of inflation differed from country
to country'. After the w'ar, countries returned to the gold standard (i.e., they
restored convertibility of their currencies into gold). For reasons of prestige,
some on returning at the prewar rates. This meant that some coun-
insisted
tries’ goods were overpriced and other countries’ goods were underpriced.
Deficits and surpluses in the balance of payments inevitably appeared, and
the adjustment mechanism required that price levels should change in each
of the countries in order to restore equilibrium. Price levels changed veiy'
slowly, however, and after five years had passed, equilibrium price levels
had not yet been attained.^
British export industries were badly depressed during the whole inter-

war period and much of the heavy British unemployment during this
period (see Figure 45.1, page 619) was due to the problems of the export
industries. Part of the problem was due to the secular decline in the demand
for many of Britain’s traditional exports. In the case of some commodities
such as coal, there was a decline in world demand, associated in the case of
coal with the increased use of oil as a fuel for ships, while in other cases,

1 Inevitably, there were short-run fluctuations and these were ironed out by movements of
deficit
short-run capital in response to changes in interest rates. Interest rates tended to rise in
short-run capital to cover any balance-of-
countries and this tended to attract sufficient

payments deficit, providing the deficit was short-lived. The way in which this is accomplished
was spelled out in the Appendix to Chapter 53.
been
2 The failure of the mechanism of price-level changes to work quickly has often
growth of price and wage rigidities in the twentieth century'. The
student
attributed to the
encounter explanation.
who consults some of the standard works on the subject may well this

Recent research has shown, however, that the flexibility of the price and svage levels in the
1920’s was no less than in the nineteenth century. Thus the success of the
gold standard in the

nineteenth century is not to be attributed to the fact that the mechanism worked well, but
that it did not have to work at all. See A. W. Phillips, ‘The Relation Between
the Level of
United Kingdom from
Unemployment and the Role of Change of Money Wage Rates in the

283-99; and R.G.Lipsey, ‘Can There Be

a
1862 to 1958’, Economica, November 1958, pp.
105-12.
Valid Theory of Wages?’, The Advancement of Science, J\i\y 1962, pages
782 THE INTERNATIONAL ECONOMY

such as cotton textiles it was associated with the nse of competing sources
of supply m
other countnes
Part of the problem was, however, without doubt associated with a
return to the gold standard at the pre-war rate of £\ = $4 86, which, due
to thewar-time inflation m Britain, was then considerably above the equi
libnum rate We have studied in Chapter 53 the consequences of pegging
the exchange rate above its equibbrium imports will be higher, exports
lower than at the equilibrium rate If the government was not prepared to
devalue the pound the only alternative was to attempt to force down the
domestic price level so that Bntish goods could again become competitive
abroad In the event, wages and prices fell very slowly at home and five
years after the return to the gold standard the necessary cut m the domestic
price level (about 10 per cent was required) had not been Wages
effected
and prices were pushed down in coal and other industries but only very
slowly and at great cost m terms of industnal strife, the legacy of which is
with us 30 years later m terms of mutual distrust between workers and
still

employers The General Strike of 1926 was a direct result of this policy, as
were umumerable other labour battles, and also the high unemployment
rate and the consequent loss of the output of 10 per cent of the country's re
sources over a period of five years
The alternative policies facing the government during the period were
to change the value of the pound to about $4 US by the mere stroke of a
pen, or to attempt to force doivn the whole price level to accommodate it

to the pre-war exchange rate of the pound The policy of forcing down
wages and pnees was chosen at tremendous cost rather than the policy that
could be accomplished by the stroke of a pen
When one looks back on this whole sordid and largely unnecessary
struggle one can be forgiven for asking what pnee ‘the prestige of the
pound’ - which was devalued m
any case after 6 lean years of attempting to
hold Its too-high value
No wonder MrL S Amery replied to the charge of the Commission in
vestigating the coal industry that the mine union leaders and mine owners
were the stupidest men in the country' by observing that the commission had
‘
omitted the prior claim of the government itself, whose financial policy
was so largely responsible for creating the situation in which both sides
^ ’
found themselves
1 See page 534
2 Quoted in David Thompson England n the Twentieth Cenlurf op cil page 1 10 There is

no doubt that the folly of the government was abetted by the majority of the professional

economists thought that the downward revision of the price level could be accomplished
who
within a year or two (although even that teemed a heavy cost to pay for mere prestige) As
with so many other issues of the day Keynes saw cleariy the folly of the government policy
See his Economic Consequences of Mr Churchill Tlic student might be interested in ponder
 international economic experience 78:

THE 1930'S: PERIOD OF EXPERIMENTATION

\Vith the abandonment of the gold standard, the exchange rates were lef
to be determined by the free market. The way in which demand for anc
supply of foreign exchange determines exchange rates on a free market ha;
already been described in Chapter 53. Various experiments were tried ^vitt
both fixed and fluctuating rates. Often rates would be allowed to fluctuati
on the free market until they had reached what looked like equilibrium, am
the rate would then be fixed at that level.
The period of experimentation coincided with the Great Depression o
the 1930’s. Trade was evennvhere being reduced due both to rising unem-
plo\Tnent and fallingdemand and to the increasing uncertainty and doubi
about the future of international markets. Furthermore, this was a perioc
during which many established ideas about the gains from trade wen
challenged, and many old-fashioned, long-discarded ideas on the desir-
ability of exporting and the undesirability of importing were re\dved. Ir
the terrible period of mass unemployment of the 930’s, go\’emments begar
1

to cast around for any measure, no matter how extreme, that might alleriatt
matters. One way of doing this was to cut back or
superficially plausible
imports and to produce the goods domestically. Prejudice against foreignen
is a potent force in most countries, and the idea that one’s troubles are due

to an influx of cheap imports from abroad has a strong appeal to the elec-
torate everywhere in the world. As a result, most countries in the 1930’s
sought to reduce imports and to encourage exports in an eflbrt to lo\vei
unemployement. After much debate Britain dropped her traditional free
trade position and introduced modest tariffs with preferential treatment foi
members of the Commonweaith and the Empire.
If one country managed to reduce its imports, then its unemployment
might be reduced because people would be put to work producing goods at
home to replace goods formerly imported. But this countr>'^’s imports are
the exports of other countries. These other countries would find their ex-
ports falling and unemployment rising as a consequence. The world level
of unemployment would not be reduced, for the first country’s success in
cutting unemployment would cause an increase in unemployment in other
countries. These other countries w'ould then retaliate by reducing their own
imports and trying to lower their unemployment by producing the goods
at home. This ^vould affect the first country, which would now find its
exports falling and unemployment rising as a result. The net effect of such
measures was to decrease the volume of trade, and thus to sacrifice the gains

ing the parallels between that situation and the one facing the British Labour Governmeni
to wha
in 1966. Surveying this whole sorry period from the VEmtage point of 1966, one wonders
consequences of the mystique for the prestigt
extent we are still suffering from the economic
of the pound.
784 THE INTERNATIONAL ECONOMY

from trade without raising employment, since the simultaneous attempts

of all countnes to cut imports svithout suffering a comparable cut in exports
*
was bound to be self defeating
The policy of discouraging imports and encouraging exports can be
achieved through two main sets of tools The first set is composed of taxes,
tariffs, subsidies, and quotas For example, a policy of tariffs on imports and

subsidies on exports will tend to discourage imports and encourage exports

The second policy is that of devaluation We have already observed that
both policies will work only as long asother countnes remain passive Con
sider what happens in the case of devaluations If a country with 10 percent
of Its labour force unemployed devalues its currency, two effects can be

expected first, exports should rise, and, second, domestic consumers should
buy fewer imports and more domestically produced goods Both of these
changes have the effect of lowering the amount of unemployment in the
country If other countries do nothing, the policy succeeds But the policy
will have raised the volume o( iinemp\o)tneni in these other countnes, since

their exports to the devaluing country will be reduced If other countnes

try to restore their positions, they will devalue their currencies as well If all
countries devalue their currencies in the same proportion, they will all be
right back where they started, with no change in the relative prices of goods
from any country and, hence, no change from the original situation^ A
situation m which all countnes devalue their currencies in an attempt to

gain a competitive advantage over one another is called a situation of

competitm devaluations
Changes m
exchange rates can be made for two quite different reasons
The first reason isto try to cut domestic unemployment by reducing imports
{and thus necessarily raising foreign unemployment) m
the manner analysed
above Wc have seen that such attempts are self defeating when they are
made by everyone The second reason is a desire to find an equilibrium
position in international payments Ifevcryone changes exchange rales with
the second reason in mind, the policy may well succeed Economic theory
predicts that (1) under quite general conditions, an equilibrium will exist,
and that (2) the approach to the equilibrium will be signalled by a reduc-
tion m
excess demands and excess supplies Thus, the search for equilibrium
through changes m
exchange rales need not be self defeating

1 The problem was that total world demand was too low to employ all the world s pro
The attempt of all countnes simultaneously to ga n a larger Tract on of this
ductive resources
inadequate demand was bound to be selfdefcaung
2 Under a paper currency system a simultaneous devaluation of all currencies has no effect
If however the devaluations are accomplished by raising the price of gold as they must be
under a gold standard then the sole effect is to ennch the gold producers of the world m
proportion to the devaluations
 INTERNATIONAL ECONOMIC EXPERIENCE 785

Were the 1930's a Fair Test of Flexible Exchanges?

The changes in exchange rates in the I930’s serv'ed to discredit flexible

exchanges as a means of reaching international equilibrium. It is important
to remember, however, that the experiment was conducted under very
unfavourable circumstances. In the uncertain and depressed circumstances
volume of trade was bound to be adversely affected. What-
of the time, the
ever system was in existence at the time was likely to become associated with
bad times and international uncertainty, even if there was no causal con-
nection between the system and the circumstances.
Without doubt, changes inexchange rates were effectively and correctly
means of getting all nations out of the slump. When every'-
discredited as a
one played with more or less equal skill at the game, the resulting competi-
tive devaluations brought no gain to anyone. Undoubtedly, this experience
has made any use of flexible exchange rates seem dubious. In particular, if

full-employment is being achieved in all countries the use of variations in

the exchange rate to remove surpluses in some countries’ payments and
deficits in thepayments of other countries is feasible. A widespread fear
of competitive devaluations persists to this day, and, whenever anyone
suggests changes in exchange rates as a means of achieving international
equilibrium, others often bring forward the evidence that changes in
exchange rates failed to reduce world unemployment in the 1930’s and
produced only futile rounds of competitive devaluations. The 1 930’s, how-
ever, was a period of particularly extreme conditions. Such conditions do
not characterise every disequilibrium in international economic affairs.

THE POSTWAR PERIOD

In order to achieve a system of orderly exchange rates that would be con-
ducive to the free flow of trade, representatives of the Allied nations met at
Bretton Woods, Hampshire, in 1944 to agree on a system of exchange
New
rates for the postwar world. The one lesson that everyone thought they had
learned from the 1930’s was that a system either of freely fluctuating ex-
change rates or of fixed rates with eeisiiy accomplished devaluations was the
road to disaster in international affairs.

The International Monetary Fund

What emerged from Woods conference was the organisation
the Bretton
called the International Monetary Fund (also referred to as the IMF or the
Fund). The Fund was designed to guarantee the maintenance of fixed
exchange rates in the face of short-term fluctuations, and to guarantee that
786 THE INTERNA-TIONA-l. ECONOMY
changes m exchange rates would occur only in the face of long-term, persis
tent deficits or surpluses m the balance of payments and that when such
changes did occur thc> would not spark offa senes of competitive devalua-
tions
Members of the IMF agree to keep a fixed exchange rate between their
currencies and the currencies of all other countnes They agree that they
will make major changes changes
in this rate e
excess of 10 per cent)
(i , m
only in the face of a persistent balance-of-paymcnts disequilibnum and
only after consultation with the officials of the Fund In order to help
governments to maintain fixed rates in the face of temporary fluctuations in
imports and exports or speculative movements of short-term capital, the
Fund IS prepared to lend foreign exchange to members m need of it A
member country that borrows gold or dollars can use it to support its
exchange rate by selling these and buying its own currency Later it hopes
to sell Its own currency and buy back the gold or dollars in order to repay
the loan ‘ The capital for such loans is subscribed by member countnes
partly in terms of gold and partly m terms of their own currencies
Without doubt, the operation of the IMF has helped many countnes out
of temporary balance-of*payments problems that might otherwise ha\c
ended in devaluations There is also no question that the Fund has been a
powerful force m favour of maintaining fixed rates of exchange To econo
mists who think fixed exchange rates are the best system, this appean as an
advantage, to those who think fixed rates (with the necessity of restonng
equilibrium by changing domestic pnee levels) is a cumbersome system, it

appears as a disadvantage This is not to say that changes in exchange rates

have not occurred Member countnes have changed exchange rates more
or less when it suited them and often without pnor notice to, let alone
consultation with, the Fund

The Internationa! Bank for Reconstruction and Development

The other major world financial msUlution emerging from the conference
at Bretton Woods was the International Bank for Reconstruction and De-
velopment (also referred to as the IBRD or the Bank) The Bank was
designed to so'lve fhe pro'b'lem o1 "long-term capita*! movements Tnere are
three ways m which this can be done The Bank has funds of its own that
It can loan to needy countnes for development purposes More important,
the Bank can underwrite loans made by developing countries from other
sources Because the Bank has considerable expertise and inside knowledge,
Its guarantee that a borrower is reasonably sound may make it possible for

a developing country to raise loans where otherwise it could not do so In

1 The way m which this works was desenbed in the Appendix to Chapter 53
 INTERNATIONAL ECONOMIC EXPERIENCE 787

the process of this, the Bank can send investigating missions to the
doing
countries concerned and can issue authoritative reports on these countries.
Finally, the Bank can itself borrow money in member countries in order to
finance its loans.
There doubt that the Bank has been a success, although it has
is little

been accused of being too conser\'ative in its loan-granting policies.

Recendy, two new institutions have been established in order to meet the
demand more risky type of loan. The International Finance Cor-
for the
poration (IFC) is empowered to make loans directly to individual firms
and not, like the Bank, only through governmental organisations. The
International Development Association (IDA) deals with governments, but
makes loans on rather riskier projects than does the IBRD.
These postwar organisations represent a really notable achievement in
the field of international cooperation. When one compares them with the
ineffectual attempts at cooperation fostered by the League of Nations after
the First World War, one realises that, in a very short time, we have come
a very long way in the field of international economic cooperation.

The General Agreement on Tariffs and Trade (Gatt)

One of the most notable achievements of the postwar world in moving back
from the high-water mark of protectionism achieved in the 1930’s was the
General Agreement on Tariffs and Trade. Under this agreement, GATT
countries get together periodically to negotiate bilaterally on mutually
advantageous cuts in tariffs. They agree in advance that any tariff cuts
negotiated in this way be extended to all member countries. Some
will
significant tariff reductions have been effected by the member countries,
but the total results have fallen far short of the hopes of the founders. None-
theless, the agreements under GATT have probably prevented nations

from solving various short-term problems by increasing tariffs. Although

are not as low as free traders might wish, it is probable that, without
tariffs

thisimaginative attempt at post-war cooperation, they would have been

higher than they now are.

The European Economic Community

In 1945, Europe seemed on the verge of famine and collapse. The war-
devastated countries were ready to forget their narrow-minded nationalism
of the past and engage in a joint effort to meet their common crisis of
insufficient food, shelter, and fuel.In 1947, America came forward with the
Marshall Plan, which gave US aid and encouragement to the devastated
continent. Less than a decade later, the nations of Western Europe were no
longer in need of any form of economic aid. Indeed, they were achieving
 1

788 THE INTERNATIONAL ECONOMY

rates of economic growth well above that of the United States, and were on
the move toward an economic union that may possibly be the first step m
an eventual political union
The European nations solved the immediate post-war balance-of-pay-
mcnts problems among themselves by an ingenious arrangement called the
European Payments Union (EPU) At the same time, they removed
national boundaries insofar as these affected the movement of coal, iron,

and steel This required that the cooperating countries harmonise their
coal, iron, and steel tax and subsidisation policies and to adopt a common
world The successes in achieving these goals
tariff policy against the outside
were critical, since the moves were test cases for the possibility of achieving
a completely tanff-free single market on all goods m Europe The Com-
munity IS now leading a satisfactory, if not altogether untroubled, existence
Tariff reductions are now being made according to a schedule that will
eliminate all tariffs on manufactured goods within the Community by 1970
If the progress continues. Western Europe will be, before the end of the
century, a single economic community with a free movement of goods,
labour, and capital among the six member countnes, Germany, Italy,

France, Holland, Belgium, and Luxembourg

Supporters of the Community suffered a great setback in 1963 when, after
months of negotiations, France vetoed Great Britain’s application to enter
the Community Nonetheless, the results have been impressive, and there
are many who hope, and more than a few who believe, that after economic
union has been achieved political union will inevitably follow
In 1965, a senous crisis developed over agriculture We saw in Chapter 1
that governments intervene in agricultural markets in many ways and for
many purposes The problems of harmonising agricultural policies have
proved more difficult than those of harmonising industnal policies The
French have proved even less willing to compromise than they have in the
past, and there are many who believe that, as the inevitable loss of some
national sovereignty to the higher European Economic Community draws
near, the French Government » becoming increasingly reluctant to let the
Community go forward Whatever the reasons, there is no question that the
Community is entering a cntical phase of its existence By 1970, it should
be possible to see if it will be allowed to develop more or less as originally
*
planned, or if its development will be severely curtailed

INTERNATIONAL PAYMENTS CURRENT PROBLEMS

The present system of foreign exchange has three basic characteristics First,
currencies are convertible into each other at fixed rates of exchange with
1 Since this passage was written another hurdle has been successfully surmounted with the
agreement on a common agTicuttural polKy
 INTERNATIONAL ECONOMIC EXPERIENCE 789

only a minimum of interference and restriction by central authorities.

Second, balances that arise out of international payments are settled by
gold flows between the central authorities of each countr}'. Third, in some,
but not all, be bought and sold freely on the open mar-
countries, gold can
being maintained by the action of the central authorities in
ket, the price

the US. This means that the currency of any country can in fact be con-
verted into gold. Gold cannot be bought freely by private citizens in the
US, but if a holder of dollars wishes gold, he merely buys pounds and then
purchases gold on the free market in London. This means that the British
central authorities mil accumulate dollars that they may eventually use to
purchase gold from the US authorities.
We discussed the problems of maintaining a fixed exchange rate in
Chapter 53. Basically, the central authorities stand prepared to support the
price of their own currency by buying it themselves whenever necessary and
supplying foreign exchange in return. This means that reserves of foreign
exchange must be held by these authorities. Normally, the reserves are held
in two forms, gold and certain key reserve currencies, mainly the US dollar

and the British pound. Central authorities are prepared to hold reserves in
the form of dollars and pounds because they are satisfied that these curren-
cies will be convertible into any required currency or into gold at a fixed

price on demand.
The problems of having one’s currency used as a reserve currency are
many. Two of the most important are discussed below. First, a constraint is
put on the central authorities that makes it difficult for them to devalue the
currency in order to adjust the balance of payments. If a country has a per-
balance-of-payments deficit due to the overvaluing of its currency,
sistent
it may need to correct the situation by devaluation.* If other countries are
holding their reser\'es in the form of this country’s currency (or claims of
any sort valued in this country’s currency), the devaluation will lower the
value of these reserves in proportion to the devaluation. If, for example, the
British pound is devalued by 10 per cent, then the dollar or gold value of
the sterling reserves held in the Indian central authorities falls
London by
by 10 per cent. If the central authorities devalue the dollar by 10 per
US
cent, then the gold value of the dollar reserves held by the French central
a.uthorities falls by 10 per cent. Thus the devaluation of a key currency
is

more serious than is the devaluation of some other currency, such as the

German mark, which is not held in large quantities as part of the foreign-
exchange reserves of any country. Second, a key currency can be put under
heavy pressure because of speculations over other currencies. For example,
throughout the mid-1950’s, Germany ran a persistent balance-of-payments

The only to restrict purchases of foreign

exchange by some
1 other thing that it can do is

exchange-control scheme.
790 THE INTERNATIONAL ECONOMY

surplus, m German central authorities were

1957, the belief spread that the
going to correct this situation by raising the exchange value of the mark
This led to a speculative move to buy marks, >vhich could be sold again
at a substantial profit if the revaluation did occur This meant that people
holding reserves in dolhn and sterling attempted to transfer these into
marks Since large volumes of sterling were held, there was a large volume
of sales of sterling, and the Bntish central authorities were hard put to
maintain the value of sterling under this heavy selling Thus we sec that ‘

the central authorities of a country whose currency is a reserve currency

need to have access to very large reserves in case there is a sudden desire on
the part of others to transfer their exchange holdings into some other
currency, even if only for a short time

Adjusting to short-term oisequilibria One of the most senous

problems ansing out of the present system of international payments is how
to maintain fixed exchange rites in the face of short term fluctuations in
the balance of payments both on current account (caused by short term
fluctuations in imports and exports) and on capital account (caused by
speculative movements of short term capital) \Ve have seen that, m order
to stabilise the exchange v atue of one’s currency, the central authorities buy
or sell foreign exchange when the pnee threatens lo move ouuide of pre*
determined narrow limits In order to be able to operate this policy suc-
cessfully the central authorities need to have adequate stocks of gold or
reserve currencies Tliese foreign exchange reserves must be adequate to
meet any temporary excess demand for foreign exchange due to an excess
of imports ov er exports or to an outward movement of capital
One of the major drawbacks of this system is that speculative movements
of capital tend to accentuate any imbalance on current account and thus
to make the system less stable than it oihcrwisc would be Ifsome temporary
increase in imports or decrease in exports makes the demand for foreign

currency unduly large and if this temporary fluctuation persists a little

longer than usual, reserves will be run down to a lev cl lower than is usual
Once reserves get so low that people begin to suspect that the authorities
willbe unable to support the existing exchange rate^ a speculative flight of
capital out of the currency will occur People with money to hold will want
to obtain some other currency whose value seems more secure, and people
who plan to buy the suspect currency will postpone their purchases
wherever possible in the hope of getting the currency cheaper after the
devaluation These quite natural speculative movements increase the dram

1 The Amencan authorities held such large gold reserve* that they could not be embarrassed
by any umporarf move to sell dollars
 INTERNATIONAL ECONOMIC EXPERIENCE 791

of foreign-exchange reserves and may force a devaluation that might other-

rvise not have been required.
One does not, of course, need to ciy' panic whenever one country loses

some of its reser\'es. This iswhat reserves are for. Their function is to offset
imbalances arising in abnormal situations. Serious problems arise only
when the reserves threaten to be exhausted before a temporary^ balance-of-
payments fluctuation is reversed. Or, to be more precise, serious problems
arise when individuals think reserves are about to be exhausted, for, once
they think this, a speculative flight wdll develop that can easily exhaust the
remaining reserves.
In order to be able to weather short-term fluctuations, central authorities
need adequate stocks of foreign-exchange reserv'es in the form of gold and
key currencies. Many countries have been getting along with very small
reserves. They have thus found themselves in situations in which quite
small fluctuations in demand for, or supply of, foreign exchange could cause
a crisis by draining away too high a proportion of total reserves.
How can tlie countries of the world solve this problem of ‘international
liquidity’ ? One cause of the problem is the fact that a very high proportion
of the world’s gold reserves are held in America. The US central authorities
have gold reserves far in excess of what they need to meet any short-term
fluctuation in the US balance of payments. For this reason, the US pay-
ments deficit in the last decade, although troublesome to US authorities,
has been a good thing for the world, since the loss of (excessive) US
reserves of gold has meant an addition to the (inadequate) reserves of other
countries.
The Monetary Fund produced a great increase in liquidity
International
by pooling the contributions of member nations and making these available
to members who are in temporary balance-of-payments difficulties. But

international liquidity is still judged by many to be inadequate, and recently

there has been a great deal of thinking about what can be done to alleviate
the problem. Probably the best-known plan for accomplishing this end is
the Triffin Plan, named after its author. Professor Robert Triffin of Yale
University.
The Triffin Plan is closely related to the proposals that were made by
John Maynard Keynes an International Clearing Union proposals
for —
that were largely rejected when the IMF was established. The Triffin Plan
proposes to convert the IMF into a world central bank. The principal
features of the plan are as follows: each member country’s reserves would
include deposits at the IMF denominated in terms of a new unit of account.
A minimum demand for these deposits would be created by requiring all
countries to hold 20 per cent of their official reserves in this form. The initial
supply of deposits would be created by converting 20 per cent of members
792 THE INTERNATIONAL ECONOMY

existing reserves - mainly m gold, dollars, and sterling - into the new
deposits and by a further deposit of all sterling and dollar balances remain-
ing in official reserves Intermember payments would normally be made in
terms of IMF deposits (Under the Keynes plan this was the only way that
deposits could be used, but the Tnffin Plan allows for the conversion of
deposits into the member’s national currency or, at the Fund’s option, into
gold To ensure that the growing demands of world trade could be met,
)

the IMF could increase its deposits by buying gold or, through loans (at
Its own discretion) and buying and selling securities, by acquiring further

amounts of member countries’ currencies The size of the annual increase in

deposits IS agreed upon in advance, and increases in excess of the agreed
level would require qualified majority votes This provision was to meet one
of the mam criticisms of the Keynes plan i e that it was inflationary As
,

long as, say, a two-thirds majority of countries have to agree on an increase

over and above the normal average increase, it is hoped that the likelihood
of irresponsible inflationary action would be rendered minimal

Adjusting to long-term disei^uilibria Fixed exchange rates also pose

a problem of how to adjust to long-run changes in trading relations The
domestic price levels (and hence the export prices) of the various trading
countnes change for many reasons internal to the countries concerned If
all pnee levels do not move in line with each other, the rates of exchange
that will equilibnate foreign payments must constantly be changing In
general, the equilibrium value of the currency of countries whose price level
is rising fastest will be falling, while the equihbnum value of the currency
of countries whose price level is rising slowest will be rising (Sec Chapter

53, page 759 )

In a system of fixed exchange rates where the equilibrium rates are

changing continuously but the actual rates arc not, we must expect a
situation of more or less continuous payments disequilibna These dis-
equilibna will be linked together, because the very act of one country
solving its disequilibrium is likely to cause a disequilibnum in some other
country This prediction has indeed been borne out by the post-war
experience
The ten years after the Second World War were unique in that the whole
world tended to be m a balance of-payments deficit to the US, which was
m a position of chronic surplus This ’dollar shortage' was a result of the
war, which left the United States as one of the few suppliers of manu-
factured goods still able to produce at a high rate Europe, with her indus-
tries in rums, needed goods for current consumption as well as capital

goods to rebuild her consumption-gcwds industries But until more goods

could be produced m
Europe, there was little to export in return for the
 INTERNATIONAL ECONOMIC EXPERIENCE 793

much-needed imports. For ten years, the dollar problem dominated the
international-payments position. Countries just could not earn enough
dollars through sales to the US
pay for all the goods they wanted to buy
to

from the US. The problem went on for so long that some economists
thought it was endemic - that it tvould be with us forever.
No sooner had other countries got used to accepting the dollar shortage
as a fact of life than it ended! During the 1950's, productive capacity grew

at a phenomenal rate in Europe and, as it did, the ability to sell to the US

grew and the need to buy from the US shrank. At the same time, US invest-
ment abroad rose at a very rapid rate. As a result of both of these develop-
ments, the nations of the world one by one found themselves no longer in a
chronic payments deficit to the US.
The dollar problem was hardly solved when the problem arose of the
US gold drain. Instead of having a chronic surplus, the US began to have
a chronic deficit on the balance of payments. We have already mentioned
that this was not disastrous in the short run. A redistribution of the world’s
gold stock helped the liquidity position of other countries without doing
any real harm to the US. The gold drain could not, of course, go on for-
ever, and, unless things changed of their own accord, something would
have to be done. In
fact, mild restrictions were placed on tourist expenditure
abroad (by cutting the value of goods a tourist could bring back duty free),
and considerable reductions were made in the volume of overseas militaiy
and aid expenditures.
balance-of-payments problems, the US did not,
In attempting to solve its

nor could be expected to, discriminate among countries. This means,

it

however, that as the US moves back into a payments equilibrium by reduc-

ing foreign expenditure and increasing foreign receipts, those countries that
were in the smallest surplus position rvill be pushed into a deficit: the dis-
equilibrium, in other words, is merely transferred abroad. The severe
foreign-exchange problems of some European countries, particularly
Britain during 1964 and 1965, were partly the direct result of the changes
that reduced the deficit in the US
balance of payments. Indeed, as long as
the world is, for better or for worse, committed to fixed rates of exchange,
and as long as the fixed rates are not the equilibrium ones (and as long as
price levels are constantly changing at different rates in different countries
for reasons unrelated to the balance of payments, we cannot expect any
fixed set of rates to
remain the equilibrium ones for any long period of time),
we must expect a permanent balance-of-payments problem to exist and to
be passed around from country to countr)'. As soon as one country solves
itsproblem, another will inherit it. When this inheritor cuts imports or
expands exports to solve its problem, some other country will find its imports
rising and its exports falling and it will move into a period of deficit.
794 THE INTERNATIONAL ECONOMY

Fixed Versus Fluctuating Rates

To what extent are all the problems described above the result of ha\mg
fixed exchange rates’ Would not the ‘problem’ of international liquidity
disappear if exchange rates were left to be determined on a free market’

Why, ma market society, where most people accept the case for leaving
prices and quantities free to be determined by demand and supply, do we
feel that this one pnee should be ngidly controlled by the central authorities ’

Advocates of free-exchange rates argue that the whole set of post-war prob
Icms, such as shortage of reserves and rcstnctions on trade caused by
chronic imbalances in payments, arc of our own making and would dis
appear overnight if we only abandoned the arbitrary government control
of this one key market - the market for foreign exchange If the free market
were allowed to operate, tlien, when a country's currency turned out to be
overvalued, the value would be reduced automatically by the forces of the
free market The value would continue to fall until the equilibnum exchange
rate was achieved No exchange reserves would be needed, since govern-
ments would not be trying to restrict movements in exchange rates There
would be no problem of international liquidity It is also argued that the
activities of speculators would tend to stabilise the rate of exchange If there
was a temporary nse in imports, demand for foreign exchange v*ou!d rise
and the price of foreign exchange would nse If speculators saw this change
to be temporary, they would sell the expensive foreign exchange and buy
the cheap domestic currency Tlvus the change m demand on the part of
speculators would be from the change m demand
in the opposite direction
on the part of importers, and the former would partially offset the effects
of the latter
The opponents of frec-market rates reply that the fluctuations in exchange
rates would add greatly to the uncertainty of persons engaged m inter-
‘
national trade and would greatly lower the \olume of trade as a result
The next step in the argument, the student might expect, would be to
cite evidence Although numerous works exist arguing the case one way or

the other on mtuitiv e grounds, there has been little significant work attempt
mg to assess the argument quantitatively All business is subject to numerous
risks, and we do not have any careful study of how significant would be the

added nsks of the possibility of fluctuations in the exchange rates To know

this, we would need to know by bow much we expected the rate to fluctuate
from week to week and how the uncertainty involved compared to the other
uncertainties of business The student who consults a senes of standard

1 Others argue that international elastiadci are so low that the free market would not work
(see Chapter 53 pages 761 3) Recent empincal research docs not lend much support to
 INTERNATIONAL ECONOMIC EXPERIENCE 795

textbooks and specialised works on the subject will find many of the authors
taking a variety of strong and mutually inconsistent positions on this prob-
lem. Some dismiss free rates as a product of the diseased imagination of
the theorist, and others accept them as the obvious cure to the world’s
problems. The only thing that the whole debate seems to illustrate is that
the less that is known, the greater is the certainty wdth which people hold
to particular views about what is and is not correct. Not all of them can

be right.
There is one bit of evidence; Canada adopted a fluctuating rate for the
decade 1952-62. Opponents of free rates are quick to argue that the experi-
ence of one country is not relevant to what would happen if the whole world
were on a free-market system.^ What the experience did show was, first,

that the day-to-day fluctuations in a free-market exchange rate could be

very small - indeed, they might be no larger than they arc when the rate
is pegged within narrow limits, and, second, that the uncertainties intro-

duced by the fluctuations actually occurring did not have a sufficient disin-
centive effect on trade as to be observable.
The next thing the student might well ask is: Tf we regard the evidence
as inconclusiveand if fluctuating rates seem to have worked once, why do
we not experiment with them again? The worst that could happen would
be that the experiment would fail, in which case we would be forced back
to the present system.’
This is a very hard question to answer. Individuals
and governmental organisations are noticeably reluctant to experiment in
economic matters. The rational case against experimentation would have
to show that the losses during the few years necessary for the experiment
(two or three) would be so large if it did not work as to overweigh the pos-
sible gains over the indefinite future if it did work. So far, however, no one
has even tried to guess what might be.
these losses
In my own view, the evidence docs not allow one
to take a strong stand
on the superiority of fixed rates over fluctuating ones, or vice versa. There
IS, however, fairly
general agreement among the policy-makers of the world
that fixed rates are superior,
and there is little doubt that the present system
willbe continued in the indefinite future. The assessment of whether or not
the system is a good one or a very
bad one will probably have to wait for
the researches of future generations of economists who can be more de-
tached in studying it than can those who are already committed to a specific
answer.

1 The economist should immediately ask, ‘Why?’ If we dismissed evidence merely because
It came from a different time or place, we would admit no evidence whatsoever. Evidence can
only be dismissed as irrelevant
if our theory specifies that, under certain conditions, one result
w'll occur and that, under other conditions, another result svill occur and if the evidence can
be shown to have been
generated by the wrong set of conditions.
796 THE INTERNATIONAL ECONOMY

The world of international economic relations is msome ways a topsy.

turvy one It is a world that awaits someone who will do for it what Keynes
did for domestic unemployment remove forever the necessity of accepting
crises as inevitable and beyond the power of go\ ernments to solve for once

and for all

 PART 10

ECONOMIC GROWTH
AND DEVELOPMENT
CHAPTER 56

ECONOMIC GROWTH^

If you look at total-output data for any Western country over a period of
a half century or more, you will be struck by the dramatic growth that has
characterised theeconomy in question. The growth experience of the
United Kingdom, Japan, the Soviet Union, Germany, France, Italy,
Holland and indeed for most Western countries, is strikingly similar. These
countries’ peoples comprise about 20 per cent of the world’s population.
For the rest of the world, for China, India, and most of Africa and South
America, a history' of growth is singularly absent over the same period of
time. In this chapter and the next, we discuss the nature of economic
growth, its causes and its consequences.

AGGREGATE DEMAND AND AGGREGATE SUPPLY

The PRODUCTIVE CAPACITY of an economy is defined as the gross national
product that can be produced when all of the economy’s resources are
fully and 8 we concentrated on short-term changes in
employed. In Parts 7
national product. We developed a circular-flow theory by taking as given
the eeonomy’s productive capacity and by concentrating on fluctuations in
the expenditure on goods and services; (this aggregate expenditure is often
referred to as aggregate demand). Such demand-induced variations in
income lead, we found, to variations in either employment or prices or both.
We now wish to consider long-term changes in national income and output.
These changes are mainly associated with the changes in the productive
capacity of the economy rather than with changes in aggregate demand.
The difference between the subject of Parts 7 and 8 and the subject
of Part 10, Growth and Development can be illustrated by considering
investment expenditure. One of the critical variables in the circular-
flow theory developed in the earlier parts is the level of investment.
1 This chapter should be read in conjunction with Chapter 21 on The Very Long Run .

They both deal with the same phenomenon from slightly different points of view.
 ECONOMIC GROWTH AND DEVELOPMENT

Variations in expenditure on investment goods will cause variations m

employment m
the investment industnes, variations the incomes of m
households employed by thest industnes, and variations the cxpcndi m
tures of these households These vanations in expenditures will cause,
through the action of the multiplier further variations m national produc
tion and in total employment This analysis concentrates on the effects of
investment expenditure on a^ngatt demand Productive capacity is assumed

to be constant {sec page 565) and national income varies according as the
percentage of this capacitj that is employed vanes

It IS important to remember, however, that investment expenditure will

after a lapse of time of anything between one and five years, also affect

aggregate supply Variations in the productive capacity of a country will be

effecied by changes in both the quantity and the quality of the capital
stock available Such variations arc brought about through investment
expenditure Once the new capital is built installed, and put in operation
the productive capacity of the country is increased, ivhcn all resources arc
fully employed output be higher than before the investment took place
will
In the theory of growth, we concentrate on the long term effects that invest
ment expenditure has on productive capacity (often called aggregate
supply) just as in the theory of the circular flow, we concentrated on the
short term effects that investment expenditure has on aggregate expen
diture (often called aggregate demand)
Because the productive capacity of a country changes slowly, it is possible

when considering variations m income over a few years to ignore such

changes When considering Ionger*run changes - over ten twenty or
fifty years say we cannot ignore changes in productive capacity Indeed
the longer the period over which we
are trying to account for variations
in national income the more important is the extent of productive capacity
and the less important is the degree of utilisation of that capacity

GROWTH RATES SOME SOURCES OF POPULAR CON

FUSION
The between the causes of short term and long term vanations
distinction
in income be clear enough ft is, fiowever, olTen ovenboilea' nr
sfioul'd'

practical discussions about the growth performance ofa particular economy

If one looks at the growth in national income of an economy over three or
four years it will be the net elTcCt of two distinct changes changes m the
productive capacity and changes in the percentage utilisation of this capa
city Changes m
national income due to changes m the percentage utilisa
tion of capacity cannot be sustained for long periods of time
A great deal of confusion would be avoided if the term the growth rate were
 ECONOMIC GROWTH 801

growth of productive capacity, and if comparisons

used to refer only to the
of national-income figures for one country’ over several years were divided
into two parts: changes due to the growth rate, and changes due to varia-
tions in the employment of productive capacity.
Even when we eliminate the problem of variations in the percentage
utilisation of capacity, there are a number of related growth concepts among
which important to distinguish. Part of any increase in the money value
it is

of full-employment output is due to a rise in prices rather than to a rise in

output. If we are concerned with variations in a countiA'’s capacity to pro-
duce goods and services, then we must try to remove the effect of price
changes from our calculations. If, for example, the level of a country’s
income from ,^1,000 million to £2,200 million while at the same time
rises

we would say that mone} national income

the average level of prices doubles,
has increased 120 per cent, but that real national income has increased by
10 per cent, since, if we value the
new output at the original set of prices, it
would only be worth ^£^1,100 million. In everything that follows, we are
concerned with variations in real national income, which is money national
income corrected for changes in the average level of prices.
Even when we are dealing wdth variations in real national income, we
may have to concern ourselves with a number of different but related con-
cepts; the total amount of output, the total output per person in the total
population, total output per person employed, and total output per hour
worked. As an example of the relations between these concepts, note that
a doubling of real national income combined with a doubling of the popu-
lation W'ouldmean no change in per capita income in spite of a large change
in total we are more concerned with per capita income
income. Very often
than with total income. If we are thinking of living standards, we are cer-
tainly concerned wth per capita income. The average living standard of a

country depends on real income per head of population. The growth in

real income per head, however, is likely to understate the growth in produc-
tive capacity for at least two reasons. First, as standards of living rise, the
proportion of the population actually employed will tend to fall as training
periods are extended, as retirement age is reduced, and as every member of
the family does not find it work in order to provide subsistence
necessary' to
for thehousehold. Thus, national income per head of population may rise
more slow'ly than income per head of working population because a pro-
gressively smaller portion actually working. Second, as
of the population is

national income rises, those who are w'orking may work fewer hours, taking
out part of their increase in living standards in terms of an increase in
leisure. If this occurs, we should expect the rise in national income per head

of employed population to be less than the rise in national income per hour
w'orked by the labour force.

26
802 ECONOMIC Growth and development

The nse m real national income per hour worked by the labour force
shows the rise in the capacity to produce goods and services through human
effort and is called a rise in Labour prodoctivitv, the rise in national in
come per head of populatioi^ shows the rise in the actual goods and services
produced per person When we speak of economic growth m this book we
shall mean the rise in the productive capaaty of a country on a per capita
basis

THE CUMULATIVE NATURE OF GROWTH

One of the most important Aspects of economic growth is that it is cumula-
tive If one country grows faster than another, the gap in their living stan-
dards will widen progressively *
By way of contrast, if one country has a

Table 56 1

EFFECT OF DIFFERENT RATES OF GROWTH’

Pereentage rate of growth per ye

1 2 3 5 7

0 100 100 100 100 100

10 110 122 134 163 197
30 135 181 243 432 761
30 164 269 438 1,147 2,946
70 201 400 792 3,043 11,399
100 270 724 1,922 13,150 86,772

‘
Nadonai income m year 0 rquah 100

of resources than another, the gap between their

less efficient allocation

national incomes will not widen progressively on this account Let us

assume that, for one reason or another, country A uses its resources 5 per
cent than does country B On this count, therefore, the real
less efficiently

national income of A will always be 5 per cent below that of B If, on the
other hand, country A uses its resources so that it grows 5 per cent faster
than docs B, then the gap in incomes will widen progressively If the two
countries start from the san^c income, and B grows at 5 per cent per year
and A grows at 5i per cent (i e A's growth rate is 5 per cent greater than
,

B's), then in 10 years ^’s iticome will be 2 5 per cent higher than B s, in
50 years it will be 12 2 per cent higher, and in 301 years it will be double B s

1 If the faster growing country warts from a lower level of income than the slower growing
one the gap will at first narrow n wiU widen only after the former has overtaken the latter
 ECONOMIC GROWTH 803

Table 56.2
AVERAGE GROWTH RATES IN REAL NATIONAL
INCOME, 1953-1962

Year in which
US will be
Percentage overtaken,
rate of growth should the
of per capita growth rates
national in Column 1
income persist

Japan 8-4 1995

Yugoslavia 6-5 2048
Germany 5-5 2005
Italy 5-4 2025
Austria 5-4 2065
France 3-6 2045
Denmark 3-3 2154
Mexico 3-1 2134
Netherlands 3-1 2142
Venezuela 2-7 2235
Belgium/ 2-4 2236
UK 2-1 2139
Philippines 1-8 2590
Turkey M 5735
Canada* 1-1 4366
USA* 1-0 —
1 Xhe growth rate for the US and Canada is understated because 1953 was a year of high
activity and a high percentage utilisation of resources, whereas 1962 was a year of low
activity

and a lower percentage utilisation of resources. The growth in full-employment productive

capacity was probably closer to 2 per cent per year over the decade than to 1 per
cent.

very low by international standards; if it grows at this rate, the US

will not
Even this rate is

for long maintain its position as the country with the highest
standard of living in the world.
Source: Calculated from the U.N. Statistical Yearbook, 1963.

Such are the long-range of tiny differences in percentage rates of

effects

growth. If we now pick hgures closer to currently observed rates of growth,

the comparison becomes even more striking. Consider two countries
start-

ing with the same level of income, but having different rates of growth. If
A grows at 3 per cent while B grows at 2 per cent per year, A s income will
be twice B's in 72 years. You may not think it matters much whether we
 ECONOMIC GROWTH 805

such as the United Kingdom or the United States, where the growth rate
is neighbourhood of 2 per cent per year. If they live in Japan, where
in the
growth has been going on at a rate of more than 8 per cent per year, the
son’s income will be about 10 times as large as his father’s.
From the point of view of raising the absolute living standards of the
poorer sections of the community, growth is clearly more important than
practically an)- other force we could mention including the redistribution
of income. A redistribution of the sort that seems practicable might cause
a maximum, once-and-for-all rise in the incomes of lower income groups of,
The same rise would be accomplished in about three years
say, 10 per cent.
with economic growth of 3 per cent per year. Of course, not everyone bene-
fits equally from growth, and many of those who are poorest will be out of
the labour force and thus least likely to share in the higher wages that are
the primary means by ^vhich the gains from growth are distributed. Thus,
redistributional policies may be required even in a growing economy. But
growth makes it much more feasible politically to do something about
poverty. If existing income is to be redistributed, then someone’s standard
of living will actually have to be lowered. If, however, there is economic
growth and if the increment in income is redistributed (though government
intervention), then it is possible to reduce income inequalities without
actually having to lower anyone’s income. It is much easier for a rapidly
growing economy to be generous to%vard its less fortunate citizens than it is
for a static one to do so.
A further reason for wanting grow'th may arise out of the problems of
national defence. If your country is competing with another for power or
prestige, then rates of growth are important. If your national income is
growing at 2 per cent while the ‘enemy’s’ is growing at 3 per cent, then all
he has to do is to wait while your relative strength dwindles. Khrushchev
had this in mind when he said to the Americans ‘We will bury you’. Also,
the expenses of an arms race, or a space race, are easier to bear, the faster
the country is growing.

THE CAUSES OF GROWTH

Our knowledge on from complete, but there is substantial
this subject is far
agreement that certain things have marked effects on the growth of an
economy’s productive capacity. The cause of grow-th traditionally empha-
sised by economists is capital accumulation. If the state of knowledge re-
mains constant and some of the society’s resources are devoted to new invest-
ment, then more and more capital equipment can be accumulated. This
capital will be productive in all sorts of ways, and it will have the effect of
raising the productivity of labour and so will cause a rise in real nadonal
806 ECONOMIC GROWTH AND DEVELOPMENT

income E\ cntually, ifthere is no growth of knowledge, all the most produc

tive investments will be used up, and, as capital accumulation continues
the marginal productivity of capital will decline and eventually will ap-
proach zero In the meantime (which is likely to be a very long time), new
capital will contribute to a nse national income m
New knowledge and inventions also contribute markedly to the growth
of national income In order to sec this, assume that the proportion of the
society’s resources devoted to the production of capital goods is just sufficient
to replace capital as it wears out Thus, if the old capital was merely
replaced m the same form, the capital stock would be constant and there
would be no increase in national income Now assume, however, that there
IS a growth of knowledge so that, as old equipment wears out, it is replaced

by different, more productive, equipment In this case, national income

will be growing because of the growth of knowledge rather than because
of the accumulation of more and more capital ^ This sort of increase in
national income can come about cither by advancing knowledge locally or
by importing it Underdeveloped countries can adapt techniques already
used abroad but not in current use at home Developed countries, which
cannot often copy superior techniques from elsewhere, must develop new
techniques by research and invention
The causes of growth mentioned so far have concerned the quantity and
quality of physical capital The productivity of labour can be greatly in*

creased by education and by improved health standards The scope is

greater the lower the existing standards When we spend money to do this
we often refer to investments in human capital
Social habits also affect economic growth Certain religious patterns, for
^ It has
instance, are more conducive to economic growth than are others
been argued that the ‘Protestant ethic’ encourages the acquisition of wealth
and IS thus more likely to encourage growth than is an attitude that directs
actmly away from the economic sphere Of course we must not jump to the
conclusion that social and religious structures should be changed so as to
maximise the possibility of growth Economists are interested in the exist-
ence of such relationships If people derive satisfaction from living in a
culture that does not produce growth, we cannot prove that they ought to

1 We do noc need (o engage in a argument about whether the quantity of capital u

fiitile

constant in tho case When a worn out is replaced by a physically

piece of capital equipment
different piece of equipment, the concept of the quantity of capital becomes a convention that
may or may not be useful for the problem at band It u neither interesung nor profitable to
argue whether the real quantity of capital has gone up or has remained constant when we all
agree about what has actually happened in the world
2 See, for example R H Tawney Rtbgmi and At Rise «f Capitalism An liisloncal Study

(Holland Memorial Lectures 1922 [London Murray, 1926]) This work is also available in a
paperback edition
 ECONOMIC GROWTH 807

alter their culture. All that we can say is that if they want growth, their
culture is likely to inhibit it and that the goals and techniques of another
culture are more likely to encourage it.

Legal institutions may likewise affect growth. The pattern of ownership

of land and natural resources may greatly affect the way such resources are
used, and thus may affect their productivity. To take but one example, if
agricultural land divided into very small parcels, one per family, it may
is

be much more difficult to achieve the advantages of modem agriculture

than if they are available for large-scale farming.^ Thus economists studying I

the problem of underdevelopment are often concerned with patterns of

land tenure. With this problem, as with so many others, we often meet a
conflict of values. If we value a society in which every man owns own
his
land, we may have to pay a price for it, because the small farm may be
much less efficient than the large one. On the other hand, the concentration
of land ownership in the hands of a few absentee landowners, who are not
concerned to maximise their can be detrimental to growth. If the
profits,

landlord’s holdings are so vast that he can obtain all the income he desires
without using his land efficiently, he may have little motivation to introduce
advanced techniques. In many societies in which land ownership is con- i

centrated in only a few hands, land reform, which usually implies the I

confiscation, nationalisation, or communisation of land, becomes a necessary

|

condition for growth. Not surprisingly, such reforms are seldom supported i

by existing governments, which tend to support the interests of the eco-

nomically powerful. Land reform can often be accomplished only in the
j

wake of a political revolution.

THE COSTS OF GROWTH

In a world of scarcity, almost nothing is free. Growth usually requires an
investment of resources in capital goods, in education, and in health. Such
investments do not yield any immediate return in terms of goods and services

for consumption. Growth, which produces more goods tomorrow, is achieved

by consuming fewer goods today. For the economy as a whole, this is the
primary cost of growth we shall consider it in detail in a moment.
;

For many individuals, there is another and more personal set of costs of
growth. If an economy is growing, it is also changing. Innovation leaves
obsolete machines in its wake, and it also leaves partially obsolete people. A
rapid rate of growth requires rapid adjustments, and these can cause much
upset and misery to the individuals affected. The decline in the number of
unskilled jobs makes the lot of the untrained worker much more difficult;

1 In many countries, collective farms appear to be more productive than the family farms
they replaced.
 ECONOMIC GROWTH AND DEVELOPMENT

and when he he may well fail to find another, particularly if

loses his job,

he IS over fifty No matter how well equipped you arc at age twenty-five, m
another twenty-five years you arc likely to be partially obsolete Many skills
become completely outdated and unneeded One aspect of this problem is
called ‘structural unemployment*, and we shall consider it later m this
chapter

The OPPORTUNITY COST OF GROWTH Let US supposc that the central

authorities of an economy can establish within limits the growth rate they
desire by choosing the allocation of resources between the production of

growth rate

Ftg 56 1 Altemaiivc growth paths in an economy with various allocations of

resources between consumption and investment

investment goods and consumption goods Further, assume that the economy
IS a fully employed one In this case, every ‘dollar’ diverted to mveslmcnt
goods means that one ‘dollar less can be spent on consumption goods

Figure 56 I shows a producDon-possibility boundary summansing the

cihAvvyopinT itT Cihreinnrtjmy' r4*TOinrGhrectntinTry’*y*'iTt\Vat*i’;?
the point X, where the rate of annual consumption is Oa, and productive
capacity growing at 2 per cent per year This means that every year the
is

production-possibility boundary will move out by 2 per cent, if the economy

continues to devote the same fraction of its resources to investment goods,

1 There are also many noneconomic costs orgrowth unspoiled landscapes become spoiled
the simplerlife of family farming gives way to urbanisation and slums the frustrations of a

complex and changing world harm some people, and so on

 .

ECONOMIC GROWTH 809

the path of growth will be along the ray labelled ‘2 per cent growth rate’. In
twenty years the economy will reach the point x'
Should the central authorities be prepared to reduce the rate of consump-
tion by the amount ab and increase the rate of investment, production
would be shifted to point y. Since more investment goods are being pro-

duced, the growth rate will be increased. In this example, the rate of growth
rises to 3 per cent, and in twenty years the economy will be at point /.

The more rapid rate of growth

is purchased at the expense of a lower rate

of current consumption. worth it?

Is it

To see the issues involved in answering this question, assume that, at

point AT, 15 per cent of the country’s resources were being devoted to invest-
ment goods and 85 per cent to current consumption. To shift to point y
would require devoting 23 per cent to investment and only 77 per cent to
current consumption. How long will it take to regain the loss in consump-
tion if the shift to point y is made ? This depends upon what one means by
‘regaining the loss’. On the assumed figures, it takes only 3y years to reach
point a, where consumption is again Oa. Thus the actual reduction in living
standards lasts only 3} years. Of course, had the re-allocation of resources
not occurred, income would have expanded along the 2 per cent growth
path and, thus, although the actual cut is restored in less than 4 years,
consumption is still well below what it would have been if the re-allocation
in favour of investment had not occurred. In fact, it takes 10 years for the
actual level of consumption to catch up to what it would have been had no
re-allocation been made. This is seen more clearly in Figure 56.2, where we
chart the two growth paths over time: after ten years the growth paths
cross. But the economy has not yet broken even: the darkened area to the

left of point z in Figure 56.2 represents the cumulative loss in consumption

over the decade. It takes an additional nine years before total consumption
over the whole period is as large as it would have been if the economy had
remained on the 2 per cent path.* From year nineteen on, the initial sacri-
fice yields bigger and bigger dividends.

Such a policy of sacrificing present living standards for a gain that one
does not begin to reap for nineteen years is hardly likely to appeal to any
but the very young. Of course, the quantities are only hypothetical (although
they are not unreasonable in the light of present knowledge), and it may be
that a smaller diversion of resources would achieve the increase in growth
rates so that the time taken to break even would be somewhat shorter. This
example is only intended to illustrate some of the costs involved in achieving

1 This comparison looks only at actual amounts of consumption. If we discount the future,
preferring to have a quantity of goods now to a somewhat larger quantity in the future, then
it will take longer than nine years to compensate consumers for the loss of goods
during the first
ten years.

26 *
810 ECONOMIC GROWTH AND DEVELOPMENT

fastergrowth through a re-allocation of resources, to suggest approximate

orders of magnitude, and to surest that the case for pursuing a faster rate
of growth by this means is possibly not as universally acceptable as it might
at first sight seem
Many governments have followed this route the Germans under Hitler,
the Russians under Stalm, and the Chinese under Mao Tse-tung adopted
four- and five-year plans that shifted resources from the production of con
sumption goods to the production of investment goods Many countnes are
using similar plans today Such plans are particularly important when

Fig 56 2 Alternative growth paths

m an economy with vanous
allocations of resources between
consumption and investment

10 30 30 40 so
Time myoars

actual growth rates are very small (say less than 1 per cent), for without
some current saenfice there is little or no jirospect of real j:rowth

WAYS OF CHANGING THE RATE OF GROWTH

Each of the factors that arc thought to affect growth can be the object of
policy action Investment can be stimulated by vanous measures, such as
special tax treatment In the US, ‘investment credits’ and accelerated
depreciation have been used An attempt can be made to stimulate innova-
tion and the growth of knowledge by encouraging research Education and
 ECONOMIC GROWTH 811

health can be improved in many ways. The effectiveness irf influencing

growth rates by these means depends on two basic things: first, how impor-
tant, quantitatively, each of these factors is in influencing growth; and,
second, how easily each of these factors can be changed by government
policy.
A
great deal of research is currently being directed to discovering the
quantitative importance of the various factors influencing economic growth. *
There are many technical problems involved in this work, and there is still
uncertainty about the validity of the procedures used to produce the esti-
mates. The work does suggest, however, that the accumulation of
existing
capital important and that such factors as innovation and education
is less

are more important than had previously been thought. In fact, the best
guess based on current knowledge is that substantially less than half of the
increase in productive capacity in the United States is due to capital
accumulation, while over half is due to other things.

GROWTH AND THE ALLOCATION OF RESOURCES

We have been discussing growth in the aggregate. Growth also has impor-
tant micro-economic implications; Growth will normally result in a re-allocation
of resources throughout the economy.

This re-allocation of resources is caused by the fact that the rate of growth
of the productive capacity of each industry \vill not exactly match the rate
of growth of product of that industry, except by a very
demand for the
unlikely coincidence. There is thus a supply and a demand side to this
re-allocation problem.
Consider first the supply side. The rate of growth of an economy is
nothing more than an average of the rates of growth of the individual
industries composing the economy. The rates of investment and innovation will
differ among industries and so the rate of growth of productive capacity
\vill also differ among industries. Thus, if the allocation of resources among
industries did not change, the rate of growth of production would vary
among industries.
Now consider the demand side. If the central authorities follow a fiscal

and monetary' policy that keeps the economy at or near full employment,
then real income will rise in line with the rise in productive capacity. What
happens to the demands for various goods wll depend on the income elas-
ticities of demand for these goods. The larger the income elasticity of
demand
for a particular commodity, the more rapidly wiU the demand for that

commodity grow as real income grows. In general, demands for commodities

1 A path-breaking work along these lines is E.F. Denison,

The Sources of Economic Growth m
the United States (New York, 1962).
812 ECONOMIC GROWTH AND DEVELOPMENT

With income of less than unity will be expanding less rapidly than
elasticities

the average rate of growth of productive capacity the economy, whereas m

the demands for commodities with income elasticities greater than unity
will be expanding faster than the average rate of growth of productive
capacity
Oneshould take care not to confuse total output with total employment
in thisgrowth context As long as an industry’s income elasticity of demand
exceeds zero (i e as long as the commodity is not an inferior good), output
,

will rise as economic growth proceeds Employment, however, will not

increase unless demand expands faster than productivity An industry in
which productivity is growing at 3 per cent per year and m which demand
ISgrowing at 2 per cent per year will be an expanding industry in terms of
output but a declining industry in terms of employment
It IS an observed fact that income elasticities are different among indus
tries and that these elasticities do not remain constant as real income nscs
It is also an observed fact that the rate of productivity increase is not the
same among industries Together, these facts lead to rather dramatic shifts

\n resources over time The decline of agncuUure and the rise of durable
goods manufacturers in the UK and the dramatic nse of the service indus
tries in the US are notable examples

Growth and the problem of structural unemployment We have

seen that growth requires a continuing re^allocation of resources Such
re allocations arc not achieved without cost The cost to human beings with
heavy investments in education and experience in particular occupations
can be enormous A man of forty-five or fifty whose skills arc special to a
declining industry may, should he lose his job, find the market value of his
services cut drastically Further, if the declining industry is a major part of
the economy of a particular gec^raphical area, then the whole area may be
a declining one, and all of its residents may suffer because of the area’s decline
m demand for all goods and services For example, if the man wishes to
move to an area where jobs are plentiful, he may be quite unable to sell his
house and furniture for anything like what their value would be in a grow
mg area He thus will not be in a position to buy a comparable house and
’.o. b.’A oAvt axca.
Argument about the effect of economic growth on employment is as old
asgrowth itself The first thing the displaced worker sees is that technological
change is destroying his job This is, understandably, what is of immediate
concern to him What he is likely to conclude is that growth destroys jobs in
general and thus raises the average level of unemployment in the economy
His conclusion is not easy to test What we can be certain of is (1) that in
all growing Western economies over the last century, lapses from full
 ECONOMIC GROWTH 813

employment have been temporary; (2) that there has been no tendency for
average levels of unemployment to rise decade by decade over the last
century in spite of a prolonged period of sustained growth; and (3) that
comparing across countries over the last few years, there has been no clear
tendency for higher grow-th rates to go along with higher unemployment
rates. Furthermore, even ifit were established that, in a laissez-faire economy,
higher growth did mean higher levels of unemployment, the macro theory
that we have studied in this book provides the central authorities with the
fiscal and monetary' whereby full employment can be pursued as a
tools
conscious object of government policy, even if it does not occur spon-
taneously. In such circumstances, the cost of growth need not be measured
in terms of higher levels of unemployed resources, but only in terms of what
the resources used to create investment goods might have made had they
been used to produce goods for current production, and in terms of the
human costs involved to those who are left behind in the march of economic
progress.
In the last few years, there has been a revival of the theory that growth
raises the average level of unemployment. The revival has occurred mainly
in an attempt to explain the disturbingly high average level of unemploy-
ment recorded in the United States since 1958, but it is not without rele-
vance to some of the depressed areas of Britain.
The theory is a complex one and in an introductory treatment we can
only consider a couple of the most interesting arguments on which it is
founded. The first argument is that the rate of growth has been accelerating
and that there has been a resulting increase in the number of individuals
displaced from those industries in which the volume of employment is de-
clining. Since it takes time for such persons to move, retrain, and enter a
new job, it is argued that the average level of unemployment will be higher
the higher the rate of growth. A major drawback to this theory is that it is
by no means certain, in spite of all the advances of the new ‘industrial
revolution’, that the growth rate is significantly higher than the average
rate achieved over the last hundred years.
The second and more important argument is that the quality of techno-
logical innovation has changed. In the first industrial revolution, so goes
the argument, technology destroyed the jobs of skilled artisans by inventing
machines to do the work, and created jobs for unskilled workers who could
operate the machines. In other words, formerly skilled jobs were broken
up into a series of unskilled ones to be done by men and machines. Although
the artisan might suffer a reduction in real income, he was not lacking in
employment opportunities, since he was always capable of performing one
of the unskilled tasks on the machines. The new industrial revolution, so

the argument continues, has reversed the technological trend. Now the
814 ECONOMIC GROWTH AND DEVELOPMENT
production process is being reintegrated The machine in the automated
factory now performs all the unskilled tasks, and only a few highly skilled
men are necessary to operate the machine and repair it when trouble
develops Thus the unskilled are the immediate sufferers from modern
technological advance, and the new jobs arc not ones into which they can
step without long preparation Thus, the argument concludes, modern
technological advance is destroying the jobs of the unskilled and, even if

demand and output are expanding everywhere, the jobs for which demand
ISexpandihg are not the jobs that the unskilled can perform At very best,
therefore, we would expect to find a rising level of (mainly unskilled) per-
sons unemployed for long periods of lime, and, at worst, there may be a
rising number of persons who can never acquire the skills necessary to fit
into the new industrial processes
The first point to make about this argument is that it cannot be ruled out

on logical grounds alone It is not self-contradictory, and

it describes a state

of affairs that could exist in the world This makes the discussion of its
relevance to today’s world a question of fact docs the argument descnbe
our world or does it contradict certain observations that we already have’
There can be no question that the demands for skilled labour have been
rising and that job opportunities for the unskilled have been shrinking
Supplies have also been changing, however, and the labour force today is

much better educatedand much more highly trained than it was twenty or
fifty years ago There is no solid evidence that the gap between the skills and

education levels demanded by employen and supplied by workers is any

greater now than it was in cartier times Second, although the statement
about the changing nature of technology may be acceptable as a generalisa-
tion, It Ignores the fact that, in terms of employment, the most rapidly
expanding sector of the US economy is not manufacturing but services
Thus, although it may be very depressing to contemplate a fully automated
manufactunng plant and wonder how to create jobs for the displaced un
skilled workers by inducing further increases in demand for the output of
the automated plant, it becomes somewhat less depressing if one realises
that a large fraction of any increment in demand will fall on service mdus
tries in which there are still many employment opportunities for the un-

skilled and semiskilled Finally, it must be recognised that workers are not
completely immobile and that the pattern of demand is not firmly fixed and
independent of the structure of relative wages and relative pnees When
general demand is high (as it was not m
the early I960’s in the US), there
will be an incentive for employers to assist workers to prepare themselves
for jobs forwhich they may not be equipped by their present training, and
to vary the demands ofjobs wherever possible Also, there may be a change
in the relative wages between scarce skilled workers and plentiful unskilled
 ECONOMIC GROWTH 815

workers. These changes in relative factor prices will provide an incentive

for employers to economise on skilled workers and to use less skilled ones
^vherever possible.
The weight of present evidence and the almost universally accepted opinion among
economists who have studied the problem is that there is no need for the policy of
achieving a low over-all rate of unemployment to conflict with the policy of achieving

a high rate of economic growth^ nor is there any significant evidence that there has been

such a conflict in the United States in the twentieth century}

1 We have concentrated on the United States because this is where the bulk of the empirical
work on structural unemployment has been done in the past decade. Although there is much
evidence that depressed areas of Britain owe their problems to changing cost and demand
structures accompanying world economic growth, there is little or no evidence that the problem
would have been any less if Britain had grown any slower in this century.
CHAPTER 57

UNDERDEVELOPED ECONOMIES

The problems of economic growth that we considered m Chapter 56 arc of

particular concern to the poorer countries of the world In our civilized and
comfortable urban life, most of us lose sight of the fact that a very short time
ago (very short in terms of the life span of the earth, that is) man lived like
any other animal, catching an existence he could from what nature
as best
threw his way It is only 8,000 years since the agricultural revolution when
man turned himself from a food gatherer into a food producer, and it is
only withm the last centunes that any significant proportion of the popula-

tion of the world could look forward to anything but unremitting work m
wresting an existence from a reluctant nature The idea of leisure as a right
to be enjoyed by everyone isvery new in human history There are S,000
million persons alive today The wealthy parts of the world, where people
work no more than forty or fifty houR per week and enjoy substantial
amounts of leisure and a level of consumption at or near that attained by
the citizens of the US, contain only about 20 per cent of the world’s popu-
lation Most of the rest struggle for their very subsistence About 2,000
million people exist at a level at or below that enjoyed by peasants in the
more successful civilisations of 5,000 yean ago
If one were studying the effect of variations from year to year in rainfall,
one would find that, for rich counines such as Great Britain or Holland,
such variations would be reflected in farm output and farm income, for
each inch rainfall fell below some critical amount, farm output and income
would vary m a regular way In poor countries, such as China and India,
vanations m rainfall are reflected in the death rate Indeed, many live so

close to the subsistence level that slight fluctuations in the food supply bring
death by starvation to large numben
The fact that fluctuations that are measured in money units m rich

countries are often measured in lives m poor ones makes the problems of the
economy look very different in different countries It also makes the prob-
 UNDERDEVELOPED ECONOMIES 817

lems of economic growth very much more urgent in poor countries than
in rich ones. Reformers in poor countries often feel a sense of urgency not
their counterparts in rich countries. To get those citizens now alive
felt by
offa bare subsistence standard in a ver>'
poor country' requires an immediate
change to a very rapid rate of economic growth.

THE MEANING OF UNDERDEVELOPMENT

We must now sec economics can shed on the problem of raising
what light
that an important
living standards in poor countries. It is often thought
economies is to find
step in dealing with the problem of underdeveloped
on underdevelopment
the ‘true’ definition of underdevelopment. Books
by raising the conceptual problem of how to define
underdevelop-
often start
meaning we attach to the
ment. The only reason we wish to make clear the
area of concern - to say
term underdeveloped is that we wish to delimit our
and not
‘We are concerned with these countries and these problems,
those countries and those problems’. Of course, different investigators

concern themselves with different groups of countries

and different P™
lems, and it would be utterly futile to argue which set of problems s m
by other terms, ne
be described by the term underdevelopment, and which set
e one
investigator, for an underdeveloped country
example, defines
to
con nes
with a per capita national income of less than S500. Another
underdeveloped resources
study to countries tvith substantial quantities of
is comp
Each of these groups of countries can be studied, and it
ope J
properly defined as underdeve
.

to argue which is the one that is

Underdevelopment has many aspects.We may measure eve

per head, t e
dozens of different ways, among them income
per hea con uc
resources unexploited; capital per head; savings
,

growt
to o oca
growth of the social system; conduciveness
e
roads, railroads, schoo
s, ,
amount of ‘social capital’ (i.e.,

education of the working classes, and so on. ,

^
of
Countries that are ‘underdeveloped’ in one
t

one ^ j^ave
underdeveloped in another of them. For examp
e,
utilisa-
have a P
a lower income per head than others, but ^ jj ^

.ion or existing Lnrnl resource, than

seek a unique ranking of the various as
ed than another
ment. We cannot say one country is more ^
of underdevelopment,
o
long as of these characteristics affect
all
country that has low
the incorne o
Furthermore, the problem of raising ^j^g^ploited natural
labour, u
capital per head, much unemployed
resources (Pakistan, say) is likely to be veiy ' ' '
'

underpopula
the income of a country like Australia,
w i
 818 ECONOMIC GROWTH AND DEVELOPMENT

many unexploited natural resources In either case, the problems will be

more difficult if the country has a religious system that places a low value on
economic activity and savings Instead of trying to define the problem of
underdevelopment, we should look for some common problems in vanous
groups of countries and be prepared to find major differences both the m
problems and the solutions m
whatever group of countnes we study

FOSTERING GROWTH PLANNING OR LAISSE2 FAIRE?

Once an underdeveloped country decide to try to raise
the authorities of
the growth rate, they will immediately face a fundamental question how
much government control over the economy is necessary and desirable^
Practically every shade of opinion from ‘the only way to grow is to get the
government s dead hand out of everything’ to ‘the only way to grow is to
get a fully planned centrally controlled economy’ has been seriously advo-
cated Such extreme views are easily refuted by factual evidence Many
economies have grown with very little government assistance Perhaps the
United Kingdom and Holland are the best examples here Others, such as
modern Russia, Poland and China, have shown sustained growth with a
high degree of centralised control of the economy Considering other coun-
tries, we find almost every conceivable mix of government and private

initiative m the growth process Government fostering of transportation

systems is very common, to take but one example of a typical form of
government intervention designed to encourage growth
What sense can we make of these apparently conflicting historical prece-
dents’ Probably the only satisfactory answer is that the appropnate action
depends on the circumstances presently ruling m the country In some
^cases, ineffective governments may have been interfering with the economy
the point of discouraging pnvatc initiative, m which case growth may
ellbe enhanced by a reduction m
government control over the economy
‘

other cases where major quantities of social capital arc needed or where
existing institutional arrangements such as land tenure are harmful to
growth active mtenention by the central authorities may be essential to
encourage growth * There are many possible mixes between state and
private initiative that have been used successfully at various times and
places On the question of what is the best mix at a particular time and in a
particular place, there is likely to be much disagreement
One of the reasons why governments wish to intervene is to produce a
higher level of savings than ivouJd ensue it saving decisions ivere left solely

! This was Adam Smith s view of what was required in Bntain in 1 776
2 This was many peoples view of what was required in Russia on the eve of the 1917
revolution
 UNDERDEVELOPED ECONOMIES 819

to private indiwduals. In a fully employed frce-tnarket economy, invest-

ment is limited by the quantity’ of sa\’ings households and firms will volun-
tarilyengage in. INTien firing standards are low, sarings are likely to be lo’.v.
Central governments can intervene and force the public to 5a\ c at a much
higher rate than they othensise ivould. This compulsoiy- saring has been
one of the main aims of most of the ‘Plans' of Communist govemmenLs,
Those ^s’ho believe in the indiridual’s freedom of choice may balk at this
compulsory' sacrifice of the firing standards of present generations for the
benefit of future ones. But presented iritb this argument, the planner would
reply that, %rith complete freedom of individual choice, growth would be
slow or nonexistent, and he tvould then go on to ask why the present genera-
tion should be allowed, by virtue of their sarings decisions, the power to
inflicta low living standard on all future generations. One of die goals of
the five-year plans of Russia, Poland and now Cliina is to raise sarings, and
thus to lower current consumption below what it would be, gi%-en complete
freedom of choice. The ulrimate goal is to make future generations better
off than they would be if they inherited only the stock of capital that would
be voluntarily left to them by present generations.
818 ECONOMIC OROWTII AND DEVELOPMENT

many unexploited natural resources In either case, the problems will be

moredifTicuIt if the country has i religious system that places a low value on
economic activity and savings Instead of trying to define the problem of
underdevelopment, we should look for some common problems m various
groups of countries and be prepared to find major differences m both the
problems and the solutions in whatever group of countries we study

FOSTERING GROWTH PLANNING OR LAISSEZ FAIRE?

Once the authorities of an underdeveloped country decide to try to raise
the growth rate they will immediately face a fundamental question how
much government control over the economy is necessary and desirable’
Pnctically every shade of opinion from the onl> way to grow is to get the
government s dead hand out of everything to *the only way to grow is to
get a fully planned centrally controlled economy’ has been seriously advo
cated Such extreme views are casilv refuted by factual evidence Many
economies have grown with very little government assistance Perhaps the
United Kingdom and Holland arc the best examples here Othen such as
modern Russia, Poland and China have shown sustained growth with a
high degree of centralised control of the economy Considenng other coum
tnes, we find almost every conceivable mix of government and private
growth process Government fostenng of transportation
initiative in the
systemsis very common to take but one example of a typical form of
government intervention designed to encourage growth
U hat sense can we make of these apparently conflicting hisioncal precc
dents’ Probably the only satisfactory answer is that the appropriate action
depends on the circumstances presently ruling in the country In some
cases ineffective governments may have been economy
interfering with the
to the point of discouraging pnvaic initiative, m which case growth may
well be enhanced by a reduction m
government control over the economy *
In other cases where major quantities of social capital arc needed or where
existing institutional arrangements such as land tenure are harmful to
growth, active intervention by the central authontics may be essential to
encourage growth ’ There are many possible mixes between state and
private initiative that have been used successfully at vinous times and
places On the question of what is the best mix at a particular time and in a
particular place there is likely to be much disagreement
One of the reasons why governments wish to intervene is to produce a
higher level of savings than would ensue if saving decisions were left solely

1 Thu was Adam Smith t \i«w of **hai wai required in Bricain in 1776
2 Thu was many people i view of what war required in Russia on ihe eve of the 1917
 UNDERDEVELOPED ECONOMIES 819

to private individuals. In a fully employed free-market economy, invest-

ment is limited by the quantity of savings households and firms will volun-
tarilyengage in. When li\'ing standards are low, savings are likely to be low.
Central governments can intervene and force the public to save at a much
higher rate than they otherwise would. This compulsory saving has been
one of the main aims of most of the ‘Plans’ of Communist governments.
Those who beheve in the indi\adual’s freedom of choice may balk at this
compulsory sacrifice of the living standards of present generations for the
benefit of future ones. But presented with this argument, the planner would
reply that, wth complete freedom of individual choice, grois'th would be
slow or nonexistent, and he would then go on to ask why the present genera-
tion should be allowed, by virtue of their savings decisions, the power to
inflict a low living standard on all future generations. One of the goals of

the five-year plans of Russia, Poland and now China is to raise savings, and
thus to lower current consumption below what it would be, given complete
freedom of choice. The ultimate goal is to make future generations better
off than they would be if they inherited only the stock of capital that would
be voluntarily left to them by present generations.

THE MEANS OF GROWTH

How can an underdeveloped country raise its growth rate ?

Education; The efficiency of labour of all kinds can be raised greatly by

education. As modem techniques are introduced, a large rise in the educa-
tional standards of the work force is necessary. A man who cannot read or
write or do simple calculations will be much less efficient in many jobs than
one who can. A
manager who knows something of modem methods of
book-keeping, inventory control, and personnel management is likely to
be much more effective in getting the most output from a given input than
one who is ignorant of these techniques.

Health A : large increase in productivity can be achieved by raising the

health standards of the labour force. ‘ Less time is lost and more effective
effort is put in during the hours that are worked when the labour force is
healthy than when it is not.

Natural resources: Natural resources are, of course, important for

growth. A country that has a large supply of easily developed resources will
find growth easier than one that has fewer and less accessible resources.
Developing such resources as are available is a means of fostering growth.

1 But health gains also lower death rates and leads to population growth. In the short run,

this makes growth more difficult, as we shall see.

 UNDERDEVELOPED ECONOMIES 821

But, at the present, the accumulated amount of productive capital is small.

And the supply of funds for capital investments is much too small to do all
the things required. It may take as much as ,{^1,000 of capital to increase
national income by 100 per year. If this
will take 100,000 million
is so, it

of capital to raise average income per year by ^^100 in a country of 100

million people. The shortage of investment funds is without doubt one of the
most serious bottlenecks on the road to development.

The vicious circle of poverty; If capital is to be created at home by

a country’s own efforts, it is necessary to divert resources from the produc-
tion of goods for current consumption. This requires a cut in present living
standards. If living standards are already virtually at the starvation level,
then such a diversion will be difficult and at best only a small proportion of
resources can be re-allocated to the production of capital goods. Such a
situation is often described as the vicious circle of poverty : because a coun-
try has little capital per head, it is poor; because it is poor, it can devote
only a few resources to creating new capital rather than producing goods
for immediate consumption; because little new capital can be produced,
capital per head remains low; because capital per head remains low, the
country remains poor. Notable examples of countries that have broken this
vicious circle by their own efforts are Britain and the Soviet Union Britain —
slowly over centuries and Russia rapidly over decades. The fact that it is

possible to break the circle does not, of course, mean that it is not a very
serious problem.

Growth through imported capital: Another way of accumulating

the capital needed for growth borrow it from abroad. If a poor coun-
is to
try, A, borrows from a rich country, B, it can use the borrowed funds to
purchase capital goods produced in B. Country A thus accumulates capital
and does not need to cut its current output of consumption goods. When
the new capital begins to add to current production, it may be possible to
pay the interest on the loan and also to begin to repay the principal out of
the increase in output. Thus, income can be raised immediately, and the
sacrifice postponed until later, when part of the increased income that
might have been used to raise domestic consumption is used to pay off the
loan. This the great advantage of allowing a poor country to
method has
have an initial increase in capital goods far greater than it could possibly
have created by diverting its own resources from consumption industries.
Examples of countries that have grown rapidly, largely at first through the
help of foreign capital, are the United States (British capital) and Canada
(US and British capital).
Many underdeveloped countries are, however, suspicious of foreign
 822 ECONOMIC GROWTH AND DEVELOPMENT

capital, for of them fear that the foreign investor will gam control
many
over their industries or their government The extent of foreign control
depends on the form that foreign capital takes If the foreignen buy bonds
in domestic companies, they do not own or control anything, if they buy
common stocks, then they do own part or all of the company , if they subsi*
disc agovernment, they may feel justihed in exacting political commit-
ments Whether or not foreign ownership of one's mdustnes carries political
disadvantages is a subject of debate ' The economic advantages arc, how-
ever, quite clear Accumulating a given amount of capital by domesUe
saving leads to a different time path of domestic living standards from
accumulating it by foreign borrowing The domestic method requires
greater current sacnficc, but pap a higher return later, foreign financing
requires no present sacrifice,* but produces lower gains m living standards
later
Getting foreign capita! is easier said than
done in the early stages of
development America and Canada were underdeveloped in the sense of
being underpopulated and having many unused resources, but they were
latent giants and held promise of nch returns to foreign investors It is
harder to see similar investment opportunities in, say, Pakistan, where
merpopulation has been a problem for ctntunes and where the soil is
severely damaged by centunes of imgauon without proper drainage The
ability of such a country to borrow from private sources is small Foreign
capital IS playing a role, but it is capital provided by foreign governments
and international agencies, not by private investors

Growth through contributed cabital Investment funds for de-

velopment are being receded today by underdeveloped countnes from the
governments of the more developed countries acting both unilaterally (as in
the US Agency for International Development and in a similar Soviet
programme) and through international agencies, such as the International
Bank for Reconstruction and Development and the Export-Import Bank
Why do the developed countnes give to the underdeveloped ones’ Some
say for political gam, some say for economic gam, and some say from com-
passion Possibly all are correct in part

} PonT/CAL There is httie doubt that m

the Cold IVar the struggle for
the political allegiance of the uncommitted countries is intense Since most
of the uncommitted countries are underdeveloped and urgently desire to

1 In Canada for example there has been much debate over the political effects of hxMng
much of Canadian industry owned by US
nationals who are presumably more open to
pressure from US central authorities than
from Canadian ones
2 See the discussion in the Appendix to Qiapier 53, pages 7&6-7, for a description of
what happens when capital is borrowed from abroad
 UNDERDEVELOPED ECONOMIES 823

grow, the political system that gives the best promise of growth is likely to
have the strongest appeal. The prospects for the vast majority of the world’s
population of obtaining any living standard above bare subsistence depends
critically on the possibility of inducing quite rapid economic growth there
;

is no doubt that political and social systems will be judged to a great extent

by their success in inducing this growth.

2 Economic: Does a country that gives aid to underdeveloped countries

thereby increase its own living standards? Or does it have an economic

interest in keeping the poor countries poor so that they can be more elfec-
tively exploited ? The theory of international trade leads us to predict that
there will be a gain if the poorer countries get richer, for as incomes rise in
poorer countries the market for our exports will grow. By and large, the
volume of international trade tends to grow - in spite of tariffs and other
forms of interference - as countries get richer. The larger the productive
capacity of the world, the greater the scope for the operation of the inter-
national division of labour, and the larger the gains from trade.
On the other hand, growth can hurt advanced countries with special
positions to protect. A country with a monopoly in the production of some
commodity can suffer when competitors arise in other countries. Also, firms
from advanced countries that operate in underdeveloped ones often have a
lot to lose from growth. The United Fruit Company of America would
undoubtedly lose profits if successful industrial development in Nicaragua
so raised the earnings of agricultural labour that the real wage paid to
workers on banana plantations doubled. There is, of course, a big difference
between having a reason to fear development and actively trying to prevent
it. At any rate, this discussion is sufficient to show that there is not a clear

case that developed nations always gain when underdeveloped countries

grow.

3 Compassionate: There no doubt that part of the interest of people

is

in developed countries in the fate of people in underdeveloped countries

stems from genuine altruism. We believe that a minimum standard of living
is necessary to preserve human dignity and to allow development of the
human personality. We are disturbed when we see the citizens of whole
nations falling below this standard. A few at least do not sit idly by; they

try to do something about it.

Perhaps the best evidence that this motive exists is the success of voluntary
appeals for food, funds, and clothing for persons in stricken areas of the
world. Although this is not a new concern, as we have grown richer, the

sizeof our contributions has grown. It is a policy of the governments of

most advanced Western countries and one more or less accepted by most
824 ECONOMIC GROWTH AND DEVELOPMENT

political parties that we devote some of our resources to alleviating poverty

throughout the world

POPULATION AND ECONOMIC DEVELOPMENT

Capital IS a necessary but not a sufficient condition for the economic growth

of an underdeveloped country An underdeveloped country must accumu

late enough capital, and of the right kinds, to break the vicious circle of
poverty Once the country does break the circle, it will generate further
investment funds from the savings that its population makes from its own
growing incomes, and it may also begin to attract private capital from
foreign investors
But first the country must begin to grow Many of the countries of the
world have a standard of living no higher than they had a hundred or
even a thousand years ago They have more total output, but they also
have more mouths to feed The average Vietnamese ts as hungry as his
great-grandfather was You might think that even modest gams m the size
of the capital stock would fsertludly add up to enough to get a sustained
growth going But this is not necessarily so since it is the amount of capital
per person that determines whether living standards will rise, and we must
thus be concerned about the growth rate per capita Population growth is a
central problem of economic development If population expands as fast
as does total output, then per capita income will not increase If population
does expand rapidly, a country may make a great effort to raise the quantity
of capital only to find that a corresponding rise m population has occurred,
so that the net effect of its growth policy’ is that a larger population is now
maintained at the original low standard of living An increase in living
standards occurs only if capital, and possibly other resources, expand faster
than the increase in population This explains the great importance
attached to population policy m many underdeveloped countries
This population problem has led economists to talk about the ‘critical

minimum that is required not merely to increase capital, but to

effort
increase itenough so that the increase m output outraces the increase
fast

in population ‘ The problem anses because population size is not indepen-

dent of the level of income If population control is left to nature, nature
solves It m a cruel way population increases until many are forced to live
at a subsistence level, further population growth is halted by famine,
pestilence, and plague In some ways, the population problem is more

I Hiivty Leibtnsitin w the jtuthoi of the phiase 'cntical rninimum effort See hw
provocative book Economic Backwardness and Ecanatme GsowVi Qohn Wiley 1957) (Also avail
able m a Science Editions paperback ) WW
Rostow has used the phrase take ofT to express
a very similai notion
 UNDERDEVELOPED ECONOMIES 825

severe today than was even a generation ago. Advances in medicine and
it

in public health have brought sharp and sudden decreases in death rates. ^
In Mexico today the birth rate is more than three times as high as the
death rate, and population is growing at more than 3 per cent per year.
A rise in production of 3 per cent per year is required for Mexico to ‘break
even’.
There are only two possible -ways to solve this problem. One is to make
such a massive push that we achieve a growth rate well in excess of the
rate of population growth. The second is to control population growth. The
problem can be solved by restricting population growth. This is not a matter
of serious debate, although the means of restricting it are, for there are
religious considerations involved. Positiveeconomics does not decide such
issues, can describe the consequences of any choice that is made. For
but it

example both Sweden and Venezuela have death rates of about 10 per
1,000 population per year. The birth rate in Sweden is 14; in Venezuela it
is 45. Thus the net increase of population per year
is 35 per 1,000 (3-5 per

cent) in Venezuela, but only 4 per 1,000 (-4 per cent) in Sweden. If each
country achieved an over-all rate of growth of production of 3 per cent per
year, Sweden would be
increasing her living standards by 2-6 per cent per
year, while Venezuela would be lowering hers by -5 of 1 per cent per
year. Today, Sweden’s standard of living is more than twice as high as
Venezuela’s.^ The gap will widen rapidly, if present population trends
continue.

BALANCED OR UNBALANCED GROWTH?

Should a developing country push expansion in all sectors of its economy
(balanced growth), or should it specialise in certain sectors (unbalanced
growth) ? The decision is an important one, and the government is surely
in a position to affect it, either because it is directing the channel of invest-
ment or because it uses its tariff and taxing power to influence the allocation
of funds.
The theory of comparative advantage provides the traditional case for

1 It is ironic that much of our compassion for the poor and underprivileged people of the

world has traditionally taken the form of improving their health, thereby doing little to avert
their poverty. We laud the medical missionaries who brought modem medicine to the savages,
but the elimination of malaria, though surely a boon to the world, has doubled the rate of
population growth in Ceylon. No one would argue against controlling disease, but we must
recognize that other things must also be done if the child who survives the infectious illnesses

of infancy is not to die by starvation in early adulthood.

2 And Venezuela is the wealthiest country in South or Central America at the present time.
It has an average income per person of about S700. (This compares with the US figure of about
S3,000, and with the figure for India of about SIOO.)
826 ECONOMIC GROWTH AND DEVELOPMENT

the desirability of unbalanced growth In reaching a decision, however, the

central authonties will also take the following considerations into account

1 The greater the degree of imbalance m growth, the greater will be

the country’s dependence upon foreign trade Too much specialisation
means putting all one’s eggs into one basket and makes the economy vulner-

able to fluctuations in world
It IS specialised and
the product obsolete
2 Specialisation
demand and supply for the product in which
also vulnerable to technological changes that may render
is going to reap the gains from trade
necessary if one is

that accrue when countries specialise m

the range of products which they m
have a comparative advantage Balanced growth pursued to the extreme
of self-sufficiency is then likely to result in a lower living standard than will

result from some degree of specialisation

3 Countries often push certain lines of production cither for prestige
purposes or because of a confusion between cause and effect Because most
wealthy nations have a steel industry, it does not follow that having a steel

industry will make a nation wealthy Indeed, if one has a serious compara
live disadvantage m steel, then having a steel industry will make one poor
Whether or not one really gams international prestige by having an un
economic steel industry or national airline is difficult to ascertain It is very
probable, however that in the long run, prestige goes to the country that
grows nch rather than to the one that stays poor but that produces at high
cost a few prestige commodities that are regarded as signs of wealth
4 In deciding what sectors of the economy to push, due regard must
be paid to potential future comparative advantages as well as to current
ones Many skills can be acquired, and the fostering of an apparently
uneconomic domestic industry may, by changing the characteristics of the
labour force, develop a comparative advantage m
that line of production
Where there is excessive concentration on current comparative advantages,
one consequence may be an excessive defence of the jiafiu quo in the pattern
of international specialisation
 PART 11

MACRO-ECONOMIC
POLICY
CHAPTER 58

MACRO-ECONOMIC POLICY

Economists still debate the causes of the Great Depression, just as historians

stilldebate the causes of the First World War. But the ovenvhelming
majority of economists agree that another episode like the Great Depression
of the 1930’s need not happen again. Indeed they feel that the more modest
episode in the US when unemployment rates stayed as high as 5 to 6 per
cent between 1958 and 1964 need not have happened. The rapid fall in
unemployment that occurred in 1964 after the US Congress finally heeded
the economists’ advice suggests that the economists are right in this
judgment. If so this is an impressive bit of exadence in favour of economic
analysis, because means that we do believe we understand the workings
it

of the economy well enough for it to be effectively controlled so as to avoid

such extremes of economic performance as depressions and violent inflations.
In this chapter we are concerned with macro-economic policy. Experience
with past policies not only constitutes our economic history, but also pro-
vides the material for testing our economic theories. Macro theories lead us
to predict how various policies will affect certain aggregate variables. If we
are now confident that we can make some correct diagnoses, it is because
the e\ddence appears to conform, at least in broad outline, with many of
the major predictions of the theories.
Just as the distinction between macro- and micro-economic theory is to
some extent arbitrary, so the distinction between macro- and micro-
economic policy is to some extent arbitrary. Of course, government pohcies
range from those that affect single households and firms to those that affect
the society as a whole. Our concern in this chapter will be with four major
aggregate variables that economic policy seeks to influence: the level of
unemployment, the price level, the balance of payments and the rate of
growth of the economy. In the first part of this chapter we shall consider
each of these policy variables asking why we are concerned about them.
830 ECONOMIC GROWTH AND DEVELOPMENT

and describing their past behawour We then go on to consider possible

conflicts between various economic policies Tins problem leads us to look
more carefully at the means available for controlling these macro variables
and this leads us finally to a discussion of some of the persistent policy
conflicts that have harassed government policy-makers since the Second
^Vorld \Var

Unemployment
^VHAT IS FULL EMPLOYMENT^ Wc have talked many times m this book
about full employment, hut every time wc have talked about the level of

employment we have spoken of such and such a proportion of the labour

force as being unemployed Surely full employment means what it say's no
unemployment^ The answer to this question is ‘No’ Various causes of
unemployment can be distinguished, and some of them arc regarded as
unavoidable aspects of the functioning of a market system Such unemploy-
ment IS called frictional unemployment One source of fnctional
unemployment is labour turnover People leave jobs for many difTcrent
reasons Some quit, and some are sacked, but almost all of them find new
jobs, though It may take time Since, at every moment of time, there will
be a group of individuals moving from one job to another, there will always
be some proportion of the labour force out of work Of course, if the volume
of fnctional unemployment stays stable over lime, it does not mean that
the same individuals are out of work For example, it has been established
that, in the US, 'between a fourth and a third of all vvorkers change jobs
’ *
either by choice or by force of circumstances in the course of a year
If the workers take an average of two weeks to change jobs, this will mean
that one per cent of the labour force will be frictionally unemployed at any
moment of time The number is possibly smaller in Bntain but it is by no
means insigniflcant This routine turnover is accentuated by economic
growth As growth proceeds, cost conditions and thus input requirements
change, the pattern of demand and thus output requirements change as
well Such changes make it necessary for people to move among occupa-
tions, industries and areas, and often necessitates retraining Thus at any

moment of time there will be an amount of unemployment which is assoa-

aterf with resource re-alTocation anrf wfiicfi cannot 6e avoirferf as fong as
there is economic growth ^ A third source of fnctional unemployment is the

seasonality of some occupations in which year-round employment is not

available for all who wish it ^fan> seasonal workers arc out of a job for
anything from a few weeks to six or seven months and, although some of

1 H S Fames Research on Labor MobUsif New York Social Science Research Council 1954

page 62
2 This problem was considered in Chapter S6 pages 812-13
 MACRO-ECONOMIC POLICY 831

them seek fill-in work, they frequently fail to find it. Finally, about 2-0 per
cent of the labour force every year are new entrants, and it is rare for any-
one to walk out of school and into a job without some delay.
For all these reasons some minimum level of unemployment must always
occur in an economy. Full employment is usually said to occur when this
minimum level is achieved. How large is it ?

In Britain, employment’ was defined in the White Paper on Employ-

‘full

ment Policy published by the Government in 1944 as existing when recorded

unemployment fell to 3 per cent of the labour force. Since the war, monetary
and fiscal policies have been used in an effort to control the level of un-
employment and, in spite of fluctuations, the rate has never risen above
2-6 per cent, and it has several times been below 1-5 per cent. The notion
of the amount of unemployment that is normal and necessary has been
revised downwards as a result of this experience. Most people now feel that
full employment requires no more than T2 per cent to T5 per cent of the

labour force frictionally unemployed. In Sweden, the post-war unemploy-

ment rate has frequently been as low as IT per cent; in Australia, it has
averaged around 1 per cent, never rising above 2 per cent. In the United
States the necessary level of frictional unemployment is generally thought
to be as high as 3 per cent. Since there has not been a prolonged period
with unemployment as low as 3 per cent this view has not really been tested.
Indeed there is no generally accepted economic theory that predicts that
frictional unemployment should be so much higher in the US than in
other countries.

Why are we concerned about unemployment?: There is virtually

unanimous agreement that anything above the necessary level of frictional
unemployment is undesirable. When resources lie idle their potential out-
put is lost forever. Unemployment for any long period of time can be a very
degrading thing, indeed when an able-bodied head of the household finds
himself without work for a period of years the experience can be soul
destroying.^

Past experience: As far back as we have any records, periods of heavy

unemployment have been observed. Until very recently, casual observers
and experts alike tended to believe that there was not very much a govern-
ment could do to affect the level of unemployment, except possibly by
interfering even less than usual in the economy when unemployment was

See pages 465-6 and 782 and the references given on page 62. An extremely interesting
1

effects on individuals of long-term unemployment can be found in Paul Jacobs,

account of the
‘Unemployment as a Way of Life’ in Employment Policy and the Labor Market, A. M. Ross (editor),
University of California Press, 1965.
832 ECONOMIC GROWTH AND DEVELOPMENT

particularly high, so that the natural marhet forces could restore full
employment From 1929 to 1932 both the British and American govern-
ments kept assuring the unemployed that prosperity was just around the
comer Until the I930’s, it was at least possible to maintain the view that
the market, left to itself, would restore full employment, because recessions,
although often sharp, tended to last only a very few^ years (Look again at
Figure 45 1, page 619 ) The Great Depression of the 1930’s dispelled foreter
the belief that the unaided free market would always restore full employ-
ment within an acceptable period of time Those who believed that the free
market did guarantee full employment were left to argue that, although
full employment had not been restored after eight years, the market would

nonetheless restore it ‘in the long run’ To those who waited from 1929 to
1937, for this problematical ‘long run’ to appear,Keynes provided the
‘Maybe you are right, but in the long run,
practical epitaph for their belief
we shall all be dead’^ In the postwar years, the governments of all
Western countnes have come to accept full employment as a goal of policy
as something that governments can achieve through their actions and as
something that they have a responsibility to attempt to achieve In 1944,
the British Government m its employment White Paper accepted a
responsibility to maintain full employment by appropriate changes m its
fiscal and monetary The American Government accepted the same
policies
responsibility in theEmployment Act of 1946, which set up the Council of
Economic Advisors, which was charged with the responsibility of reporting
each year on the state of the economy and advising the government on how
the full-employment goal could best be achieved
Since that time the Bncish unemployment rale has remained un-
precedcntly low while America did experience the period of higher rales
referred to at the beginning of this chapter By and large taking all Western
countries into account the last 20 years has by hisloncal standards been a
period of low unemployment

The Price Level

Why are we concerned about the price level^ Changes in price
levels tend to affect the distribution of income those with fixed incomes
suffer in periods of inflation and gam in periods of deflation The full effects
of changes in price levels have been discussed at many points in this book,
fora review of the effects of inflations and deflations, the student should now
re read pages 174-6
Most people agree that, cetens panbus, a stable price level is desirable
Most agree that rapid inflations and deflations are undesirable Although
there is some consensus that slow changes in the price level are also

I This IS a paraphrase of Keynes famous quote

 MACRO-ECONOMIC POLICY 833

undesirable, there is considerable disagreement as to how undesirable they

are. Some people believe that mild inflations in the order
say, 2 or 3 per of,

cent per year do not cause serious harm, while others believe that any
degree of inflation is serious and that the maintenance of an absolutely

stable price level is an important goal of policy.

Past experience: Practically every form of price-level behaviour has been

observed somewhere at some time. Probably the most dramatic occurrences
are hyper-inflations, in which prices rise by factors of a thousand-fold in a
matter of weeks. One famous example is the hyper-inflation that occurred
in Germany in the early 1920’s.. Most savings were totally destroyed in
value by the sky-rocketing prices, and the consequent disruption of the
middle classes as a social unit did a great deal to upset the social fabric of
Germany and pave the ^vay for the Second World War. It also did a great
deal to persuade many people that inflation was one of the worst imaginable
something that was to be avoided at absolutely any cost. An
social evils,

internationalcomparison of post-war behaviour of price levels is given in

Table 51.1 on page 701.

Balance of Payments
Why are we concerned about the balance of payments?:
Generally, the purpose of international trade is to take advantage of the
international division of labour: we export goods in order that we can
import those goods that are cheaper to obtain abroad than to produce at
home. Flows of long-term capital aside, a large import surplus is regarded as
undesirable, as is a large export surplus; a situation in which imports equal

exports is regarded as desirable. If a country’s present reserves of foreign

exchange are thought to be inadequate - a position in which many countries
find themselvestoday — it will regard a mild export surplus as its balance-of-
payments target, but, once reserves are at an adequate level, it will aim at
an equality between imports and exports.
Even the most enlightened government will not regard an export surplus
as anywhere near as undesirable as an import surplus. An export surplus
means that foreign exchange is accumulating, an accumulation that can go
on more or less indefinitely, as far as the surplus country is concerned. It
does mean that the country is producing without consuming to the extent
of its export surplus. Although this depresses living standards below what
they could be, it is not a cause of any crisis.Of course, some other country
must be suffering an import surplus, and that country may be forced to take
steps to remove it. An import surplus can go on only as long as foreign-
exchange reserves last. A monthly import surplus equal to, say, 5 per cent
of the total exchange reserves can last only for 20 months, and, long before

27
834 ECONOMIC GROWTH AND DEVELOPMENT

that, It will cause a speculative flight of capital m expectation of a devalua-

tion of the currency An import surplus is generally the occasion for fairly
rapid preventatne actions, unless there are reasons to believe that it is the
result of genumcl) temporary causes so that it will disappear on its own
before very long

Past experience Wc studied past balance-of-payments experience in

some detail in Chapter 55 In the past, under the gold standard, govern-
ments usually refrained from interfering with the balance of payments
Since the general advent of fixed exchange rates under a paper currency
standard (see pages 672 3), however, the central aiitliontics have had no
choice but to take a major measure of responsibility for the balance of
payments

Growth
Whv are we concerned about growth rates ^ By and large,
economic growth is It is the major cause of changes
accepted as desirable
m living standards With growth, each generation can expect, on the
average, to be substantially better off than all preceding generations The
horrors of the early mdustnal revolution to which wc alluded earlier (see
pages 460 2) are no longer with us to a great extent because economic
growth has removed the necessity of 14-hour days worked in animal like
conditions Growth, as we have seen (see page 807), is not without its costs,
and there is a limit beyond which further increases m the rate of growth
would not be regarded as desirable But that limit is probably quite high,
at the moment an increase in the rate of growth would probably be
generally accepted as desirable in almost all Western countries Certainly
a fall could be very senous If even the existing dilTcrcntial in growth rates
between Germany and the UK
should persist for a century Bntain will
seem to the German visitor of 2066 as backward as Sicily and southern
Spain now seem to the British vasitor'

Past experience Some discussion of differing growth rates was given

on pages 802-4 (see especially Table 56 2) Until recently concern in
Britain over post-war growth rates relative to the counlnes of Western
Europe had suggested that this was a new problem for Britain The tacts

are exactly the reverse By historical standards, the post-war growth rate
in Bntam has been high Throughout the nineteenth century the British
growth rate averaged between 1| and 2 per cent per annum Since the
Second World IVar the growth ntc has been between 2 and 3 per cent per
annum In fact, in quantitatively-recorded history, the British economy
has never before grown as rapidly as at present
Poor performance by international standards is not a new phenomenon
 MACRO-ECONOMIC POLICY 835

By 1914 German growth rate was very much higher than the British one.
the
W ere not for the political idiocies of two World Wars, Germany would
it

have gained complete ascendency over Britain in the normal course of

peace-time events provided that the relative growth rate of 1914 had con-
tinued to 1960. This trend was interrupted by 30 years of chaos begun and
ended by the two World Wars. But it has now reasserted itself and if it
should continue the standard of living of the average German and French
household will be very much higher than that of the average British house-
hold ^vell before the end of this century.

Prestige

An introductory textbook is not the place in which to say much about this
goal of policy. It is necessary, however, to mention that a great deal of
post-war British economic policy cannot be fully understood unless this goal
is recognised. In a full study of British policy the search for international

prestige would loom large

an explanation of much otherwise incompre-
as
hensible behaviour. and as it now seems premature, venture into
The early,
atomic power stations; the maintenance of the aircraft industry after it
became clear that the British industry was unable to achieve the economies
of scale available to the Americans and in the face of the fact that this
unprofitable industry was using the services of a high proportion of Britain’s
scarce engineering talent; the continuance of the Concord project in the
face of early financial esdmates that were obviously unrealistic; the
succession of all too predictable failures to develop rockets and other
weapons in the face of enormous development costs and the obvious
economic superiority of foreign products; the desire to duplicate the US
nuclear deterrent at the cost of a conventional mobile British force that
could have supplemented rather than duplicated the American nuclear
force ; the all too transparent desire to gain entry into the ‘nuclear club’ so
as tobe included in the top councils of the world; the desire to maintain
pound’ — all of these and many other decisions in the
the ‘prestige of the
economic and the political sphere are only made comprehensible by the
motive of trying to obtain international prestige.
In an introductory textbook it is appropriate to sidestep this problem as
much as is possible. It is necessary to note, as we have already done, that
much of the behaviour of the central authorities in Britain and in other
countries cannot be understood unless the search for prestige is given
recognition as one of the major goals governing policy decisions.^

1 Having raised the issue I would like to suggest the following two personal observations on
it. (1) It seems to me that prestige in fact goes to those who are successful rather than to those
who seek to have a few goods and services that are regarded as the manifestation of success.

Thus West Germany and France enjoy international prestige because they have in fact raised
836 ECONOMIC GROWTH AND DFVELOPMENT

CAUSES OF AND CONTROL OF CHANGES IN

MACRO VARIABLES
The causes of uncmplo>ment, balancc-of-paymcnts problems and
inflation,

economic growth have been the subject of Parts 7 to 9 of this book In

the present section wc shall merely summarise this cvrlicr discussion and
then consider how variations in these magnitudes can be controlled by the
central authorities
In order to clanf> the divcuvsion vve need to distinguish three kinds of
variables First, we liavc poLtev variadlfs vvliich arc those variables in

which the policy maker is ultimately mterestetl In the present context there
arc (ignoring prestige) four key policy varnbics unemployment (f/), the
price level [P], the balance of payments (li) and the growth rate (G)
Second, we have instrumfntal varia&lfs These arc the variables on
which our policies can act directly They include such things as the size of
the gov emmcnis budget deficit or surplus, the quantity of money and laws
of all sorts Iletvsccn these two we may have a link errated by many van-
.iblcs which wc call intfrmediate VARiAtitrs These arc variables that

we cannot affect directly and m v^hose beJnviour wc are not directly

interested except in so far as they in turn affect the behav lour of our policy
vanables The great value of economic theory to the policy maker is the
link thatIt provides, through intermediate v.anables, between the instru*

mental variables the behaviour of v%hich he can change and the policy
vanables the behuiour of which be wants to change In a simple example
a change in the instrumental vanable, govemmeiu spending, affects an
intermediate vanable, aggregate demand, which m luni alTects the policy
vanable of unemployment Using arrows to indicate a causal sequence we
cun show this link in 1 able 58 1

ihfir Iivinfi iundards rapidly Britain mjoyi ten pmti(r brcauie hrr rale of firowth hat b«n
dower and ilie pouruion of atomic (ubmannes does nothing lo mitigate thii (2) The whole
search for iniernational prniige through military power iremi a \rry odd one If any of ui
encountered two boyi one whom wai phy-tically smalt but who was w-nting Beethoim l>Ve
symphonies painting Picasso-IiLe pKiiires and insmting improved jet engines while the other
bd was doing nothmg but parading bi» physical swetigdi which he had inherited by virtue of
being bom bigger ihan the hm boy wc would have no hesitation in giving more prestige*
creating attention to the first bov rather than to the second If the Tint boy said he was going

lo take lime oft writing music painting piciurrs and inventing in order to lake a physical
fitness course so ihat he rouid gam prestige in our eves by being able lo beat up boy number 2,

we would have no hesilaiion in condemning bis drciiionand saying he would lose prestige m
our eves by showing us that he valued brute forte over genuine creative activity Since wc
would alt have no doubt as to how to assess the boy i behaviour, how can we acquiesce and
even utter encouraging noises when time aFier time our central authorities behave just as this
hvpoihelical boy v,as proposing lo behave*
But this IS a mailer on which people disagree and this footnote represents a personal opinion
that cannot claim to be bieked by the authonty of established economic theory
 MACRO-ECONOMIC POLICY 837

Table 58.1
THE LINK BETWEEN GOVERNMENTAL ACTION
AND THE ULTIMATE GOALS OF POLICY
Any number of
f An instrumental
f Government^ (The policy^
\policy intermediate
/ ^variable \ variable /
variables

Table 58.2 gives a particular example

Table 58.2
THE LINK BETWEEN GOVERNMENTAL ACTION
AND THE LEVEL OF UNEMPLOYMENT
le\el of'|
J Government jThe size of thel
\fiscal policy J tbudget deficit lemplo)-ment /
J

With these additional distinctions to aid us we can pass on to consider

behaviour of each of the policy variables.

Unemployment
Causes: The causes of changes in unemployment (other than frictional
ones) can be grouped into two main classes changes in aggregate expendi-
:

ture and structural changes in the economy. The first cause was one of the
major subjects of the whole of Part 7 (see especially Chapter 42) while the
second was discussed in Chapter 56. A change in aggregate expenditure is
caused either by a change in any one of the autonomous components of
aggregate expenditure or by a change in the division of total disposable
household income between savings and consumption expenditure. A
variation in the rate of structural change in the economy would be
associated ivith a change in the over-all rate of economic groivth or some
structural change pattern of economic growth.

Control: The only important intermediate variable through which the

level of employment can be effectively controlled is the level of aggregate
expenditure. If total spending is low, the output of firms will be loiv and the

level of employment will be low. If total spending is high, the level of output
wall be high and the level of employment -will be high as well. If the level
of employment is low there is no effective way knoim at present for the
central authorities to raise it without raising aggregate expenditure.
Therefore, if the central authorities are going to try to achieve full employ-
ment, they have no option but to take steps to vary the volume of aggregate
expenditure.
JiiS ECONOMIC GROWTH AND DEVELOPMENT

There are two major sets of instrumental vinablcs that can be used to

influence aggregate expenditure The first set is composed of all the tools
I f fisc il which uerc studied in Chapter 47 The second set is com
poIic>
prsed of all the tools of monetary policy which were discussed in Chapter 49
Ch inges in either of these sets of variables arc likely to affect policy van
ihles other than unemployment (possibly through a complex chain of

II icrmcdiale variables) A debate exists about the relative merits of

moiutars and fiscal policy This debate concerns to some extent the
( fhctiveiiess of each of these policies m influencing aggregate demand but
niULh more importantly it concerns the efrect of changes in the monetary
\ambles on policy variables other than the level
itid fiscal inslrumental of
unimploymcnt The control of unemployment is illustrated m Table 58 3

Tarlf 58 3

THF CONTROL OF UNEMPLO\ MENT

Aggregate _
< expenditure U

"K? Other ntermedidte <^7 Other policy

7 var aWes ••^7 variables

Changes tn the Price Level

< M IS In Chiptcr 51 vve studied the theory of the price level The
v.ri]trilly accepted theory links changes m the price level to changes m
tggreg itr demand and thence to changes in the demand for, or the supply
ol money There is another theory.liowcver, thatignoresaggregatedemand
ind links price level changes to the pressure on wages exerted by unions
The controversy between this tost push theory and the more orthodox
dmimd pull theory was the subject of the Appendix to Chapter 51 This
lontrovcrsy is extremely important because if the demand pull theory is

torrect then control of aggregate demand is the only method of controlling

inflation while if the cost push theory is correct other methods of control
may prove possible
In essence the demand pull theory says that changes price levels are m
lo be accounted for by disequiiibna in markets such disequilibria usually
being caused by changes in aggregate demand A nsc aggregate demand m
in a situation of nearly full employment will create excess demand m many
individual markets and prices will be bid upward The m demand for
rise
 MACRO-ECONOMIC POLICY 839

goods and sendees will cause a rise in demand for factors, and their prices
will be bid upward as well. Thus, says the demand-pull school, inflation in
the prices both of consumer goods and of factors of production is caused by
a aggregate demand.
rise in

To put the matter crudely, the cost-push theorists assert that unions
have the power to obtain significantly large increases in wages, irrespective
of the strength of aggregate demand. This union power is exercised in
varying degrees from year to year in response to such signals as the
popularity of the unions or the relation between unions and the govern-
ment. ^ When this arbitrary power is exercised, wages rise and, since wages
are a large part of the total costs of the firm, prices of commodities rise as
well.

Control of price-level changes; Inflations and deflations can be

controlled by controlling aggregate demand. A rise in prices can be stopped
by a sufficient reduction in the level of aggregate demand. The main instru-
mental variable through which the central authorities ivork is control of
the money
supply. If the money supply is contracted sufficiently any
inflationcan be brought to a halt. Other undesirable consequences might
ensue but there can be little doubt that if control of inflation is desired to
the exclusion of all else, it can be achieved by a sufficiently drastic control
of the money supply - and through that on the level of aggregate •

expenditure.
Other methods of control are possible only if the orthodox demand-pull
theory is not correct. There is still controversy over the competing theories

of cost-push and. demand-pull and in order to study the control of inflation

by means other than the control of aggregate demand, we shall consider
some attempts that have actually been made to do so under the inspiration
of the cost-push theory.
Peace-time interventions by the central authorities in an attempt to
control the price' level are referred to as wage-price policies or as incomes
policies,and they have been tried in most Western countries since the war
with varying degrees of seriousness. Believers in the demand-pull and the
cost-push theories agree that it is possible, given sufficient controls, to slow
down or to stop an inflation at least temporarily. But the kinds of controls
necessary' and the severity with which they must be applied varies with
the theory accepted.
One kind of policy may be called ‘exhortation’ ;
on appeals by the
it relies

central authorities for moderation in setting prices and wages. In America,

1 It is argued that if unions exercise too much
arbitrary power too often, they will lose
public support, and government and regulation will become inevitable. This
intervention
qualification is necessary to explain why the strength of their upward push on wages is
observed to vary from year to year.
840 ECONOMIC GROWTH AND DEVELOPMENT

the President’s price and wage guidelines represented such an attempt The
idea that prices can be controUed merely by publishing what the central
authorities would happen represents an extreme view of the
like to see

economic process It makes most sense if one is a cost push theorist, because,
since the inflation is then assumed to be caused by the exercise of arbitrary
power on the part of a few industrial or union leaders, it may be possible
to penuade them not to exercise their powers In Britain in 1949 a wage
restraint policy was initiated by the government with the full cooperation
of the unions The rise in wages was very much less than any rise that
occurred before or since in the face of the same sort of market conditions
There is little doubt that the policy succeeded temporarily When co-
operation with the unions broke down, higher than normal wage increases
occurred, so that by 1951 or 1952, the level of wages was just where one
would have expected it to be, given the market conditions that existed over
the whole period, but in the absence of a wage restraint policy ‘
An attempt
to hold prices down by and propaganda and executive power repre
talk
sents a belief that the whole process of the reaction of prices and quantities
to demand and supply, which we have been studying throughout this book,
IS so flimsy that a few words from a British Cabinet Minister or American

President can upset the process and cause the economy to stop functioning
in the way that it has functioned for a very long time in the past To those
who accept the demand pull theory this seems very unlikely When King
Canute commanded the tides to hah, he discovered the limits of his own
power and this is the fate demand pull theorists see for a policy of exhorta
tion Of course, active cooperation on the part of price - and wage -
setters can hold back inflation of excess demand for a while
in the face
Bui, according lo the demand can be only a temporary
pull iheonsis, this
rearguard action, and eventually wages and prices will nse by one means or
another
In the 1950 s two attempts were made in Britain to control inflation by
controlling wage increases These experiments reflected the current accept
ance by practically everyone, except some professional economists, of the
cost-push theory of inflation In a cost-push situation, control of wages
would be sufficient to control inflation, but the problem of how to control
•m lx iR/crtff j VfHS TitTi •jVv’wwi Ti'irt: •utrrwft "ir/i iK/i. "AT/y

cooperate in the attempt and the policy broke down, its only significant
effects were to get wages in the public sector (whose control was possible)
seriously out of line with wages in the private sector (whose control was

1 Using this as a test is complicated by the fact that the large wage increases coincided with
the abnormal market conditions caused by the Korean War But Swedish expenence is
important here since they had similar results to Britain m the absence of the upsetting effects
of an external war
 MACRO-ECONOMIC POLICY 841

impossible) . The and a minimum of effective acdon

belief that a lot of talk
by a could seriously change the operation of the economic
feiv officials
system per\'aded the whole experiment.'
The most serious attempts at wage-price stabilisation policies have been
made in some countries of Western Europe, particularly the Netherlands
and the Scandinavian countries. Their policies have been pursued with
more vigour than have those of the English-speaking countries and generally
with more active cooperation from the unions. Even these policies have been
far from unqualified successes. There is some disagreement about just what,
if any, real success they have had. The main problem that they revealed

arises from a phenomenon called ivage drift. Briefly, this is based on the
fact that it is much easier to control negotiated wage rates than it is to
control the earnings of labour. The wage rate is the amount a worker
gets per hour, and earnings are the total amount he gets per week. To see
the importance of this distinction assume, for example, that the officials
operating the incomes policy might decide that output per man-hour has
gone up 5 per cent this year, and that they will allow a rise in wage rates of

5 per cent in order to keep the increase in purchasing power in line with the
increase in output. The average wage is then raised from, say, lOr to
IOj 6d per hour. But, if labour is scarce, employers may be bidding against
one another both to attract new labour and to hold on to their existing
labour. If they are unable to do this by raising wage rates, they can offer
other inducements - such as bonuses and guaranteed overtime pay (whether
or not the overtime is worked). If, by these devices, they can raise average
earnings from, say to per week, then the rise in earnings will
greatly exceed the rise in output, and inflation will occur in spite of the
successful control over wage rates. This tendency for earnings to folloiv

aggregate demand, even though rates do not, is called the wage drift.

Unless can be controlled, a wage-stabilisation policy becomes nothing

it

more than a legal fiction with household incomes responding just as they
1 At the time of writing (early 1966), a somewhat more determined attempt is being made
to controlboth wages and prices. No serious apparatus of controls and sanctions has been
employed, however, and the demand-pull theorist must expect this experiment to have no
more significant results than the ones that preceded it. Fortunately the prediction of the
demand-pull theorists that the incomes policy is doomed to failure is testable: a few years
hence, the rate of change of wages and prices can be correlated with the appropriate indices
of demand (as detailed in the Appendix to Chapter 51) and a ‘dummy variable’ can be added
for the existence of incomes policy. The demand-pull theorist predicts that the partial

regression coefficient on this dummy variable will not differ significantly from zero. Fortunately
positive economics does provide tests such as these and does not leave us to rely on our personal
impressions and prejudices. (Since the time of writing an apparatus of sanctions has been
revealed. It remains to be seen if these can be temporarily successful and, if so far, how long
they can hold back prices in the face of a continually expanding money supply and general
excess demand.)
842 ECONOMIC GROWTH AND DEVELOPMENT

always did to market conditions, but with increases occurring because of

rises in bonuses, overtime pay and other evtra earnings rather than because
'
of changes in negotiated wage rates
We have de\oted considerable space to the control of inflation, first,
because this particular problem illustrates the important aspects of macro-
economic policy, and, second, because, as we shall see below, the policy
conflict between inflation and unemployment is one of the most serious
problems of macro economic policy
Possible policies for controlling inflation are shown in Table 58 4
Table 58 4
Instrumental fiscal

Incomes policy
variables
Ifrom exhortaton
to direct controls)

Balance-of Payments Problems

Causes Causes of balance-of-payments problems may be divided into
changes abroad and changes at home The central authortiics have little
or no control over the former so we shall concentrate on the latter There
are many
domestic causes of balance-of-paymcnis problems in a world of
fixedexchange rales A lower rate of economic development than in foreign
countriesis thought to be one such cause A rapid rise productivity helps m
in keeping costs down so that pnees can remain competitive abroad Also
innovations that arc the cause of growth mean that new and improved
products will constantly be coming into the market allowing domestic
firms to keep up with, or ahead of, foreign competition Changes in con-
sumption patterns can cause severe balance-of-payments problems as, for
example, when the change over from coal to oil firing of ship’s boilers
greatly reduced Britain’s export market for coal (and also increased her
import bill since the change over was made by Bntish as well as by foreign
ships) A third important source of problems is domestic inflation Inflation

1 If It did prove possible to control wages through an incomes policy a serious conflict would
be created between Che control of inflation and the re allocation of resources response to m
economic growth This latter problem ts described iw pages 811-15 To get the allocative
mechanism working cotceedy it u impomnttlat wages respond to demand anrf supply Wages
policy IS usuall) intended to make wages respond to supply changes (i e ,
productivity changes)
only
 MACRO-ECONOMIC POLICY 843

creates no problems for the balance of payments if all one’s competitors are
also inflating at the same rate since it is relative prices that matter in inter-
national trade, as in domestic trade. however, one country'’s price level
If,

is rising faster than the price levels of competitor countries

imports will rise
and exports will fall and balance-of-payments problems will ensue. ^

Control of balance-of-payments problems: Broadly speaking, there

are two main sets of policies that can be adopted to solve a balanee-of-
payments problem. First, national income may be reduced by raising taxes
or by lowering government expenditure. The fall in income will reduce the
expenditure of households on all goods, including imports. This policy is

called an expendilure-dampening policy, it relies on a general reduction in

aggregate expenditure to accomplish its goal of a reduction in expenditure
on imports. How successful the policy will be depends on the proportion of
income that is spent on imports. Where this proportion is small, as it is in
the US, a large reduction- in income will be needed to accomplish a given
change in imports. IVhere the proportion of income spent on imports is
large, as it is in Britain, a smaller change in income will be needed to
produce any given change in imports.^
The second major policy is an expenditure-switching one. The expenditures
of domestic households can be switched from foreign to domestic goods to
reduce imports, and the expenditures of foreign households can be switched
from goods produced abroad to goods produced at home to increase
exports. An expenditure-switching policy is accomplished by changing the
prices of foreign goods relative to domestic goods. This can be done by
taxing imports and subsidising exports, or by devaluing the exchange rate.
Expenditure-switching policies were analysed in detail in Chapter 53.
Expenditure-switching and expenditure-dampening policies can both
affect the balance of payments in the desired direction. Both of them, how-
ever, have certain side effects.
A policy of expenditure-dampening to reduce imports is somewhat like
shooting at a close target with a shotgun. The bull’s eye will usually be hit,

but so will a of other things. Specifically, a reduction in aggregate

lot
expenditure will reduce expenditures on all domestically-produced com-
modities. This means that output and standards of living will fall and that
unemployment will rise. These are side effects that are not usually regarded
as desirable, especially by those affected.

1 This is spelled out in detail in Chapter 53, page 759.

2 In general, any change in imports of an amount AM can be accomplished by changing
income by the amount AMIm, where m is the marginal propensity to import. This follows
from the relation AM=mAY. Thus if 50 per cent of income goes in imports, a change in M
of p,\ can be accomplished by changing Y by £2', if only 5 per cent of income goes in imports,
then a change in M
of ,C1 requires a change in Y of £20.
844 ECONOMIC GROWTH AND DEVELOPMENT
expenditure-switching policy will also have an effect on the level of
An
domestic income, output and employment If a devaluation of the British
pound lowers the prices of BnUsh goods relative *to the paces of foreign
ones both foreign and UK
households will buy more UK
produced goods
and fewer foreign produced goods This will raise incomes in those British
industries that produced the newly demanded goods, and, when the extra
incomes are spent, a multiplier process will be set up that will raise all
incomes m the UKOn the other hand, if the pound is appreciated, then
foreign and British households will buy fewer British goods This will lower
the income of those sectors of the UK
economy that produced those goods
the demand for which has now fallen Once the incomes of households in
these sectors have fallen their expenditures will be reduced, and incomes
throughout the UK will fall Among the important side effects of such a
policy IS a redistribution of income among sectors of the UK economy
Conditions favouring each policy We have seen that these two policies for re
moving a balance of payments deficit have opposite effects on income
Expenditure-dampening policies lower demand and income and thus tend
torecommend themselves to policy makers in situations of overfull employ
ment because in such situations it may be considered desirable to reduce
demand m order to check inflationary tendencies The use of expenditure-
dampening policies in periods when employment is less than or just equal to

the level considered to correspond to full employment will have the un

desirable effect of reducing employment and imposing a sacrifice on the
community in terms of foregone domestic output
Expenditure switching policies tend to raise demand and national income
by increasing the volume of domestic expenditure Such policies will tend to
appeal to policy-makers in situations in which income and employment are
below the desired level A successful expenditure switching policy m these
circumstances will simultaneously improve both the balance of payments
and the unemployment situation If an expenditure switching policy is to be
employment, however, it must be accompanied
successful in a period of full
by a policy of reducing domestic expenditure There is no point in switching
foreign demand onto your products if you are in a situation in which full
employment of resources already persists, because it will not be possible to
produce more output to meet the extra demand The appropriate policy
m such a situation is to reduce domestic expenditure by exactly the same
amount as the increase m foreign expenditure on domestically produced
goods This keeps national income and employment unchanged, but directs
a larger share of total output to exports, thus improving the balancc-of
payments situation
This point was not fully appreciated after the last war, and more than
one devaluation was made at times of overfull employment and already
 MACRO-ECONOMIC POLICY 845

existing inflationary pressure. For example, the Britishpound was devalued

in 1949 at a time of full employment when prices were already rising at
about 4 per cent per year. The devaluation switched expenditure into
British goods, but there was little scope to expand output. Devaluation as a
method of curing balance-of-payments problems was thus used in a set of
circumstances in which economic theory predicts that it had the minimum
chance of succeeding. It is difficult, therefore, to accept the evidence of this
period as refuting empirically the proposition that devaluations are an
effective means of alleviating balance-of-payments deficits.

Restrictions on the use of expenditure-switching policies: The

ability to adopt expenditure-switching policies has been limited to a great

extent by international agreement and even further by certain unplanned

post-war events. International agreement has greatly limited the use of
restrictions in imports as a means of switching demand from foreign to
domestic sources of supply, and it has also restricted the use of variations in
tarifis on imports and of subsidies on exports as a means of switching demand

onto domestic sources. It is probably somewhat more acceptable to the

international community to use tariff policies than to use physical controls
to accomplish this purpose, but there is still great resistance to the use of
tariffs. The latest example of such a use was in 1964 when, in the face of a
severe balance-of-payments disequilibrium, the British central authorities
instituted a 15 per cent surcharge on imports. There was a major outcry
against this, and the British Government had to give a strong assurance
that the surcharge was a temporary measure to meet a temporary imbalance
and that it would be removed within 18 months.
International agreement also restricts the use of exchange-rate variations
as a tool of expenditure-switching. By accepting a system of fixed exchange
rates, central authorities have completely eschewed the use of this policy

to remove temporary disequilibria in international payments. Most

countries do accept, however, that such changes can be made in order to
switch expenditures in the face of long-term persistent imbalances between
imports and exports.
The position of the pound and the dollar as reserv'e currencies puts an
added constraint on the central authorities of Great Britain and the United
States that restricts their use of devaluation as a means of switching
expenditure.' A devaluation of any resen-e currency such as the pound
are
means a cut in the value of the reserves that other central authorities
holding in the form of claims on that currency. To be the cause of such

losses is not something that would be accepted lightly by either of the

Bridsh or the American Governments. Reluctance is due both to the fact

I See Chapter 55, pages 788-90.

28
846 ECONOMIC GROWTH AND DEVELOPMENT

that the central authorities of thetwo countries care about public opinion
in other countries, and because, particularly in the case of Bntain, they
have encouraged the use of their currency as a reserve currency To change
the exchange rate of a currency with any frequency is to render that
currency unsatisfactory as a rcser\c currency
The policy problem of controlling the balance of payments is illustrated
m Table 58 5
Table 58 5
Instrumental Intermediate Policy
variables variaUes variable

Growth
Causes Economic growth is the policy variable about which we know least
In spite of a great deal of study, and the accumulation of isolated bits of
knowledge, we do not really understand the complex causes of growth
sufficiently well to be able to alter the growth rate as easily as we can alter
the unemployment rate No cast in which a government has intervened in a
free market society with the purpose of achieving a marked increase in its
rate of growthis generally agreed to have been a success Post-war France,

which has had a high growth rate combined with an active planning policy,
has probably come closest Many French economists are convinced that
their rather loose style of planning has had a significant effect on their
growth rate This is not an easy theory totest and it seems much less
obvious when we notice that West Germany has achieved growth rates at
least ashigh as those of France together with a policy of nonintervention
in an almost completely free market economy We can only say at this time
that the case is ‘not proven’, one way or the other Growth, therefore,
remains the enigma of macro policy , we agree that it is one of the most
important of all the variables, and we also agree that it is the one we least
know how to control
Certain theories have been put forward One is that penods of very high
 MACRO-ECONOMIC POLICY 847

demand and mild inflationary pressures are beneflcial to growth. It is

argued that such periods provide the businessman with both the incentive
to invest and the funds (which can be withheld out of profits) with which to
finance the investment. Another theory says that periods of moderate
excess capacity with an absence of inflationary pressures are most conducive
to growth. The argument is that, when there is some unemployment in
here
the economy, resources for new investment will be readily available, and
that innovations in terms of new products or cost reductions in old ones
provide the only possible promise of large profits. Research is progressing
on these theories, but as yet we really do not know enough to choose
between them.
On the negative side, it is clear that a substantial cut in demand, and a
recession of any serious degree, has at least a short-run depressing effect on
the growth rate. This follows at once from the accelerator theory of invest-
ment (see Chapter 44, pages 614—17). A
demand can be cut in aggregate
expected to cut drastically into new
argued that even if,
investment. It is

after a couple of years, a recovery in investment occurs, two years of

investment activity will have been lost forever and the growth of the
economy will always be two years behind what it otherwise would have
been. The evidence with respect to net investment appears to provide
support to this proposition.

Control: Since we are not very sure about the causes of growth, we are
not very sure about which policy measures will stimulate it. Many econo-
mists believe that anything that increases the rate of new investment will
be favourable to growth. In addition to preventing reductions in the level
of aggregate demand, investment can be encouraged by providing ample
funds at low interest rates, by providing tax advantages to investors, and by
encouraging research and development. How successful we think we shall
be depends on which of the theories of the advancement of knowledge laid
out on pages 289-90 we accept.
To encourage growth, we can also seek to change the educational system,
the health of the nation and the attitudes toward invention and innovation,
as well as to improve business practices by inventing and publicising new
techniques of business management. We shall not know whether or not any
already advanced Western nation can succeed in significantly altering its
growth rate by adopting a serious programme to do so along these or other
lines until some country really tries to do so.* If increasing the rate of
economic growth were the major objective of economic policy, it is clear
that avoiding severe depressions would be important and that stimulating
1 The policy of the British central authorities of publishing a document saying what they
would like the growth rate to be and calling it a ‘National Plan’ can hardly be regarded as a
serious programme for altering the growth rate.
848 ECONOMIC GROWTH AND DEVELOPMENT

invention, innovation and investment would likewise be desirable We do

not know enough about these processes to be confident of which policies
would best achieve the latter objectives

POLICY CONFLICTS
We have seen above that, with the possible exception of growth, economists
have knowledge of the economy to enable policy makers to achieve
sufficient

any one goal may be quite impossible, however, to fulfil all policy goals
It

simultaneously It may well be that a policy measure will bring us closer

to some of our objectives but take us further away from others In such
cases, policy objectives conflict with one another in the sense that we can
get closer to one goal only at the cost of moving further away from another
Thus, It IS not sufficient for governments to decide which objectives are
worth pursuing, they must also decide on some rate of sufistitutwii between
them, they must decide how much of one it is worth sacrificing order to m
get more of the other Many of today’s controversies over matters of
economic policy are over this issue of the relative importance of different
objectives, each objective being accepted as desirable eeleris paribus
Practically everyone, for example, accepts both a high level of employment
and the control of inflation as desirable goals of policy Where there is real
disagreement, however, over the relative importance of these two goals
is

Some people would be prepared to accept an inflation of, say, 2 per cent
per year as the price of having unemployment as low as per cent of the
labour force, other people would be prepared to let unemployment be as
high as, say, 2^ per cent or 3 per cent if that would ensure a stable price
level
A simple situation with no policy conflicts occurs when there arc a
number of separate instrumental variables each one of which aflects one,
and only one, policy variable This is shown schematically in Table 58 6

Table 58 6
THE LINK BETWEEN GOVERNMENTAL ACTION AND
POLICY VARIABLES WITH NO POTENTIAL POLICY
CONFLICTS
Gi -* - Nu ,N„-*U
/,

The Tabic reads government policy, G|, affects instrumental variable, A,

which, through a possible chain of intermediate variables, to N„, finally
 ;

MACRO-ECONOMIC POLICY 849

affects the policy variable, unemployment (and no other policy variable)

policy G2 affects instrumental variable, I2 and then, through a chain of
,

possible intermediate variables, N„ to Np, it finally affects (only) the price

level and so on for policies G3 and G^,.

In so simple a ivorld, the government needs only to vary Gj until it

secures the level of U it desires; it then varies G2 until the desired level of P
is achieved, and so on, to G3 and G4, until all its policies are fully achieved.
There are no policy conflicts.

Policy conflicts arise as soon as some improvement in one policy variable

can only be achieved at the cost of worsening the performance of another
One of the most serious sources of policy conflict in the real
policy variable.
world stems from the facts that aggregate expenditure is the only inter-
mediate variable that significantly affects unemployment while it also has
a significant affect on other policy variables.
The situation is illustrated in Table 58.7 where the symbol ? indicates
uncertainty.

Table 58.7

THE RELATION BETWEEN AGGREGATE DEMAND

AND FOUR MACRO-POLICY VARIABLES

Monetary policy Aggregate demand

^? Other intermediate
Fiscal policy "" ^ ? variables

Thus both fiscal and monetary policy can be used to increase aggregate
demand. A rise in aggregate demand in turn loivers unemployment which
we like but also raises the price level, worsens the balance of payments and
has an unknown affect on the rate of growth. Further, both fiscal and mone-
tary policies may affect other intermediate variables and through them may
have favourable or adverse affects on each of our four policy variables.

Current Policy Issues

Unemployment and prices: Unless the policy of controlling incomes and
prices through an incomes policy succeeds, there would seem to be an
850 ECONOMIC GROWTH AND DEVELOPMENT
irreconcilable policy conflict between fuH employment and stable prices
since the higher the let el of aggregate demand the higher the level of
employment but the more rapid the rate of inflation What can the econo-
mist do in the face of such a conflict’
The first thing he can do is to attempt to discover the quantitative
significance of the conflict The conflict looks very different, for example,
if 10 per cent unemployment is needed to maintain stable prices than if
only 3 5 per cent unemployment is needed The nature of '
this conflict is

shown by the curve m Figure 51 2 (page 703) Before recent empirical work
on was done, it was generally believed that something on the
this relation

order of 7 10 per cent unemployment might be needed to maintain stable

price levels As a result of recent work, our view of the nature of the policy
conflict has changed greatly In Britain, it seems fairly clear that something
in the order of 2 5-3 pel cent unemployment would be compatible with

stable prices while 1 I 5 per cent unemployment might go along with some-
thing on the order of a 3 per cent annual rate of inflation In this work, we
see the economist in one of his most important roles, that ofdiscovenng what
policy conflicts exist in the present state of the world, and quantifying them
by telling us how much of this we have to give up to get more of that
The next thing (he economist can do is to study what changes can be
made m the economy to reduce the importance of, or to remove altogether,
the existing conflict In terms of Figure 51 2, he is asking how to shift the
curve downward If it could be shifted so that, at a volume of measured
unemployment corresponding to ‘full employment’, the rate of increase in
wages was no more than the rate of increase m productivity, the conflict
would be removed altogether
One suggested method, that of direct intervention through wage-price
policy, has already been considered Other methods might require less
direct intervention In order to develop such policies, the economist has to
understand the actual behaviour of individual markets that gives rise to
this conflict A brief outline of this behaviour was given in the Appendix to
Chapter 56, and we may bnefly recall it here One reason why inflation
sets in before full employment is reached is that in a growing, changing

I To measure the extent of the conflict, wc need a full iheorx of the relevant relations m the
economy Here we can specify a simplified theory that u enough to illustrate what is involved
(1) The rate of change of productivtiy i» exogenous (2) Wages are a major proportion of
factor incomes and other incomes will change m approximately the same proportion as wage
incomes (3) The problem of maintaining a stable price level thus boils down to one of having
total wage income (which is the major detemunant of aggregate demand) rise at the same
speed as productivity (which major determinant of aggregate supply) The conflict can
is the
then be studied by observing the relation between the level unemployment and the rate of
increase in wages (More detailed work has used more adequate and necessanly quite
advanced theories linking these vanables together )
 MACRO-ECONOMIC POLICY 851

economy, the pattern of consumers’ demand is constantly changing, and

with it also the pattern of derived demand
consumers wish for factors. If
to spend an increasing fraction of their incomes on automobiles and a
declining fraction on, say, food, then an increasing proportion of the
nation’s resources will have to be in auto production and a declining pro-
portion in food. (See pages 811—15.) Such re-allocations of resources take
time and often entail the retraining of labour and its movement from one
geographical area to another. At any one time, there will be a supply of
unemployed resources that have been released from industries in which they
are not required and that have not yet moved to industries in ivhich they
are required. Tims, if aggregate demand expands, shortages and bottle-
necks and consequent infiationar}'^ pressure will develop, even though there
are unemployed resources. If movement of these resources could be speeded
up considerably, then bottlenecks and shortages would be less likely to
occur in expanding areas until over-all unemployment were at a very low
level.
Until successful policies for reducing or ehminating the conflict have been
designed, is necessar)' to choose between the Uvo competing policy
it

objectives. Hotv much inflation is it worth having in order to gain a further

reduction in unemployment? Here is an area where value judgments tend
to predominate. Reasonable men can differ in their evaluation of how bad
a bit of inflation is and of how bad a bit more unemployment is. They can
also disagree on how to w'eigh effects on different people - on how, for
example, to weigh the harm that inflation does to them against the harm
that unemployment does to someone else. These are the kinds of issues on
which political parties take stands; once the nature of the conflict has been
made clear and ways of removing it studied, the final choice of hois' we
value these t^vo goals relative to each other becomes a political one.’
In Britain it is safe to say that, since the last war, the balance of opinion
has been in favour of sacrificing some price stability in order to achieve a
ver\" high level of employment. Recently there has been a resurgence of
the view that, although no-one wants a return to the 1930’s, a somewhat
higher level of unemployment - say up to 3 per cent, would not be too high
a price to pay for removing inflation. In this case we have a current policy
debate about the relative valuations to be placed on two goals. It has also
been recognised however that the choice between unemployment and

1 This does not mean that rational enquiry ceases at this stage. If you think inflation is

more important than unemployment and 1 think the reverse, we can ask each other our

reasons for thinking so. We are then likely to come up with some positive testable statements
such as, ‘I think the effects of each will be thus and so’. Such statements are the factual basis
on which our judgments rest, and they tvill usually be testable, at least in principle. It is

important in this context that you now re-read footnote 1 on page 5.

852 ECONOMIC GROWTH AND DEVELOPMENT

inflation cannotbe taken in isolation The degree of inflation affects the

balance of payments the more rapid the inflation, ceUris paribus, the worse
the balance of payments
This link creates a serious policy conflict between full-employment and
balancc-of-payments policy, at least in a world of fixed exchange rates

Unemployment and the balance of payments With fixed exchange

rates,a major policy conflict can occur beltvccn unemployment and the
balance of payments If the country’s currency is overvalued so that at full
employment there is a balance-of-payments deficit, then an expenditure-

dampening policy will reduce imports and thus alleviate the balancc-of-
payments problem, but at the cost of raising the level of domestic un-
employment The conflict can be removed, at least temporarily, by raising
aggregate demand, while at the same time adopting expenditure-switching
policies so that international payments arc in balance when full employ-
ment is achieved If, however, the central authorities cannot, or will not,
adopt expenditure-switching policies, there vvill be no simple way out of

their dilemma, and thev will have to choose between internal and
EXTERNAL BALANCE '

Furthermore, even if no conflict exists at the moment, a policy of securing

employment at the cost of inflation makes it likely that the conflict will
full

emerge We have already seen that if a country’s price level is rising faster
than the price levels of its trading partners its balance of payments will tend
to worsen Thus full employment plus more rapid inflation than the rest
of the world will create balance-of-pavments problems and, if the country
IS determined to maintain a fixed exchange rate, a conflict between full

employment and external balance will occur, since expenditure dampening

policies will be the onlv way remaining to restore external balance
Throughout the 1950*8, coonlnes such as the United States and Germany
did not face such a conflict, since their currencies were not overvalued at
the current exchange rate when full employment ruled at home,
Indeed,
both countries have balance-of-paymcnts surpluses. Other
tended to
countries, including Bntain, were in positions m which full employment
tended to be associated with payments deficits As long as expenditure-
switching policies could not be adopted, there was pressure to reduce
aggregate demand in order to restore external balance Once external
balance had been restored, there was a tendency to worry about the
unemployment that was the direct consequence of the expenditure dampen-
ing policies In order to deal with the unemployment, aggregate expenditure
would be raised This would lead to a lowering of the unemployment rate
1 A situation of full employment is irften called a situation of internal balance while a
situation in which international payments arc in balanccis often called one of external balance
 MACRO-ECONOMIC POLICY 853

but, no sooner would the unemployment problem appear to be soK'ed,

than the rise in demand, with its consequent rise in imports, would cause a
new balance-of-payments disequilibrium. Aggregate demand would then
have to be depressed again. The history' of macro-economic policy in
Britain during the 1950’s and 1960’-s, and, indeed, in other countries
similarly situated, has been a perpetual oscillation between periods of
expanding aggregate demand to eliminate unemployment and periods of
contracting demand to eliminate balance-of-payments deficits. The policy
was given the name stop-go. Fortunately,
that led to these alterations
the level of unemployment at which payments tended to be in balance in
Britain was not high, so the oscillations were not large. They were
frustrating, however, and many economists believe that the associated
uncertainty hurt long-run investment, thus retarding the rate of growth.*
The whole problem arose because the central authorities; (1) would not
accept expenditure-switching policies to cure the balance-of-payments
problem, would not admit that this refusal created a conflict between
(2)
internal and external balance and (3) would not make a clear decision
about how much unemployment they were prepared to accept in order to
secure external balance. In such a situation the task of the economist is to
point out the policies necessary to achieve both goals, show the conflicts
made inevitable by the decision to use only some tools of policy, and press
for a clear decision in giving relative values to the ttvo conflicting goals.
Such is the force of the policy conflict that when the Labour party replaced
the Conservative party as the government of Britain after the general
election of 1964 it continued the stop-go policy of which it had been a

severe critic throughout the 1950’s. Taking power during a severe balance-
of-payments crisis and refusing to devalue the pound it had little option but
try to depress domestic demand in order to curtail imports.^
In recent times the United States has moved, as we have already seen, to a
position in which her international payments tend to be in deficit when
income is still below the full-employment level. We, thus, expect that

reducing the level of unemployment will exaggerate the payments problem.

The same conflict that other countries faced in the 1950’s has now appeared
1 This theory has not yet been satisfactorily tested.

2 It was ironic that a party that had just won an election on a policy of modernising the
country’ and ending Stop-Go began by declaring a six months’ moritorium on investment
expenditure — including university building. The effect on the balance of payments was in the
desired direction but some economists wondered about the effect on the growth rate. The
subsequent attempt to operate a serious incomes policy did, however, represent an attempt to
break out of the policy dilemma that had given rise to stop-go. Since this passage was written
the Labour government has found it necessary to initiate (in the Summer of 1966) another
severe bout of expenditure dampening in the face of another all-too-predictable balance of
payments crisis.
 MACRO-ECONOMIC POLICY 855

point but should be understood that if it were possible to do so at low

it

cost then one of the most serious policy conflicts that has dogged the UK
economy between full employment and the maintenance of
since 1925, that
satisfactory balance of payments, would be eliminated.

WHY ECONOMISTS DIFFER OVER POLICY ISSUES

In the 1940’s it used to be a common saying that if you asked twelve

economists for advice you got thirteen different answers, one from each of
any eleven economists and two from Keynes. E\'en to-day such wild
diversities can be found in many spheres. For example, about as many
economists in Britain seem to think joining the European Common Market
would be extremely beneficial for Britain as think it would be something
verging on a disaster.^
\Vhy is it that economists can differ so Uolently when called on for
professional advice ? There are many reasons and we shall only mention a
few that are most important in assessing the issues raised in the present
chapter.
1 Economists may differ on the trade off rate between various policy
variables. One economist may think 2 per cent inflation is a small price to
pay for 1 per cent unemployment and another may think that 3 per cent
unemployment is a small price to pay for a stable price level. Although it is
possible to continue a rational debate between persons holding these two
views, this is the t}"pe of issue that is intimately affected by our value
judgments and it is the kind of issue over which we will not be surprised to
find reasonable men disagreeing even after prolonged rational debate.
2 Economists may disagree over the relation beUveen variables. In
many such cases there is For example, most
substantial agreement.
economists agree that eridence has established between the relation
aggregate expenditure on the one hand and prices, unemployment and the
balance of pa^mients on the other. In other cases, however, there is no
agreement. We do not know the relation between aggregate expenditure
and economic growth. In most cases such relations are positive questions
and further research can lead to a concensus of opinion on the matters.
3 Economists can disagree over the indirect effects of some policy
change. Thus, for example, virtually aU economists agree that a devaluation
of the pound would improve the British balance of payments at least
temporarily. They disagree greatly on the other effects of such a devalua-
tion, some think them of secondary importance and some think them to be

1 See the survey conducted by the London Observer, 14 October 1962, page 5. For an
interesting discussion of the issues involved in this particular debate see T.W.Hutchinson,
Positive Economics and Policy Objectives, Allen & Unwin, 1964, pages 99 ff.
856 ECONOMIC GROWTH AND DEVELOPMENT

very senous indeed Since devaluations arc rare events it is difficult to test
these opposing views against empirical evidence
4 Economists can differ because they are asked questions for which
existing theory makes no clear predictions at all An example of such a
question is the effects on Britain of entry into the European Common
Market Members of the profession hold difTenng views on howsuch
questions should be handled Some economists, for example, take the view
that we can only speak with authonty on such issues as how to remove
US unemployment (where we agree we know the answer) if we admit that
\n other cases we arc ignorant Other economists take the view that
their authonty depends on having an answer to all questions and that the
training of an economist will give him superior intuition in answering all
economic questions so that his record of guesses will be better than the
layman s To this the first group of economists reply that true authonty
comes from honesty and that the economist’s record of guessing is no better
than the layman’s They argue that, where there is an empirically tested
theory, the economist performs ten times better than the layman, but that,
where there is none, he guesses no better than the layman and only dis-
credits himself by taking extreme positions many of which are subsequently
discredited by the facts But here is a place where presumably rational
economists disagree and only time can tell which side is correct

THE COMPOSITION OF MACRO VARIABLES

Macro economic policy is concerned with the behaviour of certain key
averages and aggregates such as the average level of all prices and the over-
all level of unemployment In fact, we care about more than just these
averages and aggregates VVe also care about their composition Is the
over-all level of unemployment made up of very unequal rates of un-
employment such as those among industries, occupations or areas, or is it

made up of rates that are very similar across all of these classifications ’We
would assess a 3 per cent over-all unemployment rate in the very UK
differently if it resulted from 3 per cent unemployment m
all mdustnes,

occupations and geographical areas than if it resulted from 12 per cent

unemployment in some areas and only 1 or 2 per cent m others
Levels of unemployment of 10 or 15 per cent are very senous matters,
indeed They are likely to mean that many people will be out of work for
very long periods of time and that there is a level of social and personal
upheaval that just does not accompany rates on the order of 3 per cent
The degree of regional and occupational mequahty in unemployment rates
in both the UK and the US has remained quite large throughout the whole
 MACRO-ECONOMIC POLICY 857

post-war period.^ Clearly, although we should be satisfied that the over-all

rate in Britain has been held at a veiy^ low level throughout the whole post-
war period, we should be very disturbed at the very high rates that persist
in some places.
Why do these localised high rates occur - why is it that in the midst of
the affluent society we have these persistent pockets of poverty, pockets
that will not respond to the cure of raising aggregate expenditure? Why
does the market not adjust to bring about approximate equality in un-
employment rates? Shouldn’t regions and occupations with excess supplies
of labour find their relative wages declining so that there is an incentive to
employers to hire more of the relatively cheap labour? Does this market
mechanism work at all ? How fast does it work ? Does the shifting pattern of
economic growth continue to disturb markets so that the adjustment
mechanism can never catch up? Would things be any better if we interfered
in the market mechanism? 'What government policies would reduce these
labour-market disequihbria ?
A similar set of ‘disaggregated observations’ could be produced for any
macro variable,and they ivould provoke a similar set of questions. But to
state just one set is enough to show that we have now gone full circle and
are back at the micro-economics with which we began our study. To tackle
these problems, we need to return to market theory, the study of which we
began in Part 2 of the book.
As we stated in Chapter 39, there is no sharp distinction between micro-
and macro-economics. There are merely higher and lower levels of aggrega-
tion and a series of questions appropriate to each level of aggregation, with
each series shading one into the other.

THE SIZE OF GOVERNMENT

In Chapter 38, we discussed micro-economic policy measures from many
points of view, and we examined some of the issues involved in the choice
between laissez faire and public intervention. In Parts 7 to 11, we have
discussed the effects of macro-economic policy, but have not said much
about the pros and cons of having the government provide this service.
Part of the reason is that the only alternative to government action in
the macro sphere is no action at all. We can argue about the relative
advantages and disadvantages of having education provided by the private
or by the public sector, butwe cannot argue about the advantages and dis-
advantages in the case of full-employment policy since the only alternative

1 But, contrary to popular belief, it has not increased steadily over the period. For the US,
see R. A. Gordon, ‘Has Structural Unemployment Worsened?’, Industrial Relations, III, May
1964, 55-77 and for the UK see a forthcoming article by F.P.R.Brechling, ‘The Measure-
ment of Structural Unemployment in Britain’.
858 ECONOMIC GROWTH AND DEVEIOPMENT

to having a full employment policy administered b> the go\ernment is to

letunemployment settle at wliale\er level the private market determines
The evidence of histor> is that this will mem accepting quite high rates
quite often
In discussing fiscal policy, we have examined government revenues and
expenditures as they affect our macro policy goals We have not discussed
the \aluation to be put on the ratio of governmental expenditures to total
national income (which we shall abbreviile as C/T) It is undoubtedly true
that many people place valuations on the size of this ratio Some think it

should be kept down, others think it should be high, and yet others think its

over-all size is unimportant as long as each particular public project is

acceptable
What can the economist say about this problem^
he can point out that the decision on C/i cannot be made over all
First,

but can be made only m

terms of individual decisions on individual projects
It does not help practical policy to say simply, ‘The size of the government
sector must be reduced by 20 per cent’, it is necessary to decide which 20
per cent must go It might be possible, for example, to obtain a sizeable
majority m favour of cutting government expenditure by 20 per cent and
yet be unable to find a majonty m favour of cutting any particular bit of
expenditure In fact, the aggregate budget of the public sector is built up
by approving particular projects Uc make piecemeal decisions on housing
and education, national defence, the health service, national parks and the
budget of the univeniiies
Those who fear the size of the government sector will be unwilling to
accept any particular nev^ measure unless persuaded that its particular
advantages overvveigh ilvcir general fears Those who favour government
activity will need less persuasion, but, unless they arc dogmatic extremists,
they will not accept any measure, no matter how dubious its advantages or
how high Its costs, just because it is to be provided by the public sector
Those who think the size of the public sector per se is irrelevant will tend to
be for or against any particular measure solely on its assumed merits and
dements
The second thing that the economist can point out is that people's re-
actions to specific government measures vary over lime When the income
tax was first introduced in ninctcenlli century Bntain, it was at a rate of
6d in the pound and the British Parliament expressed profound concern that
this rate would destroy incentives because people would not retain the full

fruits of their efforts When pubbe education was first introduced, many

persons felt that such free gifts would destroy the energies and initiatives of
the individual household Possibly some people feel the same today, but both
the income tax and public education now receive overwhelming support
 MACRO-ECONOMIC POLICY 859

What can be when people express opinion on the size of the

of concern
government sector per se? Partly it is a political matter; people have fears
and hopes about government, and they view the importance of government
in the economy in terms of their political philosophy. This is without doubt
a very important aspect, but it is beyond the scope of the present economic
inquiry.
People do debate the size of the government on economic grounds. When
they do, it is rarely that they feel a G/T of 20 per cent is better or worse
‘economically’ than a G/7 of 30 per cent. It is rather that there is usually a
large concealed theoretical structure leading to the prediction that to raise
(or lower) G/7 would raise (or lower) living standards, welfare or whatever
is of concern to us. Consider two examples.
Economist A might believe that we have been reasonably successful in
getting an ordering of the relative efficiencies with which the public sector
can handle certain jobs, but that we have been wildly wrong in judging the
absolute efficiency with which it can do any job. This economist would
argue that we have long since allocated to the public sector all those things
it could do better than the private sector and that, in doing so, we allocated
to it lots of things it does worse. He would thus believe that cutting Gj Y
substantially would raise the over-all efficiency of the economy and add to
human freedom by reducing the coercive element in our society. Economist
A also believes that the divorce between receipt of benefits on the one hand
and the payment for these benefits on the other hand that exists in the
public sector leads households to vote themselves more public goods than
they would in fact buy if these goods were sold in a free market where the
cost would be obvious. Thus A believes we over expand the pubhc sector
because at each election we vote ourselves more roads, schools, parks,
and a host of other government goods
universities, defence blissfully un-
aware of the account in terms of higher taxes that must inevitably be
rendered.
Economist Z might believe that the goods the public sector provides tend
to have high income elasticities of demand so that, if people consumed the
bundle of goods they most wanted, the size of G/7 would rise as incomes rose
because of economic growth. But, so Zbelieves, the public fears that a higher
G/7 will increase the chances of a Fascist or Communist government being
established. If the people’s assessment of the political consequences were
correct, the decision to go without the goods they want would be a rational
one. Z believes, however, that the people are wrong in this assessment of the
political effects of raising G/7. If only G/7 could be increased, the people
would see for themselves that their view of the political effects was wrong,
and they would then be happy to find themselves able to consume the goods
they really want to consume. Z will wish to raise G/7 whenever possible.
860 ECONOMIC GROWTH AND DEVELOPMENT

Faced with a need for a budget defiat to reduce unemployment, he will

want to raise G while others may want to cut taxes The effects on un-
employment in the long run will be more or less the same, but the effects
on the long-run willingness of the pubhc to accept a different level of G/T
may be very different Economist Z also argues that since there is no market
for most public goods there are no signals to show that consumers desire a
re allocation of resources into the pubhc sector If consumers want more
cars and fewer cotton shirts the consequent switching of their expenditure
sets up a senes of market signals that ends in there being more can and

fewer cotton shirts produced If consumers want more education and fewer
cars there is no market by which they can express their preferences Thus
argues Z the pubhc sector expands slower than consumers would have it
expand if there was a market in which they could express their relative
preferences for the goods produced by the pnvatc and the public sectors
Which economist, A or Z, is correct Indeed, which one of a dozen other
'*

views about the short- and long-run economic political effects of varying
GjY is correct’ Only time and the subsequent research of a host of
economists and political scientists of generations to follow ours - will tell

THE PROGRESS OF ECONOMICS

The general acceptance of the view that the validity of economic theories
should be tested by confronting their predictions with the mass of all avail-
able evidence is fairly new in economics At this point you should re read
the quotation from Lord William Bevendge given at the beginning of this
book (see pages xi and xii) The controversy that Beveridge was describing
was the one that followed the publication in 1936 of Keynes’ General Theory
of Employment Interest and Money Keynes* work gave nse to the macro
economics that we have developed m Part 7 and on which we have so
often relied in subsequent parts At many points m the present book, we
have raised the question of how various parts of macro economic theory
could be tested, we have also discussed some of the tests that have already
been conducted The student should reflect on how very different this
approach to the problem of accepting or rejecting theories is from the
Sevendge
appitiatii itsen'oed \sy
There is no doubt that since economics first began some, albeit irregular,
and halting progress has been made in relating theory to evidence in the
world of economic events This prepress has been reflected m the superior
ability of governments to achieve their policy objectives The financial
aspects of the Second World War were incomparably better handled than
those of the First World War The pathetic efforts of successive British
governments to deal with the economic catastrophe that overwhelmed the
country after the return to the gold standard and even more so after the
 MACRO-ECONOMIC POLICY 861

great crash of 1929 show measures adopted in all sincerity which in most
make things worse.^ When President Roosevelt tried
cases actually ser\'ed to
to reduce American unemployment in the 1930’s (with much more vigour
than was applied by any British Government over the same period) his
efforts were greatly hampered by the failure of most economists to realise

the critical importance of budget deficits in increasing the level of the

circular flow of income.The contrast between the unhappy period of the
1930’s and the present is great. When President Kennedy ^sashed to do
something about the high levels of unemployment in the 1960’s, his main
problem was to persuade the American Congress to adopt what most
economists agreed was an appropriate cure - a tax cut; and in 1964, when
President Johnson finally persuaded Congress to accept the tax cut, the
ensuing rise in output and employment was very close to ^vhat the econo-

mists on the President’s Council of Economic Advisors had predicted it

would be. Several times during the 1950’s and 1960’s successive British
governments have altered the Budget in pursuit of a ‘stop’ or a ‘go’ policy.
Many economists are critical of the motives behind these poHcy oscillations
but the fact that each time the economy moved in the direction predicted
by economic theor)' is impressive evidence that we have learned a great
deal about the behaviour of the economy in the last 40 years.
Such important policy areas as the running of wars and the curing of
major depressions are places in which the general tone of our theories is
tested, even if all the specific predictions of them are not. In some general
sense, then, economic theories have always been subjected to empirical
tests. When they were wildly at variance with the facts, the ensuing disaster

could not but be noticed, and the theories were discarded or amended in the
light of what was learned.
The advances of economics in the last 20 years reflect a change in
economists’ attitude toward empirical observations. Today, we are much
less likely to dismiss theories just because we do not like them and to refuse
to abandon theories just because we do like them. Today, we are more likely
to try to base our theories as much as possible on empirical observation and
to accept empirical relevance as the ultimate arbiter of the value of our
theories. As human beings, the upsetting of a pet theory may cause us much
anguish ; as scientists we should try to train ourselves to take pleasure in it
because of the new knowledge we gain thereby. It has been said that one
of the great tragedies of science is the continual slaying of beautiful theories
by ugly facts. As economists, we are all too often swayed by aesthetic con-

siderations. In the past, we have too often hung on to our theories because

they were beautiful or because we liked their political implications, as

scientists, we must always remember that, when theory and fact come into

conflict, it is theory, not fact, that must give way.

1 See, e.g., the message from King George V on page 587.
INDEX

Absolute advantage, 739 and unemployment, 852-5

accelerator theory of investment, 61 4-1 7, balance of payments accounts, 749-53
624 current, 750-2
advertising costs, 349-50 capital, 750-1, 752 3
Agency for International Development, balance sheets
822 accountants’, 255-9
aggregate consumption function, 598 bank’s, 678, 679, 681, 683, 687,
aggregate demand, 799-800 688
aggregate expenditure, 566 economist’s, 259-60
in the circular flow of income, 568 banks, and supply of capital, 240
equal to income, 581 banks, central, 684, 69{^1
graphical representation, 569 74 bankers to commercial banks,
aggregate expenditure function, 609 684-5
aggregate supply, 799 800 bankers to the government, 685
agriculture controllen and regulaton of the
problems of, 139 money supply, 685
programmes, 142
stabilisation banks, commercial, 677, 691
and increase in real income, 148 Bank of England, 677, 664
and re-allocation of resources, 150 Bank Rate, 684, 689
policy, 654 banking system, 677f
Allen. G C 459 ,
Baumol. William H , 398, 628
Allen, R G D 31, , 159, 630 Beeching Report, 533
Amery, L S 534 782 ,
Beveridge, Lord William, xi-xii, 860
antimonopoly laws, 379 biased sample, 47
arms expenditure, 659 6\ bilateral monopoly, 456
assets, value of, 446-7 Bismarck, 734
autonomous expenditures, 568 ‘black markets’, 129 31
autonomous variable, 36 bondholders, 239
average propensity to consume 596 7 booms, 622 3
average propensity to impori i 34 international propagation of, 636
average standard of living, I il Bowen, William G ,
644
boycotts, 459
Break, GF 425 ,

Backed paper money, 670 Brcchling, F P R , 857

Bam, Joe S 284, 324, 325
,
Bretton Woods Conference, 785-7
balance of actual payments, 749 53 budgets, balanced versus unbalanced,
balance of desired payments, 748-9 652 3
balance of payments, 748-53, 833-4 budget deficits, 658
effect of devaluation on, 762-3 budget line, 168
actual, 765 changes m
income, 169
equilibrium, 778 9 changes in price, 170
control, 842-6 business cycles, see trade cycles
 INDEX 863

Canada, tariffs, 776 simplified model of, 553-63

exchange rates, 795 withdrawals from, 554
capital injections into, 558
definition, 59 and saving and investment, 556-7
as input, 230 and foreign trade, 557
money and real, 238 equilibrium in, 564f
as costs, 252 lags in, 575
supply of, 427, 475 gov'ernment expenditure as an injec-
mobility 430
of, tion into, 641, 652
and economic rent, 444—5 tax revenue as a withdrawal from,
marginal productirity of, 472 641, 652
demand for, 471-75 ‘Classical Theorj'’ of employment, in-
475-6
price, and money, 610-12
terest
movements in a free market, 765-7 clearing house, 680
short-termmovements of, 766 closed shop, 459
long-term movements of, 766-7 coal mining, 533-5
human, 806 coinage, 669
importance for underdeveloped collective bargaining, 460
economies, 820-4 collective consumption goods, 512 13.
imported, 821 529
contributed, 822 commercial banks, 677, 691
capital account, balance of payments, commodities, definition, 59
750-1, 752-3 Common Market, 771-3, 787-8
capital expenditure, government, 642^ comparative advantage, 736, 738
capital goods, 556 pattern of trade determined by price
opportunity cost of, 646 system according to balance of,
capital market, 241 741-6
car industn' and trade, 631-2 comparative costs, 738
cash balances, 694 comparative static equilibrium analysis,
unavoidable, 695-6 172-3, 564
avoidable, 695, 696-7 and theories of competition, 352
speculative, 697 compassion and free market, 515
central authorities, 77-8 competition
as borrowers of funds, 474 nonprice, 335, 340
influence on rate of interest, 476 see also perfect competition, monopo-
activities of, 505 competition
listic

centrally controlled economy, 79 competitive markets

and allocation of capital, 477-9 allocation of resources, 74
and market signals, 506 supply and demand, 182
central bank, 642, 684 complements
ceteris paribus arguments, 42 demand for, 84—5
change in tastes and exchange rates, 758 shifts in demand curv'e, 91

cheque, 674 concentration ratio, 324—5

choice, 65 constant return to scale, 269
and indifference curt'es, 177 consumer
circular flow of expenditure, 539 in free market, 74
circular flow of income, 545f sov'ereignty of, 74, 77
between individual households, 545-6 satisfaction, 59
between households and firms, 546-8 taste, 71
and stages in production, 548-9 rationality', 202—3
864 INDEX

consumer durables, measurement of de- Davis, Richard G , 644

mand, 219 Debenture capital, 239
consumption expenditure, 570 debts and inflation, 176
consumption function, 594f fixed money, 625
graphical representation, 595 decreasing return to scale, 269
for the community and the individual, deficit, budget, 641

598 deficit financing,641-7

empirical observations, 599 deflation, 175, 700
long-run, 601 deflationary gap, 647, 651
short-run, 60! demand
control of shifts in, 604 shift in, 71-2
convertible paper money, 670 as flow, 80
and price, 80, 82
theory of, 247f ahd price changes, 108-12
average, 275-7 conditions of, 84
constant, 269 and 84-5
substitutes,
fixed, 254 and complements, 8f-S
marginal, 276-8 and income, 65-6
total, 263, 275 and income distribution, 89
variable, 254-5 and income changes, 118-19
short-run, 270f and inferior goods,66
opportunit) costs, 65-6, 248-51 and 87
tastes, 81,
measurement of, 248 and population size, 69
imputed cost, 249 excess, 100, 150
and payment to capital, 252 measuTemem of, 215f
social and pnvate, 253-5 effect of changes in, ICO, 249-56
changes in competition and mono- for faciors of production, 408 10
363-S
poly, elasticities of, 409-10
advertising, 349 50 for foreign currency, 753
cost curves, demand curve, 83-4
long run, 267-8 importance of, 89-90
short-run, 270-4 90-92, 102, 123
shifts.
marginal, 400-1 movements along, 92 3
cost function, 263 and supply curve, 99 100,214
cost minimisation, 262 excess, 106-7
in competition and monopoly, 381 slope of, 109
cost push, 720-30 elasticity of, I lOf
countercyclical 6scal policy, 653-4 in agnculture, I39f
credit controls, 689 and market stability, 157-8
credit rationing, 475, 478 and marginal utility, 186-7
cross elasticity, 120 substitution effect, 198-9
and changes in demand, f/5 statisticaf studies of 215~18
cross sectional data, 56, 600 in perfect competition, 296-7
cumulative movements, 624-5 inmonopoly, 315-16
current account, balance of payments, demand deposits, 675
750-52 demand elasucity
currency control, 685 definition, 112-13
currency foreign, elasticity of supply and and changes in total revenue, 116-17
demand, 753-5 determination of, 117-18
currency, reserve, 789 cross elasticity, 120
 INDEX 865

demand elasticity — continued causes of, 805

measurement of, 122-5, 218-21 opportunity cost of, 808
importance of, 137-9 change of rate, 810-11
and price fluctuations, 140 and allocation of resources, 811
at equilibrium, 360 accelerating rate of, 813
in perfect competition, 297-9 in underdeveloped economies, 818-24
for importsand exports, 754 balanced and unbalanced, 825-6
demand pull, 720-30 control of, 846-8
demand theory, testing of, 204—14 economic rent
Denison, E. F., 811 definition, 437-9
deposit money, 678 and transfer earnings, 439—42
depression, 621-2 and payment to labour, 443-^
international propagation of, 636 - and payment to capital, 444—5
Depression, Great, 829, 832 and payment to land, 445-7
deterministic hypothesis, 10 economist, role of, 518-22
deterministic relations, 24—5 education, 514
devaluation, 755 as a contribution to growth, 806
effect on balance of payments, 762-3 in underdeveloped economies,' 819
competitive, 784 efficiency, technical and economic, 231
Dicks-Mireaux, 727
L., 725, demand elasticity, in-
elasticity, see under
differential calculus, 31,38-9 come elasticity, etc.
differentiation, minimum, 387-90 employment levels and the paradox of
diminishing marginal utility, 179f thrift, 586-7

of income, 192-4 employment, full, 831-2

diminishing return, 272 equal net advantage, 428-9
discount houses, 689 equations and identities, 36-7
disposable income, 595 equilibrium
distribution, theory of, 63-4, 405-7 price in perfect competition, 99-104
criticisms of, 478-80 short-run, 308-9
disutility, 161 long-run, 309-12
dividends, 237 and changing demand, 360
dollar problem, 793 in monopolistic competition, 337-9,
Dow, C., xxi, 725, 727 360
duopoly, 343 in monopoly, 318-20
durables and nondurables, 602 equilibrium analysis
dynamic analysis comparative static, 153-4
154
definition, dynamic, 155-6
planned changes in supply, 155-6 partial, 499-501
general, 501-2 .

equilibrium differentials, 436-7

Economic activity, fluctuations in the equity capital, 239
, 618-30
level, error margins, 24—5
economic development and population, and statistical analysis, 52
824-5 European Economic Community, see
economic experience, international, 778f Common Market
economic growth, 65, 799f European Payments Union, 788
and production possibility boundary, excess capacity theorem, 339
67-8 excess demand, 100, 10^7, 150
cumulative nature of, 802 excess supply, 100
as a policy goal, 804-5 exchange control system, 764
866 INDEX

exchange rate, 741, 747f fiat currencies, 672

definition of, 748 firms
theory 753-63
of, in market behaviour, 77
changes in, caused by price changes, decisions of, 226, 264—7
753 ^ raising of capital, 239-40
equihbnutn, 756 8 goals of, 94
and changes in taste, 758 theory of, 370-1, 39lf
fluctuating and fixed, 760-!, 764-5 criticisms of theory of, 392f
unstable, 761 3 demand for loanable funds,
472-4
and national pride, 763 government, 647-57
fiscal policy,
and the gold standard, 780 definition of,648
and short term disequiltbna, 790 2 builtm stabilisers, 653
and long-term disequiltbna, 792 dynamic problems of stabilisation,
fixed versus fluctuating, 794 655
Canadian, 795 flows, 36
exchange reserves, 765 demand and supply, 80, 94
expectations self realising, 625 foreign exchange reserves, 638
expenditure foreign trade multiplier, 632, 633-6
and the circular flow of income, 559 forty-five degree line, 569f
and prices, micro economic implica- free goods, 529
tions of aggregate relations between, and the circular flow of income, 560
717 19 free markets, 65
switching 643 -6 producers and consumers in, 74
dampening, 843-6 allocation of resources in, 150
exploitation doctrine of trade, 7C8 case for, 506-7
Export Import Bank, 822 case for intervenlion in, 507-16
exports free trade
eflect on national income, 583-4, 632 versus tariffs, 769-71
good or bad, 637 and unemployment, 771
of demand, 754 5
elasticity Fnedman, Milton, 603, 714
domestic supply price of, 758 full cost pricing, 395-6
function distribution of income, 406
Factor incomes, 549 50 funclional relations, 22-4, 28
factor movements, 508 9 comparisons, 29
factors of production implications, 30-1
demand for, 408-10 terminology, 33-4
demand curve for, 415 16 notation, 37-8
supply of, 410 11, 422 maxima and minima, 41
supply elasticity, 429 30 of more than two variables, 41-2
influence on demand for, 418-21
allocation of, 428-9 General equilibrium analysis, 501-2
mobility of 429 30, 490 3 508 9 General Agreement on Tariffs and
factor prices, 265 72, 411, 414 Trade, 787
theory of, 434f George V, King, 587, 861
dynamic differentials, 436 geometric analysis, 19 20
equilibrium differentials, 436-7 George, Henry, 450
and market conditions, 486-90 Giffen case, 201, 212
factor services, 549 gold standard, 671-3, 779, 780-2
Federal Reserve System, 684 Gordon, R A , 857
Ferber, Robert, 221 government expenditure, 583-4, 641-2
 1

INDEX 867

government tax revenue, 641 diminishing marginal utility of, 192-5

government, size of, 857-60 macro-distribution of, 490
graphs, 26 equilibrium level of, 575-82
and functional relations, 29 graphical representation, 578- 9
and time series, 57 distribution, aggregate level of con-
Great Depression, 620, 783-5 sumption expenditure, 595
gross investment, 556 and consumption, 603-4
total,
gross national income, 552-3 disposable, and consumption, 603 4
gross nafionaJ product, 551, 332 income effect, !96
growth, see economic growth income elasticity of demand, 1 18-20
and increases in real income, 167f
effect on agriculture, 149
Harrod, Sir Roy, 534
and changes in demand, 215
health, and growth in underdeveloped
income statement
countries, 819
accountant’s, 258
Hines, B., 724
economist’s, 259
household
income tax, 527
deRnition, 76-7
graduated, 516
demand of, 80-81
increasing returns to scale, 269
expenditure of, 167f
indifference curs'es, 177, 199
income and price changes, 1 73
inefficiencies, quantitative significance,
behasdour of, 182, 184-5
64
observ'ed, 204-5
infant industry', 774-5
demand for loanable funds, 471
inferior goods, 86
diversibility of, 510-1
inflation, 156, 700, 723-30, 744, 759
household consumption, 594f
inflationary gap, 647, 65!
housing, local council, 509
injections into circular flow of income.
Hume, David, 733
554f
Hutchinson, T. W., 855
566
total,
hydroelectric power, 512
graphical representation, 569-74
hs-potheses, inductive and deductive,
and withdrawals, compensating shifts
54
in, 587
injection function, effect of a shift in, 583
Identities and equations, 36 innovations, 381-^1
imperfect competition, see monopolistic definition of, 289
competition rate of, 811
imputed cost, 249 inputs, 229
imports, 631—2 value of, 247
financing 632-6
of, and productive function, 261
good or bad, 637 factor combinations, 271-2, 279
and unemployment, 638-40 input-output analysis, 503^
elasticity of demand, 754-5 International Bank for Reconstruction
incentive restructuring, 527 and Development, 786, 822
income International Development Association,
and demand, 81, 86-7, 89, 210-12 787
changes 118-20, 202
in, International Finance Corporation, 787
in agriculture, 142 International Monetary' Fund, 785-6,
and household expenditure, 169 791
real and money, 174 international payments, 788-96
distribution of, 89, 405-7, 518 interest, nature of, 470, 479
 1

868 INOFX

interest rate and economic rent, 439-40

and market price, and competition, 453
and demand for fondi, 47f-5 and monopoly, 453 6
as price of funds, 475 6 and employers’ associations, 157
as income of lenders, 47G 7 and Unions, 458f
and savinfts and inscstmeni, G07 17 eaminRs and market condition, 486
and restoration of eriuihlirium, GOA 10 mobitiiy and earnings, 490 3
inventions, 289 91 costs and exchange rate, 745 6
as a contribution to growth, HOG labour force, definition, 550
inventories, 551, 55(i lan<]
inventory c>cle, 62H as natural resource, 58
investment as factor of production, 230
and circular How of income, 556- 7 supply of, 425'6
Gnancing, 606 mobility of. 429*30
determinants, 607 17 and economic rent, 415-7
and rate of interest, 607 13 e/Tect of taxes on, 449-71
and the level of income, G13 14 land tenure, 807
accelerator theory of, 627 "laws"
rale of, 81 of large numben, 12-13
investment credit, 528, 810 of supply and demand, 101-7
investment rTisendiiure, 427. 583 4 of reitini, 270
invisible trade, 731 2 legil institutions and growth, 805
(.eibenstem, Harvey, 821
I eontief. IS W503 ,

Jacobi, Paul, 829 l«ewis, J Parry, 31

Johnson, Harry. 772 limited liability, 235-6
Johnson, President 861 linear functions, 38 9
Johnston, J , 270 lapsey. R C 725, 726, 773. 781
.

joint stock companv, 237 fl

last, j-redenck, 731
and capital, 239 40 living standards, and tariffs, 774
ownenhip, 2nr loanable funds
demand, 471-5
Kaldor, N 491 supply, 475
,

Kalecki, 491 pnee, 175 7

Kennedy, John F M, 8l»l lock out. 459
.

Keynes. John Mavnard. 62, OlO 611,

612 13. 6M, 706 7. 714 Iff. 782
791, 792. 812. 855. 860 Nfacro-disinbuiion of income, 491
Khrushchev, N 805 ,
macro-economin
knowledge, as a contribviivon lo growth, distinction from micro, 537—11
506 basis of, 539-10
Kopf, David, H 6M ,
problems of, 511-2
tnacro-\ ambles, 538, 510, 836, 856
marginal analysis, 393 91
L-shaped function, 702-4, 717 19 marginal cost, 267-8
labour curve, 400-1
human resources, 58 marginal physical product, 417-18
as factor of production, 210 marginal produciiviiy
supply 418 2I
of, theory of, 482 85
mobility of, 427-9 and justice, 185 6
 INDEX 869

marginal productivity — continued reasons for holding, 694—7

and functional distribution of income, determinants of demand for, 697-9
493-4 quantity theory' of, 706-7
see also, land, labour, capital Keynesian theory of, 706-7, 714-16
marginal propensity and aggregate demand, 705-16
to consume, 596-7 monopolistic competition, 335f
to import, 634, 636 equilibrium in, 337-9, 340
to save, 636 and excess capacity, 338-9
marginal revenue product, 415 nonprice competition, 340
curve, 417 theory 341-2
of,

for competitive firm, 418-19 costs in,364—5

for industry, 419-21 response to market signals, 509
marginal utility, 177-81 and nationalisation, 533
and price, 183 monopolistic discrimination, see price
and elasticity of demand, 186 discrimination
and variable prices, 188-9 monopoly
of free goods, 529-3 demand curve, 315-16
market demand, 87-9 marginal revenue curve, 316-17
market imperfections, 507 equilibrium, 318-20
market price, 74, 80 conditions of, 322-24
market signals, 506, 507, 509 comparison wath perfect competition,
Marshall Plan, 787-8 372-3, 379-83
mathematical analysis, 20-4 policy, 378-83
use in economics, 25-6, 28, 32 and diversity of product, 366-7
Mercantilists, 733 and market behaviour, 505-6
micro-economic policy, 505-36 labour, 453-4
micro-economics, distinction from bilateral, 456
macro, 537-41 monopsony, 454—5
micro-variables, 538 multiple regression analysis, 50
minimum differentiation, 387-9 and demand, 215
minimum-wage laws, 517
Mises, L. von, 218
mixed economies, 79 National debt, 642-7
money national income
income, 174 predictions of theoiy', 583
demand for, 471-75 *
effect of shift in injection function,
nature and history, 665f 581-4
as a medium of exchange, 666-7 effect of shift in withdrawal function,
as a store of wealth,667 584-5
as a unit of account, 667-8 equilibrium, 587-9
metallic, 668-70 and investment, 613-14
quantity theory of, 670 rate of change, 614—17
paper, 670-4 stabilisation, 626
deposit, 674-5 and foreign trade, 631 f
near, 676 money, 801
supply of, 677f real, 801
operational definition of, 675 national pride and exchange rates, 763
determination of theory of supply, national product, distribution of, 63
690-2 national resources, and growth in under-
demand for, 694f developed economies, 819
870 INOFX

n'lJionaluation, 532-5 price equilibrium m, 99-104

and profit maximuaiion, 229 demand curve of firm, 296-7
natural rfsotirccj, 511 demand curve of market, 299
suppl) of, 42G-7 elasticities of, 297-9
near money, 602-3 revenue curve, 301-2
necessary condition, profit maximisation, 302-5
ncffaiise feedback, 655 6 supply curve of firm, 305-7
net in\mment, 557 supply curve of industry, 306 7
net national product, 551 iliort-nm equilibrium, 308-9
neutral equilibrium, 553 long-run rquilibnum, 309-10
new investment, 55G changes in demand, 353-B
nonprice competition, 335, 340 changes m
costs, 363-5
normal profit, 252 effect of taxes, 365 8
normative statements 4 7 and allocation of resources, 374-7
compared with monopoly, 378 81
OliROpolv and diversity of product, 386-7
comp.insons, 313 perfect divasibility, h>polhesis of, 281
joint profit maximisation. 315 51 perfect monopol), 325-8
and differentiation, IBS Phillips, A \V,C56, 727, 781
limitations of price theory, 101 2 Pliilhjis cone, 701-5
open-market operaiioiis, 68G l*h>'siocraii, 733
open shop, 459 picket lines. 459
opportunity ctMt, t»0 I, 67. 730. 73B 818 57
policv conflicts,
and household expemlilure. 169 516 22
{Milicy decisions,

and budget line, 171 523r

|)olicv tools.

and economic costing, 248 Popper K F 52 .

calculation of 25(4 I population, 89, R21

of capital goods, 616 positive statements, 4 7
of holding cash, 698 possibilit) ihrorem, 768
and prices, 741 3 (wivrrty, vicious circle of, 821
ofgrouth 808 pretJiciion in economics, 10, 519-20
optimal fictor combinations 272 i preference, 158
Orciiti. G » , 540 income and sulistimiion effect, 195-8
organisation theory, 396- 7 prestige, 835
outputs, 229 price
value of. 247 iheorv of, 63
and production function, 261 \ahdil) of, 75
and cost function, 263 uses of. 126 7
ovsnership control, 242 3 and scarcity of rrsoiircci, 234
and demand, BO, 82
Pauh, Professor frank, 657 eqoihbrritm, fO), 102
Paradox of thrift, 585 7 disequilibnum, 100-1
Parnes, H S , 8J0 analysis of, 153-4
partial equilibrium analjxis, 499 501 changes and housebold expenditure,
partnership, 236-7 153
and 239 40
capital, legssliiion, 127-33
patent laws, 382-3 effect of taxes, 136-7
patent monopolj, 384-5 fluciuaiions in agriculture, 139-40
perfect competition absolute and rehtive, 172, 175
and producers, 74 elasticity, 215
 INDEX 871

price — continued productive capacity, 799

and expenditure, micro-implications productivity of labour, 67f
of various aggregate relations be- increases in, 148, 289-92
tween, 717-19 and supply, 288-9
and opportunity cost, 741-3 predictions, 291-2
and exchange rate changes, 753-4 profit
and unemployment, 849-52 reinvestment of, 240
price discrimination, 328 definition, 248, 252
conditions (or, 329 normal, 252
effects of, 330-1 as interest, 470, 479-80
unsystematic, 334 and national income, 481
price level, 175, 830 1 profit and nationalised industries, 532
determination of, 700 profitmaximisation, 94
money, aggregate demand and, 700 motivation of firms, 227-9
rate of change, 701 and ownership, 227-9
effect of changes in one or two coun- and cost minimisation, 264
tries, 759 in perfect competition, 302-5
control, 838-42 in monopoly, 318-20
price stabilisation in oligopoly, 345-51
by producers’ associations, 142-4 criticisms of, 393-94, 397
and government policy, 127-33 long run, 399-400 *

and re-aliocation of resources, 150 and allocation of resources, 506

price system progressive tax, 524
and allocation of resources, 71-4 proportional tax, 524
and equilibrium, 100-1 protection of inditdduals from decisions
dynamic analysis of, 75-7 1 taken on their behalf by others,
as control mechanism, 181-2 514-15
and allocation of scarce commodities, proprietorship, 235-7
133-4 and capital, 239-40
determining trade according to com- public ownership, see nationalisation
parative advantage, 741-6 public utility regulations, 379
price takers, 297, 311, 360 purchasing power, 174
private cost, 253, 533
difference from social cost, 510-12
private revenue, "506, 533
Quantity theory of money, 706-14
difference from social revenue, 510-12 demand for and supply of money,
producers’ associations
707-8
in agriculture, 142
excess demand for money and aggre-
and restriction of output, 360- 3 gate demand for commodities, 709
and maximisation of profit, 363 velocity and balances, 709
,

and restriction of entr^’, 368-70 disequilibrium behaviour, 710

predictions of, 711
production
theor>' of, 225
and long-term changes in price level,

rates, 227
712-13
factors of, 229-31 and short-term changes in price level,

and 713
substitution, 233-4
and quotas, 148, 361
scarcity, 233-5
rate of growth, 8 1
production function, 261 Random sample, 47-8
production-possibility boundary, 65-8 rationing, 130
072 INDI X

ratchet efieit. 703 Said

real jncome, 174 and production, 227
recession, 023 maxtmisiiion, 3'28 9
reco\cr>. 022 sample, 47
reditcount rate, ite bank rate sstisficing, 3‘)7 8
Reel, An>eri, 467 tasings, COSr
regre^Aion anaU'SM, 50 and
circular flow of income, 556
regtessise tax. 524 deiernunant*. fi07-17
*
reluive prices, 172 and the inlerni rate, 607-10
and allocation of resource*. 474 SaHyer, 55 W , 31
rent control 518 20 scarce commodilie*
replacement intestment, 550 alioeatinn, 133 4
279
replication, (iHtnbulion tbror),4ll
reserxe reqmremenii, 005 0 scalier diagrams, 48 9, 55-6
resource* Schultz. nenr>, 215
definition of 58
allocation of.M) I.OI Ot S 011 15 economics as, 4
emplo>mfnt of 01 2 and r\ndener, 7 9
and price itatem, 71 2 and ihfories, 13 14
and iniame chanKCs ISO firrsliciions in, 15 16
and rtlaiixe prices |7f 5 and economics, 17 18
and loanable fund* >77 9 Sciinxsky, 1 , 772
and free market, 502 3 secular stagnation, hspoihesis of, 657
gain* from re'alloeaiion 73> ihareholders. 23')
and ehangn m prwliirtion 737 thori run
returns, Ian of. 2(i9 70 defnilion. 266
return to cost 279 cost curve, 270 4
pecunian 285 7 bimon, ii A , 397
return to scale, 26'> single>ias niotement 450 1
return to subsiituiion 282 5 brniili, 372. 733. 731. 818
Adam,
rrtealed preference (lieort 177 Smyth, D617 ,

social tost, 253 -4, 533

and ehsticits of demand, 1 1<> 17 dinercnte from pnvate cost, 510 12
in agriculture, 141 2 social habits and growth, 806
tlieor> of, 247 social ubbgaiiom, 515
total 301 social revenue, 510 11,515
aterage. JOl special deposiu, 686
marginal 301 standard of living, domestic, ti38
of mnno]Kilisis 3tl> 18 static crjuibbnum analysis, 153-4
pritate and sorni 510 12 slaiistical analvsis
collection cost* of, 513 idttng of theonei, 46-7
srt also i.nxes errors in, 52-3
Ricardo Duid, 01. 425 438, 733. 744 and qualification, 53-4
risk-taking. 24') iiatitiical hypotheses, 10 It

as cost.252 universal, 51
Robbins, I, , 218 rejection of, 52
Robinson, n W
432 ,
Sieuer, M D . 726
Rooseseli, President Franklin D . 860 stock, common and preferred, 239
Rostow, \V \\ ,
824 stfKks and flows, 3o
rule-making. 523 siockholdcn, 239
 INDEX 873

definition, 365
Stock Market, 241
unit tax, 365
Stone, Richard, 215
strikes, 459 long-run tax, 366-7
substitutes profits tax, 367-8
and demand, 81, 84-5 and 424-5
effort,

and shifts in demand curves, 91 on land, 449-51

substitution effect, 196 collection costs, 513-14

non-negative, 197-8 excise, 528

* sales, 528
sufficient condition, 34
supply and the circular flow of income, 559
definition, 94 and consumption function, 604
and price, 94-5, 96 progressive, 654
function, 96 techfiology
excess, 100 and supply, 95
as flow, 80, 94 quality of innovation, 813
elasticity of, 135, 140 and growth in underdeveloped econo-
•
shifts in, 73-4 mies, 820
in agriculture, 140f technical efficiency, 231
and market stability, 157-8 theories
and competition, 162-3 testing of, 43f
and foreign currency, 753-7 uses of, 221
supply curve, 96-7 Thompson, David, 534, 587, 782
shifts in, 97-8 time deposit, 675
and demand curve, 11-12 time-series data, 600
and price equilibrium, 103-4 total utility,158-9
of competitive firm, 301-2 maximisation, 181
of competitive industry, 302-3 limitation of concept, 189-91
under monopoly, 316-17 trade
of factors of production, 437-8 and the circular flow of income,
supply lag, 153 557-8
and planned changes in supply, 156 and national income, 63 If
surplus financing, 641-7 gains from, 733f
visible and invisible, 751-2
exploitation doctrine of, 768
Taft-Hartley Act, 433 andtariffs, 768
tariffs mutually advantageous, 773f
and the gains from trade, 768f trade cycle theory, 62, 623, 626-8
prohibitive, 769 Trade Unions
existing level and free trade, 769-71 organisation, 458
case for, 773-7 success of, 462
and non-output-maximising goals, and wages, 463-4, 466-8
755-7 and benefits, 464—5
tastes
and security, 465-7
and demand, 81, 87 468
necessity of,
and shift in demand curve, 91 transfer earnings,437
testing of, 206-9 and economic rent, 439-42
Tawney, R. H., 806 transfer payments, 443-6, 528-9, 560
tax
transport
and supply, 134-6 road versus rail, 511
and price, 136-9 motorways, 514

/ .
 1 1

874 INDEX

transport —conlinutd variable proportions, lawof, 274

and growth in underdeveloped econo- variables,22-3

mics, 820 magnitude of changes, 31
Tnffin Plan, 791-2 dependent and independent, 34 3
exogenous and endogenous, 35-6
Unavoidable balances, 752 dating of, 582
underdeveloped economics, 8I6f Vcblcn, Thorstein, 213
definition, 817 18 velocity of circulation of money, 710
growth in, 818 20 Verdoorn, P J ,
772
undistributed profits, 550 visible trade, 751 2
unemployment
of resources, 61 2, 67
of labour force, 550 Wages
in Great Britain, 618-21 increases, 463-4
and trade, 638*40 and benefits, 464- 5
and free trade, 77 and security, 465-6
structural 812 15 and unions, 466-9
frictional 830f wage drift, 841
causes 837 wage laws, 149
control 837 8 Walras, Leon, 501
and prices 849 52 wartime economy, 645-6
and balance of payments, 852-5 welfare economics, 64, 374-5
United States Wclmesfelder, , W
772
defence expenditure 659 in, withdrawals from ctrculai flow of in-
structural unemployment, 813 15 come, 554f
unlimited liability, 236 total, 566
utility, ste marginal utility, total utility graphical representation, 569-74
Wold, Herman, 215
Value paradox, 187-8
value judgments, 5 Zoning laws, 51