Вы находитесь на странице: 1из 97


to accompany

Ehrenberg & Smith

Modern Labor Economics:

Theory & Public Policy

Eighth Edition

Robert S. Smith
Cornell University

Robert M. Whaples
Wake Forest University

Lawrence Wohl
Gustavus Adolphus University

Copyright 2003 Addison-Wesley, Inc.

All rights reserved. Printed in the United States of America. No part of this book may be used
or reproduced in any manner whatsoever without written permission from the publisher, except
testing materials and transparency masters may be copied for classroom use. For information,
address Addison-Wesley Higher Education, Pearson PLC 75 Arlington Street, Suite 300,
Boston, Massachusetts 02116.

This Instructor's Manual is intended to summarize the content of the eighth edition of Modern
Labor Economics: Theory and Public Policy in a way that explains our pedagogical strategy.
Summarized briefly, we believe that labor economics can be best learned if students are (1) able
to see the "big picture" early on, so that new concepts can be placed in perspective; (2) moved
carefully from concepts they already know to new ones; (3) motivated by seeing the policy
implications or inherently interesting insights generated by the concepts being taught. To this
last end, we discuss policy issues in every chapter and, in addition, employ "boxed examples" to
demonstrate in historical, cross-cultural, or applied managerial settings the power of the
concepts introduced.

The text is designed to be accessible to students with limited backgrounds in economics. We do

employ graphic analyses and equations as learning aids in various chapters; however, we are
careful to precede their use with verbal explanations of the analyses and to introduce these aids
in a step-by-step fashion. To help students in the application of concepts to various issues, we
have printed answers to the odd-numbered review questions for each chapter at the back of the

We have also endeavored to put together a text that, while accessible to all, is a comprehensive
and up-to-date survey of modern labor economics. There are 9 chapter appendices designed to
be used with more advanced students in generating additional insights.

In the first part of this Instructor's Manual, we present a brief overview and the general plan of
Modern Labor Economics. We then present a chapter-by-chapter review of the concepts
presented in the text. In the discussion of each chapter we list the major concepts or
understandings covered, and in some cases suggest topics or sections that could be eliminated if
time must be conserved. We also present our answers to the even-numbered review questions at
the end of each chapter.

An important part of this Instructor's Manual are the suggested essay questions related to each
chapter. We present a few suggested essay questions for each chapter.
Table of Contents
Click on the chapter title to jump directly to that page.

Overview of the Text 1

Chapter 1 Introduction 4
Chapter 2 Overview of the Labor Market 8
Chapter 3 The Demand for Labor 14
Chapter 4 Labor Demand Elasticities 20
Chapter 5 Quasi-Fixed Labor Costs and Their Effects on Demand 26
Chapter 6 Supply of Labor to the Economy: The Decision to Work 32
Chapter 7 Labor Supply: Household Production, the Family, and the Life Cycle 37
Chapter 8 Compensating Wage Differentials and Labor Markets 45
Chapter 9 Investments in Human Capital: Education and Training 51
Chapter 10 Worker Mobility: Migration, Immigration, and Turnover 57
Chapter 11 Pay and Productivity 62
Chapter 12 Gender, Race, and Ethnicity in the Labor Market 68
Chapter 13 Unions and the Labor Market 77
Chapter 14 Inequality in Earnings 84
Chapter 15 Unemployment 88


Chapter 1 - Introduction
Appendix 1A - Statistical Testing of Labor Market Hypotheses
Chapter 2 - Overview of the Labor Market

Chapters 1 and 2 introduce basic concepts of labor economics. They are written to be
accessible to students without backgrounds in intermediate theory, and can, therefore, be
used as building blocks when a professor must "begin at the beginning." If the course is
being taught to economics majors with intermediate microeconomics as a prerequisite,
these chapters may be skipped or skimmed quickly as a review.

An appendix to Chapter 1 introduces the student to econometrics. The purpose of this

appendix is to present enough of the basic econometric concepts and issues to permit
students to read papers employing ordinary least squares regression techniques. We
strongly recommend assigning Appendix 1A in courses requiring students to read
empirical papers in the field. We also recommend (in footnote 3 of the appendix) an
introductory econometrics text that could be assigned by instructors who wish to go
beyond our introductory treatment.


Chapter 3 - The Demand for Labor

Appendix 3A - Graphic Derivation of a Firm's Labor Demand Curve
Chapter 4 - Labor Demand Elasticities
Appendix 4A - International Trade and the Demand for Labor: Can High-
Wage Countries Compete?
Chapter 5 - Quasi-Fixed Labor Costs and Their Effects on Demand

The demand for labor is discussed first primarily because we believe that the supply of
labor is a more complex topic in many ways. Before analyzing the labor/leisure choice
and household production, we first introduce students to the employer side of the market.
For instructors who desire to cover topics concerned with the decision to work first,
however, we note that Chapters 6 and 7, which deals with that decision, are self-
contained. Therefore, nothing would be lost if Chapters 6 and 7 were taught ahead of
Chapters 3, 4, and 5.

In Chapter 3 the principal question analyzed is why demand curves slope downward. In
Chapter 4 we move to a discussion of the elasticity of demand, and analyze the
determinants of the precise relationship between wages and employment. The concepts
are used to analyze how technological change and foreign trade affect labor demand.
Finally, Chapter 5 analyzes the quasi-fixed nature of many labor costs and the ways these
costs affect the demand for labor.


Chapter 6 - Supply of Labor to the Economy: The Decision to Work

Chapter 7 - Labor Supply: Household Production, the Family, and the Life Cycle

Chapters 6 and 7 analyze the decision of an individual to work for pay. The traditional
analysis of the labor/leisure choice is given in Chapter 6, while in Chapter 7 the decision
to work for pay is placed in the context of household production. The essential features
of the decision to work for pay are included in Chapter 6. In one-quarter courses or
courses in which time is scarce, Chapter 7 could be skipped; however, doing so would
eliminate analyses of family labor supply decisions as well as labor supply decisions in
the context of the life cycle.

As noted above, Chapters 6 and 7 are designed to be self-contained for the convenience
of instructors who wish to teach labor supply ahead of labor demand.


Chapter 8 - Compensating Wage Differentials and Labor Markets

Appendix 8A - Compensating Wage Differentials and Layoffs
Chapter 9 - Investments in Human Capital: Education and Training
Appendix 9A - A "Cobweb” Model of Labor Market Adjustment
Appendix 9B - A Hedonic Model of Earnings and Educational Level
Chapter 10 - Worker Mobility: Migration, Immigration, and Turnover

Once they have decided to seek employment, prospective workers encounter important
choices concerning their occupation and industry, as well as the general location of their
employment. Chapters 8 through 10 analyze these choices, with Chapters 8 and 9
focusing on industry/occupational choice and Chapter 10 on the choice of a specific
employer and the location of employment. More particularly, Chapter 8 presents an
analysis of job choice within the context of jobs that differ along nonpecuniary
dimensions. Chapters 9 and 10 analyze issues affecting worker investments in skill
acquisition (Chapter 9) and job change (Chapter 10), and both employ the concepts of
human capital theory. All three chapters contain appendices of interest to instructors who
wish to teach more advanced material.


Chapter 11 - Pay and Productivity

Chapter 12 - Gender, Race, and Ethnicity in the Labor Market
Appendix 12A - Estimating "Comparable Worth" Earnings Gaps: An
Application of Regression Analysis
Chapter 13 - Unions and the Labor Market
Appendix 13A - Arbitration and the Bargaining Contract Zone

Chapter 14 - Inequality in Earnings
Appendix 14A - Lorenz Curves and Gini Coefficients
Chapter 15 - Unemployment

Having presented basic concepts and analytical tools necessary to understand the demand
and supply sides of the labor market, we now move to analyses of special topics:
compensation, discrimination, unions, inequality, and unemployment. A complete
analysis of all these topics requires an understanding of behavior on both the demand and
supply sides of the market, and these chapters are built upon the preceding ten. No new
analytical tools are introduced in these chapters.

The chapters on unionism (Chapter 13) and discrimination (Chapter 12) deal with issues
typically covered in labor economics courses, but they are more comprehensive than
most other texts. It should be noted that the appendix to Chapter 12 includes an
application of regression analysis. The chapter on inequality is unique and can be skipped
without a loss in coverage of conventional material; however, it is written in a way that
provides a review of material in previous chapters.

Because the textbook stresses economic analysis as it applies to the labor market,
students must understand the ways economic analyses are used. The basic purpose of
Chapter 1 is to introduce students to the two major modes of economic analysis: positive
and normative. Because both modes of analysis rest on some very fundamental
assumptions, Chapter 1 discusses the bases of each mode in some detail.

In our treatment of positive economics, the concept of rationality is defined and

discussed, as is the underlying concept of scarcity. There is, in addition, a lengthy
discussion of what an economic model is, and an example of the behavioral predictions
flowing from such a model is presented. The discussion of normative economics
emphasizes its philosophical underpinnings and includes a discussion of the conditions
under which a market would fail to produce results consistent with the normative criteria.
Labor market examples of governmental remedies are provided.

The appendix to Chapter 1 introduces the student to ordinary least squares regression
analysis. It begins with univariate analysis, introduced in a graphical context, explaining
the concepts of dependent and independent variables, the "intercept" and "slope"
parameters, the "error term," and the t statistic. The analysis then moves to multivariate
analysis and the problem of omitted variables.

List of Major Concepts

1. The essential features of a market include the facilitation of contact between buyers
and sellers, the exchange of information, and the execution of contracts.

2. The uniqueness of labor services affects the characteristics of the labor market.

3. Positive economics is the study of economic behavior, and underlying this theory of
behavior are the basic assumptions of scarcity and rationality.

4. Normative economics is the study of what "should be," and theories of social
optimality are based in part on the underlying philosophical principle of "mutual
benefit. "

5. A market "fails" when it does not permit all mutually beneficial trades to take place,
and there are three common reasons for such failure.

6. A governmental policy is "Pareto-improving" if it encourages additional mutually

beneficial transactions. At times, though, the goal of improving Pareto efficiency
conflicts with one of generating more equity.

7. The concept that governmental intervention in a market may be justified on grounds

other than the principle of mutual benefit is discussed (for example, government

intervention may be justified on the grounds that income redistribution is a desirable
social objective).

8. (Appendix) The relationship between two economic variables (e.g., wages and quit
rates) can be plotted graphically; this visual relationship can also be summarized

9. (Appendix) A way to summarize a linear relationship between two variables is

through ordinary least squares regression analysis -- a procedure that plots the "best"
line (the one that minimizes the sum of squared deviations) through the various data
points. The parameters describing this line are estimated, and the uncertainty
surrounding these estimates are summarized by the standard error of the estimate.

10. (Appendix) Multivariate procedures for summarizing the relationship between a

dependent and two or more independent variables is a generalization of the univariate
procedure, and each coefficient can be interpreted as the effect on the dependent
variable of a one-unit change in the relevant independent variable, holding the other
variables constant.

11. (Appendix) If an independent variable that should be in an estimating equation is left

out, estimates of the other coefficients may be biased away from their true values.

Answers to Even-Numbered Review Questions

2. Are the following statements "positive" or "normative"? Why?

a. Employers should not be required to offer pensions to their employees.

b. Employers offering pension benefits will pay lower wages than they would if they did
not offer a pension program.
c. If further immigration of unskilled foreigners is prevented, the wages of unskilled
immigrants already here will rise.
d. The military draft compels people to engage in a transaction they would not
voluntarily enter into; it should therefore be avoided as a way of recruiting military
e. If the military draft were reinstituted, military salaries would probably fall.

Answer: (a) normative (b) positive (c) positive (d) normative (e) positive

4. What are the functions and limitations of an economic model?

Answer: The major function of an economic model is to strip away real world
complexities and focus on a particular cause/effect relationship. In this sense an
economic model is analogous to an architect's model of a building. An architect may be
interested in designing a building that fits in harmoniously with its surroundings, and in
designing such a building the architect may employ a model that captures the essentials
of his or her concerns (namely, appearance) without getting into the complexities of

plumbing, electrical circuits, and the design of interior office space. Similarly, an
economic model will often focus on a particular kind of behavior and ignore complexities
that are either not germane to that behavior or only of indirect importance.

Models used to generate insights about responses to a given economic stimulus are often
not intended to forecast actual outcomes. For example, if we are interested in bow
behavior is affected by stimulus B, with factors C, D, and E held constant, our model
may not correctly forecast the observed behavior if stimuli C through E also change.

6. A few years ago it was common for state laws to prohibit women from working more
than 40 hours a week. Using the principles underlying normative economics, evaluate
these laws.

Answer: Laws preventing women from working more than 40 hours per week essentially
blocked mutually beneficial transactions. There were women who wanted to work more
than 40 hours a week, and there were employers who wanted to employ them for more
than 40 hours a week. The restrictions upon their employment prevented these
transactions from occurring and therefore made both the women and their potential
employers worse off.

8. “Government policies as frequently prevent Pareto efficiency as they enhance it.”


Answer. Achieving Pareto efficiency requires the completion of all mutually beneficial
transactions. Ideally, government would step in to provide information is that is blocking
mutually beneficial transactions or to establish markets (or market substitutes) when
markets do not exist. However, governments also have power to prevent transactions or
distort prices, both of which can prevent the completion of mutually beneficial
transactions. Government regulations can outlaw certain transactions that the parties to
them would consider mutually beneficial (the text mentions laws that historically
prevented women from working more than 40 hours per week). Government also has the
power to distort prices by setting minimum wages, mandating premiums for overtime
work, and so forth.

Answers to Even-Numbered Problems

2. (Appendix) Suppose that a least squares regression yields the following estimate:
Wi = -1 + .3Ai, where W is the hourly wage rate (in dollars) and A is the age in years.
A second regression from another group of workers yields this estimate:
Wi = 3 + .3Ai - .01(Ai)2.
a. How much is a 20-year-old predicted to earn based on the first estimate?
b. How much is a 20-year-old predicted to earn based on the second estimate?

Answer: a. W = -1 + .3x20 = 5 dollars per hour.

b. W = 3 + .3x20 - .01x20x20 = 3 + 6 - 4 = 5 dollars per hour.

Suggested Essay Questions

1. Child labor is an issue that has been discussed a lot recently. From the perspective of
normative economics, explain the problem with child labor.

Answer: Pareto efficiency requires that transactions have mutual benefits, and this can be
assured only if the transactions are voluntary and take place with complete information.
Children may be compelled by their parents to work, and they have limited capacities to
make informed decisions even in the absence of compulsion.

2. A law in one town of a Canadian province limits large supermarkets to just four
employees on Sundays. Analyze this law using the concepts of normative economics.

Answer. There are no doubt large supermarkets that want to hire workers on Sundays
(because there are consumers who want to shop on Sundays), and there are no doubt
employees who could be induced – perhaps by higher wages – to work on Sundays. A
law preventing such work prevents a mutually beneficial transaction.

Our goal in this text is to move students along very carefully from what they do know to
the mastery of new concepts. It is our belief that students learn most efficiently if they
can associate these new ideas with an overall framework, and it is the purpose of Chapter
2 to provide that framework. This chapter has both a descriptive and an analytical
purpose. One aim is to introduce students to the essential concepts, definitions,
magnitudes, and trends of widely used labor market descriptors. To this purpose, the
chapter discusses and presents data on such topics as the labor force, unemployment, the
distribution of employment, and the level of (and trends in) labor earnings. The second
aim is to provide students with an overview of labor market analysis. To this end, we
discuss basic concepts of demand and supply so that students will be able to see their
interaction at the very outset.

We start the overview with a discussion of demand schedules and their corresponding
demand curves. Particular attention is given to the distinction between movement along a
curve and shifts of a curve. Distinctions between individual and more aggregated demand
curves are discussed, as is the distinction between short-run and long-run demand curves.
A similar discussion and set of distinctions are made for the supply side of the market.

After both the demand and supply sides of the market have been discussed and generally
modeled, we turn to the question of wage determination and wage equilibrium. Forces
that can alter market equilibria are comprehensively discussed, and the chapter's major
concepts are reinforced by discussions of the effects of unions, the existence of
disequilibrium, and the concept of being "overpaid" or "underpaid" (including a
discussion of economic rents). The chapter ends with a discussion of unemployment
across various countries.

List of Major Concepts

1. The labor market and its various subclassifications (national, regional, local; external,
internal; primary, secondary) are defined.

2. The "labor force" consists of those who are employed or who are seeking work, and
major trends in labor force participation rates are discussed.

3. The "unemployed" are those who are not employed but are seeking work (or awaiting
recall); trends in the unemployment rate are noted.

4. Changes in the industrial and occupational distribution of employment are facilitated

by the labor market, which also facilitates adjustments to the "birth" and "death" of
job opportunities.

5. The distinction between nominal and real wage rates is made, and the calculation of
real wages is illustrated.

6. Distinctions among wage rates, earnings, total compensation, and income are
depicted graphically.

7. The labor market is one of three major markets with which an employer must deal; in
turn, labor market outcomes (terms of employment and employment levels) are
affected by both product and capital markets.

8. The concepts underlying a labor demand schedule are associated with product
demand, the choice of technology, and the supply schedule of competing factors of
production; scale and substitution effects are ultimately related to these forces.

9. Underlying a supply schedule for labor are the alternatives workers have and their
preferences regarding the job's characteristics.

10. Distinctions between individual and market demand and supply curves are discussed.

11. Movements along, rather than shifts of, demand and supply curves occur when wages
of the job in question change; when a variable not shown on the graph changes, the
curves tend to shift.

12. The interaction of market demand and supply determines the equilibrium wage.

13. Changes in the equilibrium wage rate are caused by shifts in either the demand or
supply curves. Disequilibium will persist if the wage is not allowed to adjust to shifts
in demand or supply.

14. The concepts of "overpaid" and "underpaid" compare the actual wage to the
equilibrium (market) wage rate.

15. Individuals paid more than their reservation wage are said to obtain an "economic

16. The concepts of shortage and surplus are directly related to the relationship between
actual and equilibrium wage rates.

17. Unemployment rates, and especially long-term unemployment rates, have risen in
Europe relative to the United States and Canada over the recent decade; this rise may
reflect the existence of relatively stronger nonmarket forces in Europe.

Answers to Even-Numbered Review Questions

2. Analyze the impact of the following changes on wages and employment in a

given occupation:
a.) A fall in the danger of the occupation.
b.) An increase in product demand.

c.) Increased wages in alternative occupations.

Answer: (a) A fall in the danger of the occupation, other things being equal, should
increase the attractiveness of that occupation, shifting the supply curve to the right and
causing employment to rise and wages to fall.

(b) An increase in product demand will shift the demand for labor curve to the right
causing both wages and employment to increase.

(c) Increased wages in other occupations will render them relatively more attractive than
they were before and cause the supply curve to the occupation in question to shift to the
left. This will cause employment in this market to fall and wages to rise.

4. Suppose a particular labor market were in market-clearing equilibrium. What could

happen to cause the equilibrium wage to fall? If all money wages rose each year, how
would this market adjust?

Answer: Starting from the position of equilibrium, a labor market could experience a fall
in the equilibrium wage if either the demand curve shifts to the left or the supply curve
shifts to the right. While market wages are usually stated in nominal terms, their
relationship to the prices of both consumer and producer products is of ultimate
importance. Therefore, both parties to the employment relationship are, in the last
analysis, concerned with the real wage rate. The real wage rate can fall when the nominal
wage rate is rising if prices of consumer and producer products rise even more quickly.

6. How will a fall in the civilian unemployment rate affect the supply of recruits for the
volunteer army? What will be the effect on military wages?

Answer: Supply curves to a given occupation are drawn holding alternative opportunities
constant. If those opportunities become more attractive, the supply curve to the given
occupation will shift left and tend to drive up wages. Thus, a fall in the unemployment
rate will shift the army's supply curve to the left (there will be fewer recruits at each army
wage rate), and the army's wages will be driven up.

8. Suppose that the Consumer Product Safety Commission issues a regulation requiring
an expensive safety device to be attached to all power lawnmowers. This device does
not increase the efficiency with which the lawnmower operates. What, if anything,
does this regulation do to the demand for labor of firms manufacturing power
lawnmowers? Explain.

Answer: This regulation would cause the demand for labor curve of the firms that
manufacture power mowers to shift to the left. The demand for labor is in part derived
from product demand. Because it is more costly now to manufacture lawnmowers, the
prices that will be charged to consumers will rise. This price increase will move the firm
upward and to the left along its product demand curve. With less product demanded for
any given wage rate paid to workers, the end result is a leftward shift of the labor demand

curve. (If, however, consumer preferences for greater safety were to shift the product
demand curve to the right, employment losses would be mitigated.)

10. Suppose we observe that employment levels in a certain region suddenly decline as a
result of (i) a fall in the region's demand for labor, and (ii) wages that are fixed in the
short run. If the new demand for labor curve remains unchanged for a long period and
the region's labor supply curve does not shift, is it likely that employment in the
region will recover? Explain.

Answer: The initial response to a leftward shift in the labor demand curve in the context
of fixed wages is for there to be a relatively large decline in employment. This decline in
employment is larger than the ultimate decline in employment. The initial disequilibrium
between demand and supply in the labor market should force wages down in the long run,
and as wages decline firms will move downward along their labor demand curves and
will begin to employ more labor. However, employment in the region would recover to
its prior level (assuming no subsequent shifts in demand or supply curves) only if the
supply curve was vertical; if supply curves are upward-sloping, the declining wage will
cause some withdrawal of labor from the market and employment will not recover to its
prior level.

Answers to Even-Numbered Problems

2. Suppose that the supply curve for school teachers is Ls = 20,000 + 350W and the
demand curve for school teachers is Ld = 100,000 – 150W, where L = the number of
teachers and W = the daily wage.

a. Plot the demand and supply curves.

b. What are the equilibrium wage and employment level in this market?
c. Now suppose that at any given wage 20,000 more workers are willing to work as
school teachers. Plot the new supply curve and find the new wage and employment
level. Why doesn't employment grow by 20,000?

Answer: a. See the figure. Plot the Ld and Ls curves by solving for desired employment
at given wage rates. If W = 500, for example, employers desire 25,000 workers (Ld =
100,000 – 150x500); if W = 400, they would desire 40,000. Since the equation above is
for a straight line, drawing a line using these two points gives us the demand curve. Use
the same procedure for the labor supply curve.

b. To find the equilibrium, solve for the wage at which the quantity of labor supplied
equals the quantity of labor demanded: Ls = 20,000 + 350W = 100,000 – 150W = Ld.
Solve for W by adding 150W to both sides and subtracting 20,000 from both sides to
yield 500W = 80,000. Dividing both sides by 500 reveals that W = $160 per day.
Plugging W = $160 into both the labor demand and supply equations shows that L =
76,000 schoolteachers.

c. The new labor supply curve is Ls' = 40,000 + 350W. Setting this equal to Ld and
solving shows that W = $120 per day; L = 82,000 school teachers. Employment doesn't
grow by 20,000 because the shift in the supply curve causes the wage to fall, which
induces some teachers to drop out of the market.

Suggested Essay Questions

1. American students have organized opposition to the sale by their campus stores of
university apparel made for American retailers by workers in foreign countries who
work in “sweatshop” conditions (long hours at low pay in bad working conditions).
Assume this movement takes the form of boycotting items made under sweatshop

(a) Analyze the immediate labor market outcomes for sweatshop workers in these
countries, using demand and supply curves to illustrate the mechanisms
driving this outcome.

(b) Assuming that actions by American students are the only force driving the
improvement of wages and working conditions in foreign countries, what
must these actions include to ensure that the workers they are unambiguously
better off?

Answer. (a) The demand curve for low-wage workers in foreign countries shifts to the
left when the product demand for the apparel they made falls. This drives down wages
and employment (assuming a fixed supply curve). (b) To avoid the effects in (a),
students in the U.S. must be willing to buy the same quantity and quality of apparel at
higher prices – that is, they must be willing to pay a premium for apparel made by better-
paid workers.

2. Ecuador is the world’s leading exporter of bananas, which are grown and harvested
by a large labor force that includes many children. Assume Ecuador now outlaws the
use of child labor on banana plantations. Using economic theory in its “positive”
mode, analyze what would happen to employment and wages in the banana farming
industry in Ecuador. Use demand and supply curves in your analysis.

Answer. Outlawing child labor on banana plantations reduces the supply of labor to
these plantations, shifting the supply curve to the left. With a fixed demand curve, this
shift in the supply curve drives up wages and drives down employment.

This chapter studies the downward sloping nature of the labor demand curve. It begins
with a section that discusses profit maximization, and it moves deductively from the
assumption of profit maximization to the marginal conditions with respect to labor. These
conditions are expressed in simple mathematical terms, and they are also discussed
verbally. Additional insights into the marginal productivity theory of demand are
provided in a section discussing common objections to this theory of demand.

The analysis of demand begins with the assumption that both labor and product markets
are competitive; in this context, we first consider the short-run before moving on to the
long-run and the case with more than two inputs. Next we consider the demand for labor
when the product market is not competitive, and then move to an analysis of demand
when the labor market is monopsonized. In the latter context, we contrast the wage and
employment effects of "market" and "mandated" shifts in the labor supply curve.

The chapter concludes with a policy analysis of payroll taxes that demonstrates the
insights that can be derived from an understanding of the demand for labor. The principal
conceptual tool employed involves distinguishing between the wage rate employers pay
and the wages employees receive. When these two wages differ, one must be stated in
terms of the other for the demand and supply curves to be shown together. When a
payroll tax is introduced, one of the two curves must therefore shift, and there will be
related changes in both wages and employment.

The appendix to Chapter 3 is designed for students who feel comfortable using
microeconomic theory at the intermediate level. We derive the demand for labor
graphically using a two-factor model in both the long-run and short-run. Both
substitution and scale effects are graphically illustrated, and the assumptions underlying
the demand curve are more rigorously presented. Any instructors wishing to skip over the
appendix can do so without loss of concepts needed to understand the basics of the
demand for labor.

List of Major Concepts

1. The assumption of profit maximization by firms underlies the theory of labor demand.
The process of profit maximization requires considering small changes in inputs (or
outputs), and comparing the marginal revenue generated by an additional input with
its marginal expense.

2. The marginal product of labor is the added output generated by adding a unit of labor,
holding capital constant.

3. If markets are competitive, firms perceive prices as given.

4. The difference between the short-run and long-run depends on the fixity of capital.

5. The concept of diminishing marginal productivity is discussed.

6. The relationship between the demand for labor curve and the downward sloping
portion of a firm's marginal product of labor curve is analyzed.

7. The demand for labor can be stated in terms of either the real or the nominal wage.

8. The relationship between the demand curve of individual firms and the market
demand curve is briefly discussed.

9. Two principal objections to the marginal productivity theory of labor demand are
presented and discussed.

10. The conditions for profit maximization with respect to capital are relevant in the long-
run, and adjustments of capital to changes in relative prices generate substitution
effects on employment.

11. Generalizing to more than two inputs, the demand for one grade of labor is influenced
by the wages of other grades of labor.

12. The concepts of substitutes in production, gross substitutes, complements in

production, and gross complements are defined and related.

13. Product market monopoly affects the profit maximization conditions, and thus, the
demand for labor.

14. Monopsony in the labor market affects wages, employment, and the labor conditions
for profit maximization. The reason for this derives from the upper-sloping labor
supply curve facing the individual firm, and the resultant increase in the marginal
expense of labor above the wage rate.

15. In the context of monopsony, wage increases accompanying market shifts (to the left)
in the labor supply curve produce the conventional expectations of decreased
employment. Mandated wage increases, however, flatten the labor supply curve and
reduce the marginal expense of labor, leading to ambiguous expectations regarding
employment changes, at least in the short-run.

16. The imposition of payroll taxes on the employer will shift the demand for labor curve
(when drawn as a function of employee wages) to the left, causing worker wages
and/or employment levels to fall.

17. (Appendix) The graphical depiction of a production function is presented.

18. (Appendix) The demand for labor in the short-run is graphically derived.

19. (Appendix) The demand for labor in the long-run, showing both substitution and
scale effects of a wage change, is graphically illustrated.

Answers to Even-Numbered Review Questions

2. Suppose that the U.S. military is having difficulty recruiting volunteers and is
considering one of two options: raising pay or reinstating the draft system. Analyze
the opportunity costs of lost civilian production when volunteers are used as
compared to those associated with drafting civilians using some random method of

Answer. In choosing employers, pay is an important consideration. Thus, many of those

who choose a military job are those whose civilian job opportunities pay less than the
military. Conversely, many of those who choose to remain civilians are workers whose
civilian pay is higher than their military pay offer. Because profits are maximized when
workers’ marginal revenue productivities (MRPL) are equal to the wage (W), we can
assume that those with higher pay also have higher civilian MRPL. Thus, when society
relies on military volunteers, it will lose less civilian output than it would by drafting an
equal number of civilian workers randomly. (It should be noted that pay is not the only
consideration in choosing a job, and that workers are really trying to maximize utility.
Those who choose civilian life over the military will be those who would get the least
utility from performing military duties. If some of the latter are forced into the military,
there is also an opportunity cost to society of lost worker utility!)

4. Suppose that prisons historically have required inmates to perform, without pay,
various cleaning and food preparation jobs within the prison. Now suppose that
prisoners are offered paid work in factory jobs within the prison walls, and that the
cleaning and food preparation tasks are now performed by non-prisoners hired to do
them. Would you expect to see any differences in the technologies used to perform
these tasks? Explain.

Answer. When inmates were required to work without pay, their wage was essentially
zero – and we would expect that prisons to have adopted labor-intensive technologies
(using the argument inherent in equation 3.8c). When wages rise, the cost of expanding
output using labor becomes greater, and we expect prisons to adopt the use of more
capital in the production process.

6. Suppose the government were to subsidize the wages of all women in the population
by paying their employers 50 cents for every hour they worked. What would be the
effect on the wage rate women received? What would be the effect on the net wage
employers paid? (The net wage would be the wage women received less 50 cents.)

Answer: Consider a simple competitive labor market in which the demand and supply of
women are both expressed in terms of the wage received by women (which, in the
absence of any subsidy, is assumed to be equal to the wage paid by employers). Given
the demand curve, D0, and the supply curve, S0, market clearing wage and employment
levels will be W0 and E0, respectively.

Suppose the government now subsidizes employers by paying them 50 cents for every
hour women work. Viewed in terms of the wage received by women, the employers'
demand curve will shift up by exactly 50 cents (reflecting the fact that this amount will
be paid by the government). At the old market clearing wage received by women, W0,
the number of women employers want to hire, E2, exceeds the number who are willing to
work, E0. This puts upward pressure on the wage received by women, and this wage rises
until the excess demand for labor is eliminated. This equilibrium occurs at the wage rate
W1, and the employment level E1.

It is clear from the figure that the wage received by women increases by less than 50
cents as long as the supply of labor curve is not vertical (i.e., as long as labor supply is
responsive to wages). Indeed, the more responsive labor supply is to the wage rate, the
less the women's wage will rise. Since the wage paid by employers now equals the wage
women receive less the 50-cent subsidy, it is also clear that the wage paid by employers
declines (by 50 cents minus the increase in the wage women receive).

It is important to stress to students that one would reach identical conclusions if one
analyzed the subsidy in terms of the wage employers pay. If supply and demand curves
are drawn in terms of this variable, a 50-cent-an-hour subsidy for women would shift the
female labor supply curve down by 50 cents. At the old wage paid by employers, the
supply of female labor would now exceed the demand. Downward pressure would be
placed on the wage paid by employers and it would fall by less than 50 cents (as long as
labor supply was responsive to the wage). As a result, the wage received by women
would rise by 50 cents less the fall in the wage paid by employers.

8. In 1999, the U.S. Bureau of Labor Statistics reported that hourly compensation costs
per U.S. manufacturing worker were $19.20, while those in Mexico were $2.12.
Recognizing that the analysis leading up to equation 3.8c can be used to understand
the choices firms make between any two factors of production, explain why a
growing firm with facilities in both Mexico and the U.S. might still expand its output
using U.S. workers. (Hint: consider U.S. and Mexican workers to be substitute
factors of production.)

Answer. The profit-maximizing firm will choose to expand production in the least costly
way. To do so, it will continue to substitute one factor of production for another until the
costs of expanding production using the two factors are equal (see equation 3.8c). In
choosing between U.S. and Mexican workers, profit maximization means that firms will
substitute one for another until the ratio of their wages to their marginal productivities are
equal. Mexican wages may be much lower than in the U.S., but if the relative marginal
productivity of Mexican workers is even lower, firms would decide to expand output
using U.S. workers. Put differently, even though wages are lower in Mexico, the ratio of
wages to marginal productivity – which is the critical datum – could be higher there than
in the U.S.

Answers to Even-Numbered Problems

2. The marginal revenue product of labor in the local saw mill is MRPL = 20 - .5L,
where L = the number of workers. If the wage of saw mill workers is $10 per hour,
then how many workers will the mill hire?

Answer: The mill will hire workers until MRPL = W. 20 - .5L = 10 when L = 20

4. The output of workers at a factory depends on the number of supervisors hired (see
below). The factory sells its output for $.50 each, it hires 50 production workers at a
wage of $100 per day, and needs to decide how many supervisors to hire. The daily
wage of supervisors is $500 but output rises as more supervisors are hired, as shown
below. How many supervisors should it hire?

Supervisors Output (units per day)

0 11,000
1 14,800
2 18,000
3 19,500
4 20,200
5 20,600

Answer. The firm needs to compare the marginal cost to the marginal revenue of hiring
an additional supervisor. The marginal cost is always $500 for each extra supervisor.
The marginal revenue is the number of additional units produced times the price of output.

Number of Supervisors MC MR
1 $500 $.50x3800 = $1900
2 $500 $.50x3200 = $1600
3 $500 $.50x1500 = $750
4 $500 $.50x700 = $350
5 $500 $.50x400 = $200

The firm will hire three supervisors since the marginal revenue generated from hiring the
third supervisor exceeds $500 but the marginal revenue generated from hiring the fourth
supervisor is less than $500.

Suggested Essay Questions

1. Assume that wages for keyboarders (data entry clerks) are lower in India than in the
United States. Does this mean that keyboarding jobs in the United States will be lost
to India? Explain.

Answer. Indian data entry clerks will be substituted for American ones only if the ratio of
their wage to their marginal productivity is lower. Thus, it is not wage alone that affects
the incentives to substitute; marginal productivity is also critical.

2. American students have organized opposition to the sale by their campus stores of
university apparel made for American retailers by workers in foreign countries who
work in “sweatshop” conditions (long hours at low pay in bad working conditions).
If this movement is successful in raising pay and improving working conditions for
apparel workers in foreign countries, how will these changes abroad affect labor
market outcomes for workers in the apparel and retailing industries in the United
States? Explain.

Answer. If increased labor costs abroad are not accompanied by increases in marginal
productivity, then there will be incentives to substitute for these foreign workers (with
capital or workers elsewhere, including the United States). However, increased costs of
manufacturing university apparel also would be expected to reduce sales and the scale of
output, which will put downward pressure on employment in the American apparel and
retailing industries. The presence of both substitution and scale effects – working in
opposite directions – implies that the ultimate effect on American workers in these
industries cannot be predicted by theory alone.

3. “Despite free trade and the need to compete with American and Canadian
manufacturers, most Mexican factories continue to use outdated equipment and
inefficient (labor-using) work systems.” If true, does this indicate that, in the face of
very low wages in Mexico, plant owners there are making mistakes?

Answer. The choice of technology is affected by the marginal costs of producing using
labor (W/MPL) compared to the marginal costs of producing using capital (C/MPK).
When wages are low and capital is costly, other things equal, economic theory leads us to
expect that firms would use labor-intensive methods to produce.


While Chapter 3 dealt with the downward sloping nature of labor demand curves,
Chapter 4 deals with the magnitude of the employment response to a change in the wage
rate. We begin the chapter by defining and discussing the own-wage elasticity of demand.
In this regard the Hicks-Marshall laws of derived demand are explained, with each of the
four laws being related to the substitution and scale effects (concepts that were
introduced in Chapters 2 and 3).

After discussing the laws of derived demand in the context of own-wage effects, we
move to a discussion of the cross-wage elasticity of demand. Here we stress the concepts
of gross substitutability and gross complementarity (as distinguished from substitutes or
complements in production). Another section is devoted to a discussion of the empirical
evidence on both the own-wage elasticity of demand and cross-wage elasticities.

The chapter concludes with sections that apply the concepts of demand elasticity to
analyzing the effects of minimum-wage legislation and technological change. The
appendix to Chapter 4 analyzes the labor-market effects of international trade.

List of Major Concepts

1. The own-wage elasticity of demand is the percentage change in employment of a

class of labor induced by a one-percent change in the wages of that class.

2. Cross-wage elasticities of demand are the percentage change in employment of a

class of labor induced by wage changes in another class; they may be positive or

3. The four Hicks-Marshall laws of derived demand are introduced and related to the
substitution and scale effects of a wage change.

4. The concepts of gross substitutability and gross complementarity are defined and
distinguished from substitutability or complementarity in production.

5. Empirical evidence concerning the own-wage and cross-wage elasticities of demand,

based on both statistical studies and inferential analyses, is presented.

6. Standard labor demand theory predicts that an increase in the minimum wage will
result in the loss of employment.

7. Actually measuring the employment effects of minimum-wage increases requires that

we distinguish between nominal and real changes in the rate, that other things
influencing employment levels be controlled for, and that the presence of uncovered
sectors and intersectoral shifts in product demand be built into the design of the study.

8. The results of studies estimating the effects of minimum-wage increases are sensitive
to the specification employed, with some studies finding the "conventional" negative
effects and some finding none. Even those studies with negative employment effects
generally find labor demand elasticities that are much smaller than those summarized
earlier in the chapter.

9. It is possible that the generally small effects of minimum-wage increases are the
result of the studies' focus on short-run effects, but they might also derive from labor
markets that are characterized by monopsonistic behavior (for which theoretically
expected short-run employment effects of mandated wage increases are ambiguous).

10. Technological change in product markets can change the slope and placement of
product demand curves, thereby shifting and/or changing the elasticity of labor
demand curves.

11. The labor-demand effects of technological improvements in capital depend on

crosselasticities; in attempting to analyze the likely dominance of the substitution or
scale effect in this case, the Hicks-Marshall laws applicable to own-wage changes
cannot be slavishly applied.

12. Technological change causes total employment to be reallocated, not permanently


13. (Appendix) International trade is based on comparative advantage, and while trade
may shift employment across industries, it is not true that trade will cause permanent
job loss in high-wage countries.

Answers to Even-Numbered Review Questions

2. Union A faces a demand curve in which a wage of $4 per hour leads to demand for
20,000 person hours and a wage of $5 per hour leads to demand for 10,000 person
hours. Union B faces a demand curve in which a wage of $6 per hour leads to demand
for 30,000 person hours, while a wage of $5 per hour leads to demand for 33,000
person hours.
a. Which union faces the more elastic demand curve?
b. Which union will be more successful in increasing the total income (wages times
person hours) of its membership?

Answer: (a) As noted in the text, the elasticity of demand for labor is not necessarily a
constant along a given demand curve. Indeed, when we speak of changes in wage rates
that are not infinitesimal, the actual value of the elasticity depends on the wage rate from
which one is starting. Given the data on union A and the formula for the elasticity of
demand, %E/%W, union A's elasticity when one increases its wage rate from $4.00 to
$5.00 is given by (20,000-10,000)/20,000 divided by ($4.00-5.00)/4.00, or (1/2)/(-1/4),
which equals -2. In contrast, when one decreases union A's wage from $5.00 to $4.00, its
elasticity is given by (10,000-20,000)/10,000 divided by (5.00-4.00)/5.00 or (-1)/(1/5) or
-5. Its elasticity over the interval $4.00 to $5.00 depends on which wage we use as a base.

To prevent this type of result, economists often define the average elasticity over the
wage interval W1, to W2 as


Note that this elasticity estimate does not vary with the end of the wage interval (high or
low) at which one starts. In the present question the average elasticities for union A and
union B are given by

Elas. (A): [(20,000-10,000)/15,000]/[(4.00-5.00)/4.50] = (2/3)/(-2/9) = -3

Elas. (B): [(33,000-30,000)/31,500]/[(5.00-6.00)/5.50] = -.524

Given the above data, union A faces the more elastic demand curve.

(b) One cannot say which union will be more successful in increasing its members' total
earnings. This depends upon a number of factors, including the bargaining power of the
two unions and the firms with which they deal. It is true, however, that the union with the
more elastic demand curve will suffer a larger percentage employment loss for any given
percentage increase in wages, and this is likely to reduce its incentive to push for large
wage gains. Thus, one's inclination is to say that the union facing the less elastic demand
curve is likely to be more successful in raising its members' wages.

(This answer assumes that wage/employment contracts under collective bargaining lie on
the demand-for-labor curve. As shown in the appendix to Chapter 12, this need not
always be the case.)

4. Clerical workers represent a substantial share of the U.S. work force -- over 15
percent in recent years. Concern has been expressed that computerization and office
automation will lead to a substantial decline in white-collar employment and
increased unemployment of clerical workers. Is this concern well founded?

Answer: Offices have become more computerized in recent years because the cost of
using computers has fallen relative to labor's price (the wage rate). This causes a
substitution effect, tending to shift the labor demand curve to the left for categories of
labor that are substitutes in production with capital. However, there is also a scale effect
tending to increase employment for the above categories, so we cannot tell in advance
which effect will dominate. (For labor categories that are complementary with capital in
the production process, the labor demand curve clearly shifts to the right.) Therefore, it is
not necessarily true that white-collar employment will fall; the scale effect may prevail
for many of these jobs (a dominant scale effect is more likely if product demand is elastic,
if it is difficult to substitute capital for labor, and if the share of capital in total cost is

Even if labor demand shifts left for a particular occupational category, unemployment
will not be the long-term result unless wages are rigid. Adversely affected workers would
have to shift to other occupations and may experience some transitional joblessness, but
only if wages are rigid and employees refuse to shift to lower paying jobs will their
unemployment be permanent.

6. In 1942 the government promulgated regulations that prohibited the manufacture of

many types of garments by workers who did the sewing, stitching, and knitting in
their homes. If these prohibitions are repealed, so that clothing items may now be
made either by workers in factories or by independent contractors doing work in their
homes, what effect will repealing the prohibitions have on the labor demand curve for
factory workers in the garment industry?

Answer : Repealing the prohibitions enables garment manufacturers to substitute home

workers for factory workers. Assuming that the 1942 regulations were constraining, one
can presume that there will be at least some substitution of home workers for factory
workers; this substitution will tend to shift the labor demand curve for factory workers to
the left. However, there may be a favorable scale effect for certain factory workers
performing tasks (such as packaging and shipping) complementary with home production.

Besides the shift to the left of the labor demand curve, the new substitution possibilities
opened up by repealing the 1942 regulations should serve to make the labor demand
curve for factory workers more elastic. Just as the greater ability to substitute capital for
labor will tend to make the labor demand curve more elastic, so too will the ability to
substitute home labor for factory workers.

Answers to Even-Numbered Problems

Ed: the answer to problem 2 is to be changed:

2. Professor Pessimist argues before Congress that reducing the size of the military will
have grave consequences for the typical American worker. He argues that if one million
individuals were released from the military and were instead employed in the civilian
labor market, average wages in the civilian labor market would fall dramatically.
Assume that the demand curve for civilian labor does not shift when workers are
released from the military. First, draw a simple diagram depicting the effect of this
influx of workers from the military. Next, using your knowledge of a) the definition of
the own-wage elasticity of labor demand, b) the magnitude of this elasticity for the
economy as a whole, and c) the size of civilian employment in comparison to this flood
from the military, graph these events and estimate the magnitude of the reduction in
wages for civilian workers as a whole. Do you concur with Professor Pessimist?

Answer. Because you were asked about the effects on civilian wages as a whole, you will
probably not concur with Professor Pessimist. Own-wage elasticity of demand for labor =
%(quantity demanded)/%(wage) = (Ld/Ld)/(W/W). In this case Ld = 1 million, Ld
= about 135 million employed workers, and the own-wage elasticity of demand for labor is

approximately -1. Thus, -1 = (1 million/135 million)/(W/W), so W/W will be very
small -- about -1/135 (or -0.0074). This implies that wages will fall by 0.74 percent.
However, the military recruits in a very narrow segment of the labor market--
mostly high school grads who do not attend college, and who are between ages 17-21.
Thus, downsizing would have the greatest effect on this segment of the market. If there
were only 13.5 million, say, in this age group, a labor demand elasticity of –1 would yield
a wage effect of the military downsizing of closer to –7.4% on this group of the

4. (Appendix) The production possibilities curve for the United States is linear and
allows it to produce a maximum of 500 million units of clothing or 300 million units
of food. The production possibilities curve for France is also linear and allows it to
produce a maximum of 250 million units of clothing or 150 million units of food.
Which good will the United States export to France?

Answer: Neither. The two countries have the same opportunity cost, so neither has a
comparative advantage in either good.

Suggested Essay Questions

1. The public utilities commission in a state lifts price controls on the sale of natural gas
to manufacturing plants and allows utilities to charge market prices (which are 30%
higher). What conditions would minimize the extent of manufacturing job loss
associated with this price increase?

Answer. This question involves the cross-elasticity of demand. A higher price of natural
gas will have a substitution effect that could favor increased employment, and a scale
effect that tends to reduce employment. Factors that minimize the extent of job loss are
those that make for a robust substitution effect and a small scale effect. A large
substitution effect will tend to occur if labor is easily substituted for natural gas in the
production process, and if the supply of labor is relatively elastic. A small scale effect
would be created if natural gas is a small part of the overall cost of production, and if the
demand for the products made using natural gas is relatively inelastic.

2. One anti-terrorism expert proposes the development of two capabilities that would
protect shipments of hazardous materials by truck. One is to maintain continuous
satellite monitoring of all such shipments, and the other is to install devices that
automatically shut down any truck that has been hijacked or deviates from its
approved route. Discuss how implementing this proposal is likely to affect the
demand for truck drivers, noting especially the conditions under which this effect is
likely to be largest.

Answer. This proposal is an attempt to monitor truck drivers, and it really raises the cost
of labor (trucks now require both a driver and monitoring equipment). Thus, the four
factors underlying the elasticity of labor demand are relevant. Where it is easier to
substitute capital for labor, then trucks will tend to get bigger and the number of drivers

needed will go down more. The substitution effect will also be larger if the supply of
capital (in the form of larger trucks) is elastic. If product demand is more elastic, the scale
effect will be larger and product demand will go down more. Finally, where the
monitoring equipment represents a larger share of overall cost, the scale effect will be

3. (Appendix). One observer of the North American Free Trade Agreement (NAFTA)
claims that, contrary to expectations, jobs in Mexican agriculture have been destroyed
while jobs in the industrialized cities of northern Mexico have expanded. Assuming
the facts on job loss and employment gains are accurate, are they consistent with
economic theory?

Answer. Free trade allows countries to specialize in producing goods and services that
have the lowest internal opportunity cost (that is, to specialize in goods for which they
have a comparative advantage). If we think of two generalized goods (agricultural goods
and manufactured goods), a country becomes more efficient in the production of
manufactured goods will, by the definition of opportunity cost, become less efficient in
the production of agricultural goods. If Mexico has a comparative advantage in the
production of manufactured goods, it must have a comparative disadvantage in the
production of agricultural goods. Thus, the assumed facts in the question are quite
consistent with economic theory.

This chapter is designed to analyze the effects of quasi-fixed costs on the demand for
labor. We begin the chapter with a descriptive section on the magnitude and growth of
nonwage labor costs, because the quasi-fixed costs of labor are generally nonwage in
nature. In this section we discuss employee benefits (not all of which are quasi-fixed in
nature), and we also introduce the concept of hiring and training costs.

One implication of the existence of both variable and quasi-fixed labor costs is that there
arises a trade-off between increasing employment through hiring added workers and
increasing employment through hiring workers for longer hours. This trade-off is
discussed in the second section of the chapter, and the importance of distinguishing
between employment and hours is highlighted in our policy analysis of the overtime pay
premium and mandated benefits for part-time workers.

In the third major section we move from a general discussion of quasi-fixed labor costs to
an in-depth analysis of one particular kind of quasi-fixed cost: firms' labor investments.
An investment is a type of expenditure that occurs primarily in some initial period and
then does not recur. While the firm hopes to recoup the investment over a period it
expects the worker to be with the firm, the cost of investment becomes a "sunk cost."
This section analyzes the implications of these labor investment characteristics for the
demand for labor (after first introducing the concept of present value and modeling the
labor investment decision by the firm).

The fourth and fifth sections offer detailed analyses of the two principal types of labor
investments: training investments and hiring investments. In the section on training
investments the student is introduced to the notion of general and specific training, as
well as to the implications of training investments for the demand for labor. While
Chapter 9 also covers aspects of education and training, it is our belief that this
introduction to human capital theory in Chapter 5 is useful. In this chapter, as throughout
the text, we introduce particular concepts or tools as they are called for by the larger
context of analysis, because by maintaining a clear view of the overall context of analysis,
the student is better able to learn the insights that economics has to offer. In this
particular case, we deliberately chose to spread the concepts of human capital theory
across different chapters--using these concepts as necessary and maintaining the overall
substantive organization of the text (built around demand and supply).

For similar reasons, the section on hiring investments includes a discussion of credentials
and signaling, as well as an introduction to the concept of internal labor markets. These
topics are also discussed elsewhere in the text (notably in Chapters 11 and 12), but we
felt that a complete discussion of the effects of quasi-fixed costs on the demand for labor
was impossible without a discussion of these concepts. Again, we wanted to maintain the
organizational overview in the minds of the students. (We also firmly believe that
discussing concepts or phenomena in several contexts and at different points in the book
reinforces the learning process.)

List of Major Concepts

1. The distinction between variable and quasi-fixed labor costs is made.

2. The relative growth of wage and non-wage costs is presented.

3. The essential characteristic of an investment is that resources are expended in the

current period and returns are received later; the principal types of labor investments
that firms undertake relate to training and hiring.

4. There are both explicit and implicit costs of job training.

5. Employee benefits are categorized and the types typically received are listed.

6. The presence of quasi-fixed costs causes an employment/hours trade-off, and the firm
must determine its optimum mix of employment and hours per worker.

7. Increased overtime pay premiums that might be required under the Fair Labor
Standards Act would tend to reduce the use of overtime, but whether they increase the
number of workers employed depends on the size of the reduction in total labor hours

8. The concept of present value and the need for discounting when economic decisions
are made in the context of several time periods are discussed.

9. The multi-period demand for labor, the way this demand is affected by investment
costs in the initial period of hire, and the way investment costs alter profit-
maximizing conditions with respect to labor are all generalizations of the single-
period analysis in Chapter 3.

10. The distinction between general and specific training is defined, and the effects of
specific training on the relationship between wages and marginal productivity is

11. Training investments are recouped through the creation of a "surplus" (a gap between
marginal product and wage) that also cushions the worker from layoffs over the
business cycle.

12. The presence of hiring costs induces firms to use credentials and internal labor
markets in the recruiting, selection and promotion processes.

13. Like training costs, hiring investments increase the productivity of selected job
applicants (by distinguishing among them on the basis of productivity), and they are
recouped by paying wages less than productivity.

Answers to Even-Numbered Review Questions

2. When plants close, firms usually must incur various costs associated with laying off
its workers, including processing necessary forms, helping them find other jobs, and
paying them severance allowances. Suppose that industry X finds itself in a much
more competitive product market than it used to face, and that firms in the industry
now have a greater probability of closing than they used to have. How might this
change affect (a) the number of employees hired in the industry, and (b) their average
hours of work?

Answer. Firing costs are quasi-fixed, because they are associated with workers, not hours
of work. When they are increased, as they are in industry X, this will induce firms to (a)
hire fewer workers, and (b) work those they hire for more hours.

4. Workers in a certain job are trained by the company, and the company calculates that
to recoup its investment costs the workers’ wages must be $5 per hour below their
marginal productivity. Suppose that after training, wages are set at $5 below
marginal productivity, but that developments in the product market quickly (and
permanently) reduce marginal productivity by $2 per hour. If the company does not
feel it can lower wages or employee benefits, how will its employment level be
affected in the short-run? How will its employment level be affected in the long run?
Explain, being sure to define what you mean by short-run and long-run!

Answer. In the short run (that is, when training investments have already been
concluded, so all that is variable is the employment levels of trained workers), marginal
revenue product still exceeds wages by $3 per hour, so it is advantageous for the
company to continue employing workers it has already trained. The company is not
making back enough to make the training be a good investment, but making back $3 per
hour is better than laying off the workers and making back nothing! Thus, workers will
not be laid off.

In the long run (that is, when the company is deciding about investing in new workers),
the $3 payback per hour is not sufficient to justify the training investment if wages
remain as they are. Thus, the firm will not hire and train new workers under the current
circumstances. Employment will fall as the firm fails to replace those who leave, and the
decline in employment will eventually serve to raise the marginal productivity of labor.
The decline in employment will stop when the marginal revenue product of labor is once
again $5 greater than the wage rate.

6. Suppose that the United States adopts a policy requiring employers to offer 600 hours
of paid leave for mothers of newly born babies. Assuming wages remain the same,
analyze the labor demand effects of mandated paid child-care leave on women of
childbearing age and on women past childbearing age.

Answer: This policy clearly increases the expected cost of employing women of
childbearing age by imposing on employers a quasi-fixed cost (equal to 600 hours of

normal earnings). This increased cost, with wages remaining equal, will reduce the
demand for younger women; the quasi-fixed nature of the cost implies that their
employment will fall more than their average hours of work.

For older women, for whom the costs of employment are unaffected, there will be both
scale and substitution effects. The former will tend to reduce demand for their services,
while the latter will tend to increase it. The overall effects of this policy on the demand
for older women cannot be predicted from theory alone (however, the four factors
affecting the elasticity of demand for labor can be used to analyze when the substitution
effect will be large relative to the scale effect).

8. Major league baseball teams scout and hire younger players whom they then train in
the minor leagues for a period of three to five years. Very few of their trainees
(perhaps 5%) actually make it to the major leagues, but if they do they are bound to
the team that owns their contract for a period of six years. After six years, the player
can become a "free agent" and choose any major league team on which to play.
Keeping in mind that the major league teams pay the costs of, but derive no revenues
from, their minor league teams, what would be the most important predictable effects
of allowing players to become free agents immediately upon entry into the major

Answer: During the training period, teams are paying the salaries of their minor league
players and expending other resources on their training without receiving any revenues in
return. These costs represent investments in general training. A firm has no incentives to
offer general training at its own expense unless it can somehow tie the trainee to the firm
for a period long enough to recoup its investment expenditures. The rule under which
players are tied to the major league team owning their contract is intended to offer teams
a period over which to recoup these general training expenses.

If players were able to become free agents immediately upon making it to the major
leagues, teams that did not train these players would bid their wages up to a level equal to
their marginal productivity. Teams offering the training would therefore have no way of
recouping their investment expenditures, which can only be done by paying a wage less
than marginal productivity. Thus, with immediate free agency, teams would no longer
have incentives to scout and train their own players, and they would tend to adopt a
strategy of "raiding" players already trained by other teams. The major effect of
immediate free agency would therefore be to destroy the current minor league
arrangements for training players. The major league teams might give up their minor
league teams and rely solely on colleges for training professional baseball players.
Immediate free agency might also cause independent baseball training schools to arise,
with tuition charged directly to the trainees. A final alternative might be for the major
league baseball teams to collectively operate a minor league system that is financed by
assessing each team an equal share of the total costs of running the training operation.

Answers to Even-Numbered Problems

2. Suppose that a firm is considering training a worker. The worker's MPL is $100
during the training period, but rises to $200 in the post-training period. The worker's
wage is $100 during the training period, the cost of training is $50 and the discount
rate is 10%. What is the most that a profit-maximizing firm can afford to pay the
worker in the second period?

Answer: The firm will undertake the training if the discounted net benefits from the post-
training period exceed the net expense from the training period, i.e., if W0 + Z - MP0 <
(MP1 - W1)/(1 + r). Plug in the values to solve for W1 at the breakeven point. $100 + $50
- $100 = ($200 - W1)/1.1, or $50x1.1 = $200 - W1, so W1 = $145. If the post-training
wage is less than $145, the firm will make a profit.

Suggested Essay Questions

1. The manager of a major league baseball team argues: “Even if I thought Player X
was washed up, I couldn’t get rid of him. He’s in the third year of a four-year, $24-
million deal. Our team is in no position financially to eat the rest of his contract.”
Analyze the manager’s reasoning using economic theory.

Answer. A baseball team that has committed itself to a four-year contract has made an
investment, in the hopes, of course, of receiving a return. The cost has been “sunk,” so it
is of no relevance to any decision about how to use the player during the contract period.
The only thing of relevance is the player’s marginal revenue productivity as compared to
the marginal revenue productivity (less marginal cost to the team) of an alternative player.

2. One recent magazine article on economic recovery from a recession argued, “Labor
productivity growth usually accelerates in the first year of an expansion, because
firms are slow to hire new labor.” Comment.

Answer. One reason firms are slow to hire in expansions is that they are slow to lay off
workers during a recession. Workers in whom the firm has made an investment are paid
less then the value of their marginal product, so that the firm can recoup investment costs,
and this difference offers employment protection when productivity falls in a recession
(because investment costs are sunk and the firm will continue to employ a worker in the
short run as long as productivity exceeds the wage). As productivity rises during
expansion, firms will not hire workers (which involves an investment) until the gap
between productivity and wages is again large enough so that the firm can recoup
investment costs.

3. An author recently asserted, “Low wage jobs provide fewer hours of work than high-
wage jobs.” Using economic theory, is this statement likely to be correct? Why?

Answer. Low wage jobs involve less training than high wage jobs, and if the training in
high wage jobs is firm-specific, employers will want to substitute longer hours of work

for hiring more workers. Thus, it is consistent with economic theory for employers to
require longer hours of work for workers with more skills.


Beyond introducing some descriptive material on labor force trends in this century, the
primary purpose of Chapter 6 is to present an analysis of an individual's decision
concerning whether and for how long to work. The context of this decision is the
traditional labor/leisure choice framework and the chapter is carefully constructed to
build the concepts necessary for this analysis. The analysis begins with a section that
discusses the choice process verbally, building upon what students know concerning
product demand. It then moves to a specific analysis of the demand for leisure time
(which in this context is the obverse of the supply of labor), and introduces the concepts
of income and substitution effects (they are more rigorously dealt with later in the
context of a graphic analysis).

Our graphic analysis is intended to accomplish two ends. One is to fix and define more
precisely the concepts of income and substitution effects. The second is to equip students
with a tool necessary to analyze many policy issues affecting work incentives. A
sampling of such policies and their analyses is given in the final section of the chapter
(following a section that discusses empirical findings concerning labor supply to the

List of Major Concepts

1. Measures of aggregate labor supply generally focus on labor force participation rates
and weekly hours of work; trends in these measures are presented and discussed.

2. The relationship between the demand for leisure, the demand for other goods, and the
supply of labor is the focal point for beginning our analysis of labor supply theory.

3. The substitution effect is defined as the change in hours supplied attendant on a

change in the wage (price of leisure), holding income constant.

4. The income effect is the change in hours supplied for a given change in income,
holding the wage constant.

5. The major forces affecting labor supply are preferences, wages, and income; these
forces can be graphically depicted.

6. The five assumptions underlying indifference curves (a graphic depiction of

preferences) are discussed.

7. The incorporation of information on wages and income into the drawing of budget
constraints is illustrated.

8. Graphical analyses of the income and substitution effects are presented.

9. The concept of "reservation wage" is defined and illustrated graphically.

10. Empirical findings with respect to the labor/leisure choice, from both
nonexperimental cross-section data and experimental studies, are presented.

10. Analyses of the budget constraints created by several government income support
programs are presented. Analyzed are those with "skikes," those with zero net wage
rates (including those with work requirements), and those with positive effective
wage rates (as illustrated by an analysis of the Earned Income Tax Credit program).

Answers to Even-Numbered Review Questions

2. Evaluate the following quote: “Higher take-home wages for any group should
increase the labor force participation rate for that group.”

Answer. This quotation is correct, because for labor force participation decisions, the
substitution effect dominates the income effect. The strength of the income effect is
relatively weaker when the initial hours of work are smaller. When initial hours of work
are zero – as is the case when a person is out of the labor force – then the income effect is
zero if leisure is a normal good (increased resources cannot induce one to increase the
consumption of leisure, since leisure hours are already at their maximum).

4. The way the workers' compensation system works now, employees permanently
injured on the job receive a payment of $X each year whether they work or not.
Suppose the government were to implement a new program in which those who did
not work at all got $0.5X but those who did work got $0.5X plus workers'
compensation of 50 cents for every hour worked (of course, this subsidy would be in
addition to the wages paid by their employers). What would be the change in work
incentives associated. with this change in the way workers' compensation payments
are calculated?

Answer: This change in workers' compensation has two effects. First, it reduces the
subsidy for people who do not work from $X to $0.5X. This reduction in income by itself
would produce an income effect that tends to induce the injured worker to work more (he
or she is poorer if not working than under the previous workers' compensation system).
On the other hand, for those who work, the wage rate is increased by 50 cents an hour.
(We assume here that the change in workers' compensation payments is not so large as to
influence market wages.) The increased wage by itself would tend to induce injured
workers to work more because the cost of leisure has risen by 50 cents an hour; however,
the eventual outcome is theoretically unclear.

The effects of these changes can be seen in the figure below.

Along segment DE there is a clear-cut strengthening of work incentives. Segment DE has
a steeper slope than the previous budget constraint (BQ and it also lies to the southwest of
BC. Thus, along segment DE there is a substitution effect inducing more work and an
income effect that also induces more work. To the left of point E, however, along
segment EF, there are income and substitution effects that work in opposite directions.
Along segment EF the 50-cents-an-hour increase in the wage rate is sufficient to increase
the injured worker's income under workers' compensation, thereby creating an income
effect that reduces work incentives, other things equal. However, the substitution effect
of the increased wage continues to exert an increase in work incentives and the outcome
of the two effects is not predictable in advance.

Thus, if the tangency point between the worker's indifference curve and the full budget
constraint used to be along BC but to the right of point E, the worker faces a clear-cut
strengthening of work incentives under the new program. If, however, the worker's
tangency point along BC was to the left of point E, the new program would have an
unpredictable effect on work incentives.

6. Suppose the Social Security disability insurance (DI) program was structured so that
otherwise eligible recipients lost their entire disability benefit if they had any labor
market earnings at all. Suppose, too, that Congress was concerned about the work
disincentives inherent in this program, and that the relevant committee was studying
two alternatives for increasing work incentives among those disabled enough to
qualify for it. One alternative was to reduce the benefits paid to all DI recipients but
make no other changes in the program. The other was to maintain the old benefit

levels (for those who receive them) but allow workers to earn $300 a month and still
keep their benefits. Those who earn over $300 per month would lose all DI benefits.

Analyze the work incentive effects of both alternatives. (The use of graphic analyses
will be of great help to you.)

Answer: The proposal to reduce the average DI benefit may cause recipients to seek
work or it may not, depending on their preferences and the extent of the cut. Compare,
for example, cases a, b, and c below.

The proposal to allow DI recipients to keep their benefits until a certain earnings level is
reached will induce some of those now not working to work at least a little (case d).
Others may have preferences that preclude work (case e). However, some of those who
medically qualify for DI but would now work may decide to cut their hours of work (case
f). Thus, it is not clear from theory which proposal would have the stronger work

8. The Tax Reform Act of 1986 was designed to reduce the marginal tax rate (the tax
rate on the last dollars earned) while eliminating enough deductions and loopholes so
that total revenues collected by the government could remain constant. Analyze the

work incentive effects of tax reforms that lower marginal tax rates while keeping total
tax revenues constant.

Answer: Reducing the marginal tax rate has the effect of increasing the wage rate,
because workers are allowed to keep more from any extra hours worked. Keeping tax
revenues constant suggests that workers' after-tax incomes also remain constant. Thus,
the Tax Reform Act tended to increase the wage while keeping workers' incomes
constant -- creating a pure substitution effect that tended to increase hours of work.

Answers to Even-Numbered Problems

2. Nina is able to select her weekly work hours. When a new bridge opens up, it cuts
one hour off Nina's commute to work. If both leisure and income are normal goods,
what is the effect of the shorter commute on Nina’s work time?

Answer. When the new bridge opened, Nina’s budget constraint shifted to the right in a
parallel fashion as the amount of available time for either work or leisure (as opposed to
commuting) was increased. This shift in her constraint created an income effect (she can
now work more and consume more leisure). Because both income and leisure are normal
goods, both would increase. The only way income can increase in this case is for her to
work more, so we must conclude that her extra hour per day from the shorter commute is
divided in some way between more work and more leisure. Therefore, she works more.

Suggested Essay Questions

1. In 2002, a French law went into effect that cut the standard workweek from 39 to 35
hours (workers got paid for 39 hours even though working 35), while at the same time
prohibiting overtime hours from being worked. (Overtime in France is paid at 25%
above the normal wage rate.) (a) Draw the old budget constraint, showing the
overtime premium after 39 hours of work. (b) Draw the new budget constraint. (c)
Analyze which workers in France are better off under the 2002 law. Are any worse
off? Explain.

Answer. In the drawing below, the old (pre-2002) constraint is ABC, where slope of BC
is 25% greater (in absolute value) than the slope of AB. The constraint created by the
new law is ADE, where earnings at D are equal to those at B, and the slope of DE is
horizontal (workers cannot get paid for more than 35 hours of work).



39 35 0  Hours of Work

close to B (that is, they worked close to 39 hours before), will also be better off if their
original utility-maximizing indifference curve passed below point D. However, for those
whose original utility-maximizing indifference curves passed above point D (almost
surely the case for most of those with original tangencies along BC), utility will fall under
the new law.

2. Country X cuts the income tax rates applicable to those with the highest incomes, and
it newly adopts a wealth tax – a tax that is based on the value of family assets
(personal assets, real estate and financial assets) above a certain threshold. Discuss
the likely work incentive effects of these tax changes on high-income workers.

Answer. The new law changes the constraint from ABC to ADEF.


 Hours of Work

Clearly, the income tax rate reduction increases the slope of the budget constraint
(increases the net wage rate). If the wealth tax reduces a person’s overall command over
resources (which happens along segment DE), then work incentives are clearly
increased – wages are increased while wealth falls. If the effect of the two tax changes
serve to increase both the wage rate and the command over resources (compare segment
EC with EF), then the tax changes have an ambiguous effect on work incentives, because
the substitution and income effects have opposite effects on work incentives.

Chapter 7 analyzes the labor supply decision (the decision to work for pay) in the context
of household production theory. In this chapter, the primary alternative to working for
pay is not assumed to be leisure, but household production. This framework quite
naturally leads the discussion of labor supply into the context of families, thereby raising
the issue of family labor supply decisions. Further, since one's household productivity
varies considerably across the life cycle (as, of course, do wages), the concepts of
household production also lead to a discussion of labor supply over the life cycle.

Instructors facing severe time constraints may wish to skip this chapter. The insights
provided by the analysis in Chapter 7 are refinements of the basic concepts introduced in
Chapter 6, and they do not contradict the insights or predictions of Chapter 6. However,
Chapter 7 summarizes some recent directions in which labor supply theory has been
going, and to sacrifice Chapter 7 would mean forgoing concepts and empirical work
close to the frontiers of economic analysis.

The chapter begins with an introduction to the concept that households combine time and
goods to produce commodities that are consumed at home. The graphic analysis of
household production and the choice of household production technology is shown to be
completely analogous to the graphic analysis and fundamental implications of the
labor/leisure choice discussed in Chapter 6. The household production context of the
labor supply decision, however, yields insights about that decision that go beyond those
of Chapter 6. These insights are discussed after our brief introduction to household
production theory in the first section.

In particular, we point out the tripartite choice between market work, household work,
and leisure in analyzing why the substitution effects for women might be expected to be
larger than those for men. We discuss such family labor supply decisions as who stays
home to care for children (if anyone does), whether both spouses will work for pay, and
the interdependency of the spouses' labor supply decisions. The "additional worker" and
"discouraged worker" hypotheses are also discussed in this context.

Our discussion of the life-cycle aspects of labor supply begins with the observation that
household productivity does indeed vary over the life cycle. The traditional interrupted
careers of married women cannot be explained without reference to the shifts in
household productivity that take place when children are born and as they grow older.
Labor supply over the life cycle is also affected by the way wages typically vary with age,
causing intertemporal substitution effects; in this context, we discuss the important issue
of choice of retirement age (including data on the way lifetime Social Security benefits
vary with age of retirement).

The chapter concludes with a policy analysis of "child support assurance" programs.

List of Major Concepts

1. The basic concepts of household production theory include the combining of goods
and time to produce commodities that yield the family utility.

2. Household commodities may be produced by time-intensive methods or by goods-

intensive methods; the method chosen is in part a function of the price placed on time.

3. The principal predictions associated with the income and substitution effects in the
labor/leisure model are unchanged in the context of the household production model.
The latter model, however, adds a third dimension of choice about time usage (market
work, household work, leisure).

4. As wages change, there will be changes in the time intensity of commodities

consumed as well as in the time intensity of household production technologies.

5. Joint household production decisions (which spouse, if either, should remain home
instead of working for pay) have yet to be completely modeled, but they must clearly
take account of the partners' marginal productivities at home and the wages they can
command in the "market."

6. The "discouraged worker hypothesis" and the "additional worker hypothesis" are
discussed in the context of household production theory.

7. Labor supply decisions over the life cycle are affected by household productivity
changes and predictable changes in wages over the life cycle that create intertemporal
substitution effects without corresponding income effects.

8. Graphic analysis of the choice of optimum retirement age is presented, emphasizing

how delaying retirement by a year can affect the present value of one's total income
over the remaining years of expected life.

9. Child support assurance programs ensure transfer payments to custodial parents based
on the age and number of children, not on income. In contrast with welfare programs,
which tend to create budget constraints with zero net wage rates, child support
assurance programs preserve incentives to engage in market work. However, for
those who worked for pay in the absence of such programs, the pure income effect
created by support assurance programs should tend to induce fewer hours of paid

Answers to Even-Numbered Review Questions

2. A recent study of the labor force participation rates of women in the post-World War
II period notes:

Over the long run women have joined the paid labor force because of a series of
changes affecting the nature of work. Primary among these was the rise of the
clerical and professional sectors, the increased education of women, labor saving
advances in households, declining fertility rates, and increased urbanization.

Relate each of these factors to the household production model of labor supply that
was outlined in Chapter 7.

Answer: One of the central aspects of the household production model of labor supply is
the importance of the relative productivity in paid employment as compared to household
production. Increased opportunities in the clerical and professional sectors, as well as
increased educational levels, serve to increase productivity in paid employment (that is,
to increase the wage rate that women can command). Declining fertility rates tend to
reduce the productivity of hours spent at home, while the invention of labor saving
devices in household production make it easier to substitute goods purchased with cash
for time at home; both of these factors flatten the household utility isoquants (an hour of
household productivity forgone can be replaced more readily by goods purchased with
money). Increased urbanization also tended to make it easier to substitute goods for
household production. All these factors tended to raise market productivity relative to
household productivity, and some of them served to increase the strength of the
substitution effect relative to the income effect.

4. Is the following statement true, false, or uncertain? Explain.

"If a married woman's husband gets a raise, she tends to work less, but if she gets a
raise, she tends to work more."

Answer: Ignoring the question of joint labor supply decisions, if a married woman's
husband gets a raise, that raise (to her) has an income effect. This increased income
without a corresponding increase in her wage rate tends to induce her to work fewer
hours. However, if her wage rate rises, she will experience both an income and a
substitution effect, and if she already works, theory cannot predict which one is dominant.
If she is out of the labor force, a wage increase will increase her chances of labor force

The text pointed out, though, that spouses may make their labor supply decisions jointly.
For example, if the husband's wage increase caused him to work more, the wife may also
decide to work more if they are complements in household production (or consumption).
Thus, the answer to this question really depends upon whether one assumes the two
spouses have household productivities that are interdependent; if so, they must make
their labor supply decisions jointly.

6. Several studies have indicated that for prime-age males, the income effect of a wage
increase tends to dominate the substitution effect. Other recent studies point out that
hourly wages tend to rise over the early stages of the life cycle (the young receive
lower wages than the middle-aged) and that young males tend to work fewer hours
than middle-aged males, other things equal. Employing a theory of life-cycle
allocation of time, explain the apparent discrepancy.

Answer: Studies showing that for prime-aged males the income effect of a wage increase
tends to dominate the substitution effect look either at wage increases that have occurred
as society has become wealthier and more productive or at wage rates across individuals
in a population. In both cases there are both substitution effects and income effects of
wage changes. However, studies of the life-cycle effects of lower wages in the early
stages of one's working career with higher wages later on are examining these wage-
change effects over an individual's lifetime, holding constant the individual's expected
lifetime wealth. With these studies there is a substitution effect – leading to more work as
wages rise – but no corresponding income or wealth effects. The latter studies are in the
pure life-cycle mode of analysis, where at a given time individuals have an expected
lifetime wealth and also face predictable changes in their wage rate as they age.

8. Suppose that, under state law, the financial settlement in a divorce case that does not
involve dependent children depends upon the economic contribution each marriage
partner made up to the date of divorce. Thus, if the wife earned an income equal to
her husband's throughout the years, she would be determined to qualify for half of the
assets at the date of divorce. Based on what you have learned in Chapter 7, how could
an equitable settlement be determined in the case of a woman who stayed home,
raised the family's children, and never worked for pay?

Answer : A wife who did not work for pay nevertheless contributed to the family's
income by performing household production services that would otherwise have had to
be purchased in the market at some cost. Put differently, a woman who performs
household services saves the family money that it would otherwise have had to spend.
For a discussion of how these services can be valued, see Example 7.2.

Suggested Essay Questions

1. Assume that a state government currently provides no child care subsidies to working
single parents, but that it now want to adopt a plan that will encourage labor force
participation among single parents. Suppose that child care costs are hourly, and
suppose the government adopts a child-care subsidy that pays $3 per hour for each
hour the parent works, up to 8 hours per day. Draw a current budget constraint for an
assumed single mother (net of child care costs), and then draw in the new constraint.
Discuss the likely effects on labor force participation and hours of work.

Answer. If the old budget constraint is AB below, the new one will have a steeper slope
(reflecting a net wage that is $3 per hour higher) for the first 8 hours of work (see AC);

after that, the budget constraint is segment DB. Among those single parents not working
before the subsidy is adopted, the higher wage rate will tend to increase labor force
participation (the substitution effect dominates for participation decisions). For those
already working (tangencies along AC), the income effect and substitution effects of this
wage increase will have opposite tendencies on the hours of work, so the net effect is not
predictable. However, some people working over 8 hours a day before may reduce their
supply of hours and move to point C on the constraint.



8  Hours of Work

2. Assume that a state government currently provides no child care subsidies to working
single parents, but that it now want to adopt a plan that will encourage labor force
participation among single parents. Suppose child care costs are hourly, and that the
government adopts a child-care subsidy of $20 per day if the single parent works 6 or
more hours per day. Draw the current budget constraint (net of the hourly child care
costs) for an assumed single mother, and then draw in the new constraint. Discuss the
likely effects on labor force participation and hours of work.

Answer. In the drawing below, the pre-subsidy constraint is AB. The subsidy of $20
per day (CD) begins at 6 hours of work and continues for all levels of work hours beyond
6 (segment DE, which is parallel to AB). Thus, the new constraint is ACDE. Those
single parents who were out of the labor force before (maximized utility at A) and who
have very steep indifference curves will tend to remain at point A; however, those with
flatter indifference curves will find that their utility is maximized at point D. Thus, some
workers who were out of the labor force before will now join, and those who do will

desire jobs offering exactly 6 hours of work per day. For those along AC before
(working less than 6 hours per day), the tendency also will be to move to 6 hours of work
(although it is possible that some will have such a steep indifference curve to the left of
their tangency along segment AC that they will not be better off by working 6 hours).
For those working more than 6 hours per week before, the income effect of this subsidy
will create a tendency for them to desire fewer hours of work (as long as the hours do not
fall below 6).


6 Hours of Work

Chapter 8 introduces students to the concept of compensating wage differentials.
Following the practice in earlier chapters, it seeks to move students from concepts they
are familiar with to new concepts and tools. Again, the analysis begins with a verbal
exposition of occupational choice and the wage outcomes that flow from this choice
when jobs differ along nonpecuniary dimensions. Once the essential assumptions and
predictions of economic theory in this context are explained, we introduce students to a
graphic analysis that is intended to yield additional insights. The graphic analysis of the
issue of occupational choice is also intended to provide students with a tool for analyzing
the effects of government policies on the labor market.

We first apply the concepts of hedonic theory to a "bad" job injuries. Policy implications
are related to occupational safety and health legislation. We then apply the theory to an
analysis of how elements in the employment "package" on which employees place a
positive value affect the wage rate. The application in this section of the chapter relates to
the regulation of employee benefits, particularly pensions.

For those who wish to enrich the coverage in Chapter 8, we have added an appendix that
analyzes worker choice of jobs that have different probabilities of layoff. This appendix
offers another application of the theory of compensating wage differentials to an
interesting policy problem, and in so doing elucidates certain issues not commonly
understood. The analysis also introduces the student to the notions of "risk aversion" and
the willingness to pay for insurance ("certainty").

List of Major Concepts

1. In the context of full information and choice, worker behavior will generate
compensating wage differentials for job characteristics that are unpleasant or costly.

2. Compensating differentials play a dual role in allocating labor to unpleasant jobs and
in compensating those who accept unpleasant work.

3. The prediction that there will exist compensating wage differentials for unpleasant
work rests on assumptions of utility maximization, worker information, and worker

4. Employee preferences are graphically expressed in the concavity and slope of

indifference curves.

5. Employers with different costs of eliminating unpleasant job characteristics can be

graphically represented.

6. A market equilibrium curve (or offer curve) is derived from the zero-profit isoprofit
curves of the employers in the market.

7. If the market is working properly, employees who are least averse to an unpleasant
job characteristic become employed with firms that find it most expensive to
eliminate that characteristic.

8. The theory of compensating differentials can only be tested using techniques that
control for other influences on job characteristics.

9. Government attempts to regulate the outcome of labor market decisions that are made
in a perfectly functioning market could lead to a reduction of utility for the workers
the government is intending to help.

10. Government intervention into the labor market can increase worker utility if the
market is not functioning perfectly (that is, if not all costs or benefits of the decision
are borne by those making them).

11. The mix of wages and benefits in the compensation package depends on both
employee preferences and the trade-offs employers are willing to make.

12. (Appendix) Some job characteristics normally considered bad may be considered
good by some workers (layoffs may be preferred if they are known in advance).

13. (Appendix) There are two issues relating to the undesirable characteristics of layoffs:
the degree to which yearly layoffs (known in advance) constrain a worker's hours of
work to lie below those otherwise desired, and the degree to which layoffs cause the
worker's income each year to fluctuate.

14. (Appendix) The concept of risk aversion is related to the hypothesis that the expected
utility of a level of income ($X) received with certainty is greater than the expected
utility of a stream of income that may fluctuate over time but yield an expected yearly
value of $X.

Answers to Even-Numbered Review Questions

2. Statement 1: "Business executives are greedy profit maximizers, caring only for
themselves." Statement 2: "It has been established that workers doing filthy,
dangerous work receive higher wages, other things equal." Can both of these
statements be generally true? Why?

Answer. Both statements can be simultaneously true. If workers are informed about job
hazards and have a choice about the jobs they take, their behavior will force even the
most greedy executives to pay higher wages for filthy, dangerous work. Even the
greediest profit maximizer must obtain a work force, and to do so must pay a wage that
workers will accept. If workers have alternative job offers that pay the same wage but
offer better working conditions, they will accept those offers and turn down work at the
more dangerous or filthy workplaces. Their behavior then will force owners to either pay
the compensating wage differentials or clean up the workplace.

4. Suppose highway workers in a certain city are required to give the Supervisor of
Highways an under-the-table payment of $X per year. Would you expect wages paid
to highway workers by this city to be higher or lower than the market wage? Would
you expect the salary paid by the city to its Supervisor of Highways to be above or
below market? Explain.

Answer. We would expect wages for highway workers in the city to be high relative to
the market wage. In order for workers to be attracted to highway jobs in the city they
must take home wages that are at least as great as they could obtain elsewhere. This
means that the wages they receive, less any bribes they must pay, must be at least equal
to the prevailing wage. If the Supervisor is able to extract under-the-table payments from
these workers, it is an indication that the wages paid by the city are relatively high.
Alternatively, one can look upon the bribe as an undesirable element of employment for
which the workers would have to receive a compensating wage differential.

The Supervisor's wage, by analogy, can be low relative to the market. Under-the-table
payments provide a supplement to the Supervisor's regular salary, so that even if the
Supervisor's salary is low the job may attract a satisfactory flow of applicants.

6. Suppose that Congress were to mandate that all employers had to offer their
employees a life insurance policy worth at least $50,000 in the event of death. Use
economic theory, both positively and normatively, to analyze the effects of this
mandate on employee well-being.

Answer. From the perspective of positive economics, mandating that employers offer at
least $50,000 in life insurance will obviously have no effect on those who are already
offering that much or more, but it will add to the costs of those who were previously
offering less. To be competitive in the labor market, those previously offering less must
have been compensating their workers in some other way (to make their jobs as attractive
as those of their competitors). It is thus likely that low-insurance employers were paying
higher wages than those offering more insurance. To now compete with their
competitors in the product market, the affected employers must reduce their wages (to
keep overall costs in the competitive range). Thus, the wages in firms previously offering
less insurance will decline. Of course, if wages do not, or cannot, decline by enough to
fully offset the added costs of more insurance, then employment among these employers
will fall.

From the perspective of normative economics, we would like to know if this mandate
improves the welfare of the workers affected. If the labor market is perfectly functioning,
workers are able to obtain the combination of wages and life insurance that maximizes
their utility. If the mandate forces them to take some other mix, then their utility will
decline. If the market is not allowing workers to “buy” (in the form of lower wages) the
life insurance they want, then mandating increased insurance could improve the welfare
of affected workers (as long as the mandate does not require workers to buy more than
they are willing to pay for).

8. "The concept of compensating wage premiums for dangerous work does not apply to
industries like the coal industry, where the union has forced all wages and other
compensation items to be the same. Because all mines must pay the same wage,
compensating differentials cannot exist." Is this statement correct? (Assume wages
and other forms of pay must be equal for dangerous and non-dangerous work and
consider the implications for individual labor supply behavior.)

Answer. This statement is not correct. To understand how the market would adjust, let us
assume that we have a set of relatively safe coal mines and a set of relatively dangerous
coal mines. Both sets of mines must pay the same wage rate and offer the same fringe

They both advertise for help and, assuming workers quickly find out which mines are
safe and which are dangerous, the safe mines receive many more applications than the
dangerous mines. The safe mines can thus be highly selective about the applicants they
choose, and they will tend to hire the most dependable, hardest working, most motivated
employees. The dangerous mines, with very few applicants, will have to take who they
can get (those workers not chosen to work in the safe mines). Safe mines will have high
quality, highly productive workers getting wage $X, while the dangerous mines will have
lower quality workers obtaining the same wage. Thus, workers of unequal productivity
would receive the same wage, and this is tantamount to the receipt of a compensating
wage differential.

Put differently, the theory of compensating wage differentials says that people of equal
skill will receive different wages when working conditions differ. But a natural corollary
of this is that, when working conditions differ, people of different skills might receive the
same wage. In both cases workers in less desirable circumstances receive higher wages
than they would otherwise receive.

Answers to Even-Numbered Problems

2. Consider the conditions of work in perfume factories. In New York perfume factories,
workers dislike the smell of perfume, while in California workers appreciate the smell of
perfume, provided that the level does not climb above S*. (If it rises above S*, they
start to dislike it.) Suppose that there is no cost for firms to reduce or eliminate the smell
of perfume in perfume factories and assume that the workers have an alternative wage,

Draw a diagram using isocost and indifference curves that depicts the situation. (The
New York and California isocost curves are the same, but their indifference curves
differ.) What level of perfume smell is there in the New York factories? In the California
factories? Is there a wage differential between the California and New York workers?

Answer: See the figure below. The California workers are paid exactly the same as the
New York workers. This wage equals W*. The level of smell in California is S*; in New
York it is 0.

Suggested Essay Questions

1. A country passes a law, applying only to large firms, that cuts the standard workweek
from 39 to 35 hours (workers are paid for 39 hours even though working 35), while at
the same time prohibiting overtime hours from being worked. This law does not
apply to smaller firms in the country. (Overtime had been widespread in both sectors,
and is paid at 50% above the normal wage rate.) Using economic theory, how do you
think this law might affect wages and employment in the two sectors?

Answer. The immediate effect on workers in large firms is to give them a wage increase
of around 10% (work 35, are paid for 39), but the law also prevents them from working
overtime. How wages and employment respond depends on the value workers put on
overtime work. If they value the ability to work overtime, some in the large-firm sector
may seek jobs in the small-firm sector (where it is still possible to earn a 50% premium
for hours in excess of 39) – which would drive down wages in the small-firm sector and
drive them higher in the large-firm sector. The higher wages in the latter sector would
reflect a compensating differential for the restricted access to a valuable workplace
characteristic. However, if workers now see jobs in the large-firm sector as more
desirable than before (given the immediate wage increase and the inability of employers
to demand overtime work of their employees), then more would seek work in that sector.
In this latter case, employment shrinks and wages rise in the small-firm sector, while
employment grows and wages fall in the large-firm sector.

2. A recent article stated, “Workers in low-wage jobs lack the basic security, the health
benefits, and the flexibility in their work lives that most American workers take for
granted.” Assuming this statement if true, do these facts contradict the theory of
compensating wage differentials?

Answer. The theory of compensating differentials predicts that, other things equal, jobs
with low non-wage benefits would have to pay higher wages. This statement is implicitly
comparing those in low-skilled jobs with those in high-skilled jobs, where clearly “other
things” are not comparable. Thus, the facts in this statement do not contradict the theory
of compensating wage differentials.


Chapter 9 introduces students to the concept of human capital and treats in detail
education and training investments. The chapter begins with a section on the demand for
education by workers, in which a theory of human capital investment is formulated and a
formal model of choice is presented. Implications of this model for both individual and
aggregate (market) behavior are then derived.

The second section of the chapter analyzes the relationship between education and
earnings. We introduce age/earnings profiles and discuss the reasons for their convexity.
Included in this section is an analysis of the differential convexity among such profiles
for men and women.

Next, we consider the question of whether education is a good investment. We analyze

this question from both an individual and a social perspective. The major findings of the
literature with respect to the individual rates of return to education are summarized, and
we discuss possible biases (including selection biases) inherent in these findings. When
discussing education as a social investment, we introduce both the traditional answers of
the "human capitalists" and the more agnostic views of those who see education as purely
a signaling device. In this context of evaluating education and training as investments, we
devote a section to evaluations of government job training programs.

Appendix 9A presents and explains a "cobweb" model of labor market adjustment, in

which the need for educational investments slows down the supply response to changes
in market demand. Appendix 9B presents a hedonic model of education and wages that
uses the graphic tools of Chapter 8. This hedonic model is useful in explaining several
empirical facts about the relationship between education and wages, and the discussion
also serves to integrate the concepts in Chapters 8 and 9.

List of Major Concepts

1. Investments in human beings are part of the general category of investments.

2. Investments entail costs in the current term with returns flowing in over later periods.

3. Costs of human capital investments include out-of-pocket expenses, forgone earnings,

and psychic losses.

4. Because investment returns flow in over several years, an analytical tool to convert
future sums to present value is required (the concept of present value and discounting
future sums is explained in some detail).

5. Human capital investments are more likely to be made by people who are not present
oriented, by people who are young, in situations in which the costs of human capital

investments are lower, and in situations in which the returns to these investments are

6. Variations in the returns to human capital investments call forth supply responses by
individuals, affecting college enrollments in predictable ways.

7. Because education is costly, jobs that require more education or training must pay a
higher wage to attract workers (that is, to compensate them for the cost of investment).

8. Age/earnings profiles are flatter for less educated workers, reflecting smaller human
capital investment costs in their early years and lower growth of productivity.

9. Post-schooling investments in on-the-job training can help account for both the
convexity and the fanning out of age/earnings profiles.

10. Post-schooling investments reduce actual earnings below potential earnings, and as
such investments decline over age, one's actual earnings approach potential.

11. Some differences between men and women in the acquisition of education and
training (including university majors) can be explained by lower rates of return to
some human capital investments among "traditional" women, who expect interrupted
labor market careers.

12. Evaluations of whether education is a good individual investment typically present

rate of return estimates that involve three sets of biases: upward biases associated
with the correlation between education and ability, downward biases associated with
the failure of monetary earnings to reflect all the benefits of a college education, and
selectivity biases arising from the fact that people who choose one career may be
more productive in that career than a comparably trained person who does not choose
that career.

13. Evaluations of whether education is a good social investment must consider the
hypothesis that education acts as a screening device, rather than an activity that
enhances productivity.

14. If the full cost of education is inversely related to ability, and if ability is positively
related to on-the-job productivity, then firms can use educational attainment as a
screening device (workers will sort themselves out according to ability in choosing
their level of educational attainment).

15. Public sector job training programs have created demonstrable earnings gains only for
adult women, and the present value of these gains typically exceed program costs.

16. (Appendix 9A) Delays in supply responses associated with the long gestation periods
of some human capital investments can create periods of oversupply followed by
periods of shortage (the "cobweb model" as it applies to the labor market).

17. (Appendix 9B) The hedonic model implies that those who obtain the most education
are least averse to learning and probably most able to learn quickly.

Answers to Even-Numbered Review Quesitons

2. "The vigorous pursuit by a society of tax policies that tend to equalize wages across
skill groups will frustrate the goal of optimum resource allocation." Comment.

Answer: As indicated in Chapter 1 and elsewhere, the optimum allocation of resources

requires that all mutually beneficial transactions be accomplished. If wages are forced
toward equality by government fiat, potentially beneficial human capital transactions
may be discouraged. That is, because the acquisition of training and education is costly,
human capital investments will not be undertaken unless there is a future return to them.
These returns normally are in the form of higher wages paid to those with the higher skill
levels, and if these higher wages cannot be paid, human capital investments that might
have been made will be discouraged. Thus, the pursuit of wage equalization across skill
levels will discourage human capital investment and may result in too few workers
entering skilled occupations.

4. When Plant X closed, Employer Y (which offers no training to its workers) hired
many of X’s employees after they had completed a lengthy, full-time retraining
program offered by a local agency. The city’s Equal Opportunity Commission
noticed that the workers Employer Y hired from X were all young, and it launched an
age-discrimination investigation. During this investigation employer Y claimed that
it hired all of the applicants from X who had successfully completed the retraining
program, without regard to age. From what you know of human capital theory, does
Y’s claim sound credible? Explain.

Answer. Y’s claim is consistent with human capital theory in two respects. First, its own
hiring and training costs appear to be negligible (we are told that it offers no training on
its own, and that its hiring standards consist of taking successful graduates of another
program). Because it makes no major investments in its workers, it therefore has no
reason to prefer younger workers. Second, because the retraining program to which X’s
former employees had access was “lengthy,” it may well be that only the younger
workers from X decided to invest in this retraining. All workers have to decide whether a
human capital investment opportunity will have expected benefits (properly discounted to
the present) that are at least equal to the costs, and a shorter period over which benefits
are received reduces these benefits. Thus, older workers are less likely to have decided to
invest in retraining – with the result that only the younger workers became qualified to
apply to Employer Y.

6. Suppose that the government, in an effort to upgrade the quality of mechanics,

promulgates legislation requiring all new mechanics to take three years of post-high
school training and to pass a competency test. Those who are currently mechanics
will not be subjected to these requirements. What are the likely labor market effects

of this legislation? Which labor and consumer groups would gain and which would

Answer: The overall effect of this requirement to license mechanics will be to reduce the
supply of mechanics and to increase their wages (average quality will rise as well.) The
labor market groups that gain are current mechanics, whose wages will be increased but
who are not subject to the licensing procedure themselves, and future mechanics of high
quality (who will not have to compete with low quality, lower cost mechanics in the
future). The labor market groups that lose include those who would have become low
cost, low quality mechanics; these potential mechanics will either be barred from the
market or will have to make training investments that they would not otherwise have
made. Put differently, the returns to training for those who would have obtained it
anyway will rise, but mandating such training will prevent some from entering the
profession and force others to undertake such training even though it might be very
costly for them in economic or psychic terms.

Among consumers, the gainers will be those who want high quality repairs and
previously had to spend time and effort to distinguish the high quality from low quality
mechanics. Losers will be poorer consumers, who may have preferred low cost, low
quality repairs to doing the work themselves, to paying the higher cost for high quality
repairs, or to having no repairs at all.

8. Many crimes against property (burglary, for example) can be thought of as acts that
have immediate gains but run the risk of long-run costs. If imprisoned, the criminal loses
income from both criminal and noncriminal activities. Using the framework for
occupational choice in the long run, analyze what kinds of people are most likely to
engage in criminal activities. What can society do to reduce crime?

Answer. Committing a crime like burglary is essentially the mirror image of a human
capital investment, because with an investment costs are borne in the present and the
returns come later. Characteristics that tend to reduce the expected costs of committing a
crime are a high discount rate (a "present orientation") and relatively poor earnings
prospects in the labor market (less to lose by being jailed).

To reduce crime society needs to reduce the immediate benefits or increase the expected
future costs of committing a crime. Reducing the benefits could be accomplished by
installing protective devices that make burglaries less likely to succeed. Increasing the
costs can be done by increasing the likelihood of catching thieves, increasing the length
of incarceration, or raising the labor-market earnings potential of those currently with the
least to lose.

Answers to Even-Numbered Problems

2. (Appendix) Suppose the supply curve for optometrists is given by Ls = -6 + .6W,
while the demand curve is given by Ld = 50 - W, where W = annual earnings in
thousands of dollars per year and L = thousands of optometrists.

a. Find the equilibrium wage and employment levels.

b. Now suppose that the demand for optometrists increases and the new demand
curve is L'd = 66 - W. Assume that this market is subject to cobwebs because it
takes about three years to produce people who specialize in optometry. While this
adjustment is taking place, the short-run supply of optometrists is fixed. Calculate
the wage and employment levels in each of the first three rounds and find the new
long-run equilibrium. Draw a graph to show these events.

Answer: a. Initial equilibrium W = $35, W = 15. (Find this by setting Ls = -6 + .6W = Ld

= 50 - W and solving for W.)

b. First round: L is still 15, so W = $51. This is point A in the figure. (Find W by
plugging L = 15 into the new Ld equation.)

Second round: Labor supply reacts to first round wage, L = 24.6, but this pushes W down
to $41.4 (at point C). Find this by plugging W = $51 into the Ls equation to find L = 24.6,
and then plugging L = 24.6 into the new Ld equation.

Third round: Labor supply reacts to second round W, L = 18.84, but this pushes W up to
$47.16 (see point E). Find this by plugging W = $41.4 into the Ls equation to find L =
18.84 and then plugging L = 18.84 into the new Ld equation.

Long-run equilibrium, W = $45, L = 21. (Find this by setting Ls = -6 + .6W = Ld = 66 - W

and solving for W.)

Suggested Essay Questions

1. Assume that a developing country with a labor force that includes many children,
ages 10-14, wants to reduce the use of child labor. Suppose that it decides to open
more schools, so that children who do not now have a school near their home can
attend a school without a long commute. Analyze how opening new schools will
affect child labor, explaining these effects fully.

Answer. Human capital theory suggests that those making decisions about investing in
schooling weigh the present value of expected future benefits against the near-term cost.
One element of cost is forgone earnings, which will be nonzero for children who could
work. A second element, however, is the cost of commuting (time and expense, perhaps
involving living away from home) to school. Making schools more geographically
accessible, then, lowers the cost of investing in human capital – and theory predicts that
more young students would attend school rather than work.

2. A study shows that, for American high school dropouts, obtaining a General
Equivalency Degree (GED) by part-time study after high school has very little payoff.
It also shows, however, that for immigrants who did not complete high school in their
native countries, obtaining a GED has a relatively large payoff. Can signaling theory
be used to explain these results?

Answer. Graduating from high school is more or less the expectation for American
students, and those who drop out may be viewed as having a low aptitude (or low
tolerance) for learning, even if they later obtain a GED. Immigrants may come from
countries in which high schools are either more demanding or less available, so dropping
out may not send the same signal of low aptitude or tolerance. If, though a GED, these
immigrants are certified as knowing the equivalent of American high school graduates,
employers may prefer them to American GED recipients, other things equal.

This chapter employs the human capital framework to analyze the phenomena of
geographic mobility and worker turnover. It demonstrates how the insights of human
capital theory can explain the observed patterns of mobility and turnover, including the
personal characteristics of those most likely to exhibit either kind of mobility.

The chapter also describes and analyzes immigration policy in the United States. It
considers the problem of illegal immigration and uses economic theory to identify the
gainers and losers from a more restrictive immigration policy. Finally, the section on
immigration concludes with an analysis of the overall effects of immigration (including
illegal immigration) on the "native" population.

The lengthy section on U.S. immigration policy is, of course, motivated by our human
capital analysis of individual migration. However, the section does not directly employ
human capital analytics. Thus, instructors who want their students exposed to human
capital analysis as it applies to geographical mobility and turnover, and who are willing
to forgo an analysis of the very topical issue of illegal immigration and what to do about
it, could save some time in the course by eliminating the section of the chapter on
immigration policy.

List of Major Concepts

1. Worker mobility can be viewed as a human capital investment, in which the benefit is
added utility in the future and the costs are the direct and psychic costs of quitting one
employer and seeking work elsewhere.

2. People will move from jobs or areas where pay is relatively low to jobs or areas
where pay is relatively high unless such mobility is inhibited by a short time horizon
(or high discount rate), costs of finding out about alternatives elsewhere, or high costs
of the move itself.

3. There is an element of self-selection in immigration because those who are most

likely to migrate are those for whom the net benefits of migration are largest
(although the benefits are often initially depressed by unfamiliarity with the language
or customs of the area to which they have moved).

4. Countries in which the earnings distributions is more compressed than in the United
States will tend to send relatively skilled workers to the United States, while those
with distributions that exhibit greater variance will tend to send less-skilled workers
to the United States. U.S. immigration has tended to become less-skilled in recent

5. Like other investments, migration investments can fail to work out as expected,
resulting in "return migration."

6. A large influx of immigrants will tend to lower wages in the relevant labor markets
and create more employment, but only in special circumstances would an influx of X
immigrants take X jobs away from "natives."

7. The argument that immigrants fill jobs no "native" would take overlooks the fact that
inducements to work in a particular occupation are not independent of the wage being

8. Immigration may increase product demand and the demand for other skill grades of

9. If immigrants receive wages equal to their marginal product, the native population as
a whole will not experience a loss of income unless the immigrants receive
government services whose value exceeds the taxes they pay.

10. "Matches" between employer and employee are improved through the process of
voluntary quits and involuntary layoffs. The human capital model can be used to
model quit behavior.

11. Human capital theory can shed light on the cyclical pattern of quit rates, the pattern of
quit rates across age groups, and international differences in quit rates.

12. The costs of quitting and searching for a new job may produce an upward sloping
supply curve to individual firms, leading to monopsonistic behavior.

Answers to Even-Numbered Review Questions

2. One way for the government to facilitate economic growth is for it to pay workers in
depressed areas to move to regions where jobs are more plentiful. What would be the
labor market effects of such a policy?

Answer. Clearly, interregional migration would be increased as a response to this

migration subsidy, and employment would grow in areas of high economic opportunity.
The effects on employment in areas of reduced economic opportunity and the effects on
wages in both areas depend on assumptions made about the labor market.

If it is assumed that the labor markets in both areas are in equilibrium and that wages in
the area of reduced economic opportunities are lower than they are elsewhere, then there
would be incentives for workers to leave the low wage area and migrate to the higher
wage area. If there are high costs of migration, some workers would not respond to a
moderately large wage differential because the increase in wages would not be sufficient
to cover the cost of migration. However, when the costs of migration are reduced by the
subsidy, more people will respond to a given wage differential and move from the
depressed area to the area of better opportunity. This move will shift the supply curve for

the depressed area to the left, shift the supply curve for the area of better opportunities to
the right, and thus reduce the wage differential between the two areas.

If the problem in the sending area is one of unemployment due to wage inflexibility in
that area (that is, the wage rate lies above the equilibrium wage), then an increase in out-
migration will reduce unemployment in the sending area, although it may not affect the
wage rate.

4. Suppose the United States increases the penalties for illegal immigration to include
long jail sentences for illegal workers. Analyze the effects of this increased penalty on
the wages and employment levels of all affected groups of workers.

Answer. Jail sentences for illegal immigrants would reduce the net gains from
immigration, even if the probability of being apprehended is small. The reduced net
benefits from migration should reduce the number of illegal immigrants coming into the
country and, assuming these immigrants are unskilled workers, this reduction should
raise wages in the unskilled labor market. Thus, unskilled native workers benefit by the
imposition of jail sentences on illegal immigrants who are apprehended. Employment in
the unskilled market goes down, although unskilled employment among natives increases.

Skilled workers may be adversely affected by this new policy because reduced
immigration may mean reduced product demand (and therefore have a scale effect on the
demand for their services). Moreover, if skilled and unskilled workers are gross
complements, the reduction of unskilled employment and the increase in the unskilled
wage may cause the demand for skilled workers to decline. If there is a leftward shift in
the demand for skilled workers, there will be a reduction in wages and employment in the
skilled markets. (Of course, if skilled and unskilled workers are gross substitutes, the
decline in demand brought about by the reduction in product demand would be mitigated
or even offset by the increased substitution of skilled workers for the now-more-
expensive unskilled labor.)

Workers from outside the United States who would otherwise have become illegal
immigrants (but who are now deterred from immigration) will be worse off. However,
those who decide to immigrate and are not apprehended will be better off because they
will receive higher wages in the unskilled market.

6. The last two decades in the United States have been characterized by a very wide gap
between the wages of those with more education and those with less. Suppose that
workers eventually adjust to this gap by investing more in education, with the result
that the wages of less-skilled workers rise faster than those of the more-skilled (so
that the wage gap between the two falls). How would a decline in the wage gap
between the skilled and the unskilled affect immigration to the United States?

Answer. Immigrants at least implicitly compare the wages they can expect in the United
States with those they can expect in their place of origin. Thus, if the American unskilled

wage rises relative to the skilled wage, the United States should experience a rise in
unskilled relative to skilled immigration.

Further, because this declining gap between the educated and the less-educated serves to
make the American distribution of earnings more equal. Skilled workers tend to come
from countries with more equal earnings distributions than found in the United States
(many European countries, for example), while unskilled workers tend to come from
countries (often developing countries) with less equal earnings distributions. When the
American earnings distribution becomes more equal, there will be reduced incentives for
skilled immigration from Europe, say, and more incentives for unskilled immigration
from countries with less-equal distributions.

Answers to Even-Numbered Problems

2. Suppose that the demand for "rough laborers" is Ld = 100 – 10W, where W = wage in
dollars per hour and L = number of workers. If immigration increases the number of
rough laborers hired from 50 to 60, by how much will the short run profits of
employers in this market change?

Answer: The short-run profits of the employer equal the area below the MPRL curve and
above the wage rate. See the figure. This will increase from area W1AB in the figure to
area W2AC. Find the wage rates by plugging the employment levels into the Ld equation.
50 = 100 – 10W initially, so W = $5 per hour. 60 = 100 – 10W after the immigration, so
W = $4 per hour. The area of the triangle is .5(base x height), so the area of W1AB is
50x$5x.5 = $125 per hour, and the area of W2AC is 60x$6x.5 = $180 per hour. Thus,
profits have risen by $55 per hour.

Suggested Essay Questions

1. Two oil-rich middle-east countries compete with each other for the services of
immigrants from India and Pakistan, who perform menial jobs that local workers are
unwilling to perform. Country A does not allow women to work, drive or go out of
the house without a chaperone. Country B has no such restrictions. Would you
expect the wages these two countries pay for otherwise-comparable male immigrants
to be roughly equal? Why or why not? Explain.

Answer. The supply of immigrants to A will be restricted, partly because female

immigrants who want to work will not go there and partly because married men will
avoid A because of the restrictions on their wives. The wages in A will have to be higher
in order to attract immigrants.

2. A particular manufacturing company employs low-skilled workers in a developing

country in which turnover among such workers is generally high. The company’s
president states that it is her goal to ensure that “workers have a serious emotional
connection to our company.” The company is unusual in that it offers air
conditioning, showers, tuition for English courses, and monthly parties. In what ways
might these policies make it difficult to retain employees, and in what ways might
they help?

Answer. The provision of these non-wage benefits is costly, and the theory of
compensating differentials suggests that their presence will serve to reduce the wages that
their workers could otherwise command elsewhere. Lower wages, by themselves, will
tend to increase turnover at the firm, especially among workers who do not value the non-
wage benefits very highly. However, these benefits will also serve to attract those who
value the amenities of parties, education in English, air conditioning, and so forth – and
because these amenities are unusual, those who place a high value on them will tend to
remain with the firm (the could not get them elsewhere).

This chapter explores in detail the relationship between compensation and productivity. It
begins with a discussion of the employment contract, which is largely implicit and
legally unenforceable. How this contract can be made self-enforcing within the context of
asymmetric information is the principal focus of this section.

After discussion of some general issues relevant to worker motivation, we turn (in
sequence) to an analysis of how motivation is affected by the basis of pay, the level of
pay, and the sequencing of pay. Thus, we discuss in turn issues related to piece rates,
commissions, profit sharing, and hourly pay (including merit pay); efficiency wages; and
deferred payment schemes, promotion tournaments and the issue of "career concerns. "
The chapter ends with a short section on two puzzles: why earnings increase with tenure
and why they increase with firm size.

List of Major Concepts

1. Productivity varies across workers and over time for a given worker, and it involves
taking the initiative in a myriad of hard-to-observe ways that advance the employer's

2. Contracts can be both formal and implicit, with the latter being incompletely
specified and hence legally unenforceable.

3. For an implicit contract to be self-enforcing, the parties can rely on signals that they
are contracting with the "right kind" of person; alternatively, they can structure the
contract so that the other party derives more from honest continuation of the
employment relationship than from reneging on their promises.

4. Dividing a "surplus" between marginal revenue product and the alternative wage is
critical to a self-enforcing employment contract; this surplus can be created by labor
investments of one sort or other.

5. Workers can be motivated to be highly productive by close supervision or by having

their earnings tied to their performance; the latter method requires that the measures
of their performance be correlated with their effort and the employer's objectives.

6. Because of group considerations, motivation techniques must take account of the

perceptions of fairness and issues of group loyalty.

7. Compensation schemes are jointly chosen by employer and employee.

8. Schemes that tie pay to individual productivity must take account of worker risk
aversion, but they are useful in eliciting signals about worker characteristics.

9. Output is not normally one-dimensional, and if pay is based on objectively-measured
aspects of output ("quantity"), workers will put forth little effort to increase output
along the other (subjective, or "quality") dimensions.

10. Group incentive schemes run the risk of creating "free rider" problems.

11. Time-based pay with merit increases can be based on either absolute or relative
output; absolute measures do not correlate as closely with individual effort, but
relative measures can induce counterproductive behaviors among employees.

12. Employees who feel generously treated by their employers may put forth greater
effort; hence, by increasing their wages employers can increase the productivity of
their workers (the efficiency wage).

13. Efficiency wages are most effective when the employer-employee tenure is expected
to be long.

14. With long expected tenures, the sequencing of pay is also an option; this option
promises handsome future rewards for current effort.

15. One scheme involves a period of underpayment followed by later overpayment,

which has both signaling value in obtaining future-oriented, hard-working employees
and offers incentives for current workers to put forth effort.

16. Promotion tournaments also have signaling and incentive value.

17. Employee "career concerns" (which can involve future payoffs with other employers)
can both distort and enhance efforts with one's current employer.

18. Concepts in this chapter contribute to the cluster of hypotheses that seek to explain
why wages rise with tenure and why large firms pay higher wages.

19. As firms increase in size it becomes increasingly costly to monitor worker effort, and
one way to cope with this monitoring problem is to pay higher wages. The higher
labor costs associated with greater firm size suggests that the labor supply curve to a
firm may be upward sloping, and this may explain monopsonistic behavior among
firms in the labor market.

Answers to Even-Numbered Review Questions

2. The earnings of piece-rate workers usually exceed those of hourly paid workers
performing the same tasks. Theory suggests three reasons why. What are they?

Answer. First, if stability of hourly pay is preferred to the potential fluctuations of piece
rate pay, then part of the "earnings premiums" enjoyed by piece-rate workers is a
compensating wage differential. Second, piece-rate pay will appeal most to workers of

above-average productivity; therefore, such workers are most likely to accept jobs paid
by the piece. Third, piece rates increase the incentives of given individuals to expend
effort, which increases their pay.

4. Suppose two soft-drink bottling companies employ drivers whose job it is to deliver
cases of drinks to stores, restaurants and businesses. One company pays its drivers an
hourly wage, and the other pays them by the number of cases delivered each day
(which can be affected by efforts of drivers to visit and sell to new customers). Which
company is more likely to experience higher rates of traffic accidents among it
drivers? Why?

Answer. Drivers paid by the piece are motivated to deliver more cases per day. They will
probably drive faster so they can make more calls, and they therefore might have a higher
accident rate.

6. Some real estate brokers split the commission revenues generated by each sale with
the responsible agent. Others, however, require their agents to pay them (the brokers)
money up-front, and then allow the agents to keep the entire commission from each
sale they make. Which agents would you predict to have the larger volume of sales,
those who split all commissions with their employer or those who pay an up-front fee
to their employer and then keep the entire commission? Explain.

Answer. There are two reasons to expect that agents paying an up-front fee will generate
more sales. First, having paid the fee, agents can keep the entire commission from each
additional sale. Their incentives to make an additional sale are therefore stronger, and
one would expect them to work longer hours and engage in more intensive efforts to
increase their sales volume. Second, the requirement to pay the broker an "employment
fee" in advance of sales will restrict interest in that kind of pay scheme to those agents
whose experience and skills give them reasonable assurance they will have a high sales
volume. Thus, the "up-front fee" scheme calls forth signals about the agents' own
expectations concerning volume (in this regard it has effects analogous to those of the
"underpayment-then-overpayment" compensation scheme described in the text).

8. An amusement park open only in the summer hires teenagers to operate its rides and
concession stands, paying them $4.00 per hour and putting aside $1.50 per hour into a
fund that they will receive as a lump-sum payment if they work through Labor Day
(typically, its biggest day of the year).

(a) What problem is the amusement park apparently trying to solve with its
compensation plan, and in what two ways does this plan help to solve the problem?
(b) Suppose the government rules that the compensation plan violates minimum
wage laws because workers who quit before Labor Day receive only $4.00 per hour.
How can the park now address the problem mentioned in your answer to (a)?

Answer. (a) The park has apparently had trouble keeping its teenage workforce through
Labor Day. This compensation plan will appeal most to job applicants who intend to

work through Labor Day (that is, it has signaling value), and it also gives current workers
incentives to stay on through then.

(b) If the plan is struck down, the park has limited alternatives owing to the seasonality of
its labor demand (which probably rules out efficiency wages and other plans based on
long-term attachments). It could invest more in screening applicants, substitute capital for
labor, or devote more resources to supervising employees (with an eye toward persuading
them to stay on the job or capitalizing on "career concerns" by providing detailed
recommendations to future employers).

Answers to Even-Numbered Problems

2. A firm is considering the adoption of a plan in which it would pay employees less
than their MRPL early in their careers and more than their MRPL late in their careers.
For a typical worker at the firm MRPL = 10 + .1T, where T = the number of years
which the worker has been employed at the firm and MRPL is measured in dollars per
hour. The worker's wage per hour is W = 8 + .2T. Assume that this wage is high
enough to attract workers from alternative jobs, that the discount rate for the firm is 0,
and that the expected tenure of a typical worker is 35 years. If workers retire after 35
years, will this plan be profitable for the firm? Explain. For how many years will the
firm "underpay" it workers?

Answer: The graph shows the wage and MRPL lines, which cross after 20 years. The
firm will adopt the plan because it expects to profit from it -- the early underpayments
exceed the later overpayments. (Because the discount rate is zero, we can find the
present value of underpayments and overpayments by simply adding them up.) Triangle
A (the initial underpayment) and triangle B (the later overpayment) are equal when T =
40 years. Triangle A exceeds triangle B if retirement comes before tenure equals 40

Suggested Essay Questions

1. Firm X is working with Firm Y on creating innovative educational materials that will
be used to create courses to be taught using the internet. Firm X is supplying the
computer specialists, while Firm Y is supplying the course content materials. Both
firms will share in the profits of the courses they jointly produce. Firm X wants Firm
Y to adopt financial incentives for its employees working on this project to finish
their work according to a deadline. Under what conditions will incentives be most
effective? Are these conditions likely to hold in this case?

Answer. Production incentives are most effective when they perfectly align the interests
of workers and owners, and when a worker’s output is only affected by his or her effort
(and not affected by factors outside the control of the individual). In this case, we must
question whether either condition is met. First, producing a high-quality course in a
brand new environment requires creativity in dealing with issues that cannot be foreseen,
so tailoring incentives to deadlines might not align the interests of employer (who wants
quality) and employee (who wants to meet a deadline). Second, X and Y are jointly
producing these programs, so delays or mistakes made by one team will affect the ability
of the other to meet the deadline; put differently, the output of Y is not completely under
the control of its workers.

2. A recent magazine article on Japan’s economic problems stated that, “As the post-war
baby-boomers reach their 50s, Japan’s lifetime-employers are carrying the cost of

paying their senior workers more than they are worth.” Is this comment consistent
with economic theory? Explain.

Answer. Yes, it is consistent. Employers offering lifetime jobs have the ability, and
often the incentive, to offer a payment scheme that underpays workers in their early years
(wage less than marginal product) and overpays them later on (wages greater than current
marginal productivity). The later overpayment serves to compensate workers for their
earlier underpayment, and it is necessary to attract workers to the firm. The purpose of
this scheme is to provide incentives for employers to stay with the firm and work
diligently (and honestly), lest they be discovered shirking and lose the opportunity for

This chapter represents a comprehensive inquiry into wage differentials across gender,
racial and ethnic groups. It begins with a section on earnings differences by gender, in
which the overall differential is broken into two parts: that associated with measurable
productivity differences and that associated with unobserved (unexplained) differences.
The latter differences are associated with (but not confined to) current market
discrimination. Discrimination is defined and problems of its measurement are discussed
in the context of analyzing gender differences in earnings.

Black-white earnings differentials are analyzed next in a subsection that includes a brief
treatment of differences in the ratios of employment to population. Earnings by ethnicity
are also discussed. In each case, the analysis includes a review of attempts to estimate the
effects of discrimination, with special emphasis on the effects of such hard-to-observe
factors as English language proficiency, cognitive achievement, and school quality.

The second major section of the chapter analyzes theories of market discrimination.
Becker's theories of employer, customer, and employee discrimination are discussed, and
the theory of statistical discrimination is explained, along with noncompetitive models of
discrimination (occupational crowding, dual labor markets, search-based monopsony,
and theories involving collusive action).

The chapter concludes with in-depth discussions of governmental efforts to reduce or

eliminate market discrimination: the Equal Pay Act of 1963 and the Civil Rights Act of
1964. Included in our discussion of the last are the evolution of the disparate impact
standard by the courts (as opposed to a disparate treatment standard), legal decisions
involving seniority, and the emerging comparable worth remedy. The chapter closes with
an analysis of the federal contract compliance program, including the standards against
which affirmative action plans are judged and the results of studies that have tried to
assess the effects of the program.

The appendix to Chapter 12 contains an introduction to the problems of estimating

comparable worth "earnings gaps." The purpose of this appendix is twofold: to give
students a brief illustration of the use of regression analysis and to show them how
comparable worth comparisons are made.

List of Major Concepts

1. Income disparities between men and women may have their roots in different
incentives to acquire productive characteristics.

2. Current labor market discrimination is said to exist when the market places values on
personal characteristics of workers that are unrelated to productivity.

3. Earnings differentials caused by differences in productive characteristics are termed

4. Occupational segregation is one form of discrimination, and it can be measured by an

index of occupational dissimilarity; however, it is difficult to distinguish between the
effects of occupational choice and those of employer discrimination.

5. To measure the extent of wage discrimination, one must determine what the earnings
ratio would be if the protected class and white males had the same productive
characteristics. However, the adjusted differential is in reality an unexplained
differential, and it could reflect the effects of unmeasured worker characteristics as
well as market discrimination.

6. Much of what appears to be labor market discrimination against women takes the
form of occupational segregation, which, while still rather marked, seems to be
declining somewhat recently.

7. When productive characteristics are controlled in an analysis of earnings differentials,

they account for all but roughly 10 percentage points of the gender wage differential.

8. Differences in the black-white employment-to-population ratio are a function of both

higher unemployment rates and lower labor force participation rates among blacks.

9. Studies using convention ally-measured variables for productive characteristics

suggest that about 11 percentage points of the observed disparity between black and
white males may be due to current labor market discrimination. Studies that control
for cognitive achievement scores as well suggest that black men earn from 8 percent
more to 8 percent less than white men with comparable productive characteristics.

10. Human capital and language-proficiency differences account for A but 3 to 7

percentage points of the Hispanic wage differential.

11. If employers discriminate against some group of workers, they will act as if they
believe the marginal product of those workers is lower than it really is. Thus, they
will hire fewer such workers than would be called for by profit maximization, and
those who are the most discriminatory will make the least profits.

12. Under employer discrimination, the behavior of prejudiced employers will reduce
demand for the minority group and cause a wage differential to exist. The size of the
differential depends on the size of the minority population relative to the distribution
of prejudiced employers in the market.

13. The implication that prejudiced employers will be less profitable suggests that
discrimination ought to be eliminated over time as nonprejudiced (profitable)
employers buy out less profitable, prejudiced employers.

14. Like employer discrimination, customer discrimination implies a shift to the left of
the demand curve for the services of a protected class. However, with customer
discrimination, a reduction in productivity is, from the employer's perspective,

15. Employee discrimination generates supply-related behavior that might cause

employers to segregate their plants by race or sex if possible. If not, wage
differentials will arise as a result of the need of employers to retain workers in the
prejudiced group.

16. Statistical discrimination arises from a screening problem in which job applicants are
evaluated both on their individual characteristics and on average characteristics of the
group to which they belong. Statistical discrimination should be reduced in situations
in which the variance of individual characteristics around the group average widens.

17. Both the crowding hypothesis and theories emphasizing the dual labor market suggest
the presence of noncompeting groups, but they do not satisfactorily explain the
creation of these groups.

18. If search costs create upward-sloping labor supply curves to individual employers,
and if discrimination raises the search costs of certain groups of workers, then
monopsonistic behavior will create wage differentials among otherwise identical

19. Some theorists use collusive action on the part of employers to explain the creation
and persistence of noncompeting groups. Employers are seen as deliberately dividing
the labor force to guard against cohesive collective action by workers, but the theory
does not explain how an employer cartel is maintained in the face of clear-cut
incentives to cheat.

20. Anti-discrimination programs by the government must set standards for both
employment and wages. If employment standards are the only ones used, prejudiced
employers may comply by paying protected-class workers less than white males. If a
wage standard is the only one applicable, then prejudiced employers will respond to
increased wages for protected classes by reducing employment.

21. A disparate treatment standard imposed under the Civil Rights Act judges that
discrimination has occurred if different procedures are used for different groups of
people and if it can be shown that there was an intent to discriminate. Proving intent
is difficult, and policies that may appear to be neutral on the surface may nevertheless
perpetuate the effects of past discrimination.

22. Courts have moved towards a disparate impact standard, by which it is labor market
results, not motivation, that counts. Under this standard, policies that lead to different
effects by race and sex are prohibited unless a business "necessity" can justify their

23. Because of occupational segregation, men and women often occupy dissimilar jobs.
The comparable worth remedy is based on comparing the skill content, responsibility,
and working conditions in jobs for purposes of pay comparisons; however, mandating
wage increases for women could reduce the incentives of employers to hire them.

24. The Federal Contract Compliance Program seeks to shift the demand curve for
protected classes to the right. Federal contractors are required to file affirmative
action plans that state their goals for hiring and promoting members of protected
classes (taking account of "availability").

25. Realistic estimates of availability should account for the compensation policy of the
firm, the willingness of workers to commute to the firm, the degree to which the firm
has incentives to train new employees, and the extent to which job applicants can be
induced to move to the firm's labor market area.

26, Studies to evaluate the effects of government anti-discrimination efforts have focused
on time series analyses of earnings ratios and effects on federal contractors
(emphasizing changes in employment levels, wages, and quit rates for protected-class

27. (Appendix) Estimating comparable worth earnings gaps typically involves evaluating
characteristics of jobs for men and women and estimating the relationship between
these characteristics and compensation for white males. This relationship can then be
used to estimate what women would receive if they were paid on a basis comparable
to men.

28. (Appendix) A precise and informative way of estimating the relationship between
point scores and compensation would be to use ordinary least squares regression
techniques to fit the "best" line through the observed points on the graph. The
estimated coefficient on the point score variable is an estimate of how much a unit
change in that variable affects earnings.

Answers to Even-Numbered Review Questions

2. “In recent years, the wage gap between skilled and unskilled workers in the United
States has grown. This growth means that measured labor market discrimination
against unskilled Mexican immigrants is also growing.” Comment on whether the
second part of this statement is implied by the first part.

Answer. Labor market discrimination is said to exist when workers who are productively
equivalent are systematically paid different wages based on their race or ethnicity (or
some other demographic characteristic unrelated to productivity). It is true that rising
inequality causes a greater gap between the average wages of native whites and unskilled
Mexican immigrants, because native whites are better educated and more skilled, on
average. However, the existence (and size) of labor market discrimination depends on
the wage gap between unskilled native whites and unskilled Mexican immigrants (that is,

between two productively equivalent groups) – so the facts quoted in the statement are
not sufficient to determine if labor market discrimination is growing.

4. Assume there is a predominantly black, central-city school district surrounded by

white, suburban districts that recruit teachers from the same pool. Black teachers are
equally willing to teach in both places, but white teachers are reluctant to take jobs in
the central city (assume their prejudice extends only to students, not to black teachers).
Assume further that there are not enough black teachers to fully staff the central city
schools. If the law requires teachers in the same district to paid equally, but allows
salaries across districts to vary, will black teachers earn more, the same, or less than
white teachers? Why?

Answer. Because there are too few black teachers to completely staff central city schools,
these schools must attract white teachers who, by hypothesis, prefer to teach in suburban
schools. Thus, to attract enough white teachers to fill the job slots, the central city schools
must raise wages (this wage increment would act as a compensating wage differential,
compensating white teachers for taking jobs they might otherwise find "distasteful").
Because salaries for all teachers within a school district must be equal, the black teachers
in the central city schools must receive the higher wages there, also.

Wages in the suburban schools will be lower than the wages in central city schools,
because these schools can attract the white teachers without offering a compensating
wage differential. Because black teachers, by hypothesis, are equally willing to work in
central city or suburban schools when wages are equal, suburban schools will not be able
to attract black teachers (who will find the higher wage at the central city schools more

6. You are involved in an investigation of charges that a large university in a small town
is discriminating against female employees. You find that the salaries for professors
in the nearly all-female School of Social Work are 20 percent below average salaries
paid to those of comparable rank elsewhere in the university. Is this university
exhibiting behavior associated with employer discrimination?

Answer. There are different relative demands and supplies by academic field that are
reflected in differential salaries. Professors of social work, therefore, may receive
relatively low wages because the supply of labor to that field is greater relative to demand.
Whatever the cause of the large relative supply of women to the social work field, it is
not obvious that this particular university is engaged in the behavior we could attribute to
employer discrimination.

Employer discrimination can be observed in two instances. One occurs when, with equal
wages for men and women, the employer clearly prefers to hire men over women of
comparable productive characteristics. The other occurs when, given a lower market
wage for women relative to comparable men, the employer fails to hire an all-female
work force. Because the university has apparently hired a nearly all-female work force in
the School of Social Work, it does not seem to exhibit this latter behavior.

8. In the 1920s South Africa passed laws that effectively prohibited black Africans from
working in jobs that required high degrees of skill; skilled jobs were reserved for
whites. Analyze the consequences of this law for black and white South African

Answer. The effects of the law on black Africans were unambiguously adverse. Blacks
were crowded into low-paying unskilled occupations, for which the wage was driven
down still further by the requirement that blacks could not do other work. Those who
would have chosen to obtain training for skilled positions were not able to do so.

The effects of the laws on white workers were ambiguous. Skilled white workers were
helped, in the sense that they received higher wages than they would have received
otherwise (had blacks been allowed into the skilled trades). Unskilled whites, however,
were probably made worse-off by this law because of its effects on the unskilled wage.
Some unskilled whites, however, reacting to the increased wage differential between
skilled and unskilled jobs, would have elected to obtain the training necessary for
entrance to a skilled trade.

Answers to Even-Numbered Problems

2. Suppose that MRPL = 20 - .5L for left-handed workers, where L = the number of left-
handed workers and MRPL is measured in dollars per hour. The going wage for left-
handed workers is $10 per hour, but employer A discriminates against these workers
and has a discrimination coefficient, D, of $2 per hour. Graph the MRPL curve and
show how many left-handed workers employer A hires. How much profit has
employer A lost by discriminating?

Answer: See the figure. A non-discriminating employer will hire left-handers until
wage = MRPL. Because 10 = 20 - .5L when profits are maximized, then L = 20 workers
for a profit-maximizing employer. Employer A, however, will hire left-handers until
wage + D = MRPL. Since 10 + 2 = 20 - .5L, then L = 16 for employer A. Lost profits
equal triangle ABC, whose area is 4 x $2 x .5 = $4 per hour.

4. (Appendix) In the market for delivery truck drivers, Ls = -45 + 5W and Ld = 180 –
10W, where L = number of workers and W = wage in dollars per hour. In the market
for librarians, Ls = -15 + 5W and Ld = 180 – 10W. Find the equilibrium wage and
employment level in each occupation and explain what will happen if a comparable
worth law mandates that the librarian wage be increased to equal the delivery truck
driver wage. Use a graph.

Answer: To find the equilibrium wage for truck drivers, set Ld = Ls and solve for W:

-45 + 5W = 180 – 10W, or 15W = 225, so W = $15 per hour

Plugging this into the two equations shows that, for truck drivers, L = 30. The calculation
for librarians is as follows:

-15 + 5W = 180 – 10W, or 15W = 195, so W = $13 per hour

For librarians, L = 60. If the librarians' wage were increased to $15 per hour, employers
would move back along the Ld curve from point A to point B on the figure below and hire
fewer librarians, reducing their employment from 60 to 40.

Suggested Essay Questions

1. Will government-mandated requirements to hire qualified minorities (at non-

discriminatory wages) in the same proportions they are found in the relevant labor
force reduce the profits of firms that formerly engaged in employer discrimination?
Fully explain your answer.

Answer. Firms that engage in employer discrimination forgo profits in order to indulge
their prejudices. Thus, requiring them to hire and pay qualified minorities in proportion
to their availability will not reduce profits. (It will, however, reduce the utility owners
derive from their businesses.)

2. Will government-mandated requirements to hire qualified minorities (at non-

discriminatory wages) in the same proportions they are found in the relevant labor
force reduce the profits of firms that formerly faced customer discrimination? Fully
explain your answer.

Answer. If customers are prejudiced, they will tend to avoid businesses hiring workers
from groups they are prejudiced against; thus, profit-maximizing employers will prefer to
hire workers from groups that customers do not have distaste for. If the law requires
employers to hire from all groups proportionately, firms previously attracting prejudiced
customers will lose business (either to firms that might not be covered by the law, or
when consumers substitute other goods or services for the one in question); however,
their costs could go down now that there is no reason for paying a premium to workers

from “favored” groups. Firms previously attracting non-prejudicial customers will now
be unable to capitalize on the lower demand (and lower wages) for minority workers, so
their costs will rise.

The major focus of this chapter is on the economic effects of unions, in both the private
and public sectors. It begins with some necessary definitions and descriptions of
unionism in the United States compared to elsewhere in the world, before turning to
elementary coverage of major pieces of labor legislation in the United States.

In seeking their objectives, unions are constrained by the demand for their members'
services. Unions facing relatively inelastic demand curves are better able to raise wages
without adversely affecting employment levels very much. The simplest model of unions'
objectives is the "monopoly union" model, in which the union sets the wage and the
employer adjusts by setting the employment level. A more complex model is the
"efficient contracts" model, in which the union and firm jointly bargain over wage and
employment levels.

In attempting to explain the major activities of unions, we have organized the analysis
around the demand for unions by workers and the supply of union services by labor
unions. We use this analysis to help understand the major trends in American
unionization. No discussion of union behavior would be complete without an analysis of
strike activity and (for the public sector) interest arbitration. The section on strikes
include an exposition of the Hicks model of bargaining and the Ashenfelter-Johnson
political model of strike behavior (including the effects of the Landrum-Griffin Act). The
section on arbitration discusses the contract zone in the context of both conventional and
final-offer arbitration.

Having analyzed some key characteristics of union behavior, we turn to an analysis of the
effects of unions on wages and other workplace outcomes. The measurable effect of
unions on wages is the relative wage advantage, found by comparing union wages to
wages in the nonunion sector. The true (or absolute) effects of unions on wages are not
measurable, and the possible biases inherent in measuring the absolute effects using
relative-effect measures are discussed at length. We close the chapter by summarizing
empirical evidence on the effects of unions on relative wages, total compensation,
employment, productivity, and profit, and discuss both "traditional" and alternative views
of union effects on the overall social welfare.

The appendix to Chapter 13 analyzes how the uncertainty of arbitrators' decisions affects
bargaining outcomes. As in the appendix to Chapter 8, the discussion defines and
illustrates the concept of risk aversion..

List of Major Concepts

1. Union membership as a fraction of the population in the United States is low relative
to the other major industrial countries, and American union activities are relatively

2. Unions are constrained in their objectives by the labor demand curve.

3. The monopoly-union model assumes that the union sets the wage and the employer
sets the employment level.

4. The "efficient-contracts" model emphasizes that if unions and employers bargain only
over wages, the resulting employment/wage outcome will be inferior to another set of
outcomes that would improve the welfare of both parties. Under this model, the two
parties bargain over both wage and employment levels.

5. The decline in American union membership can be understood, in part, by reference

to shifts in the demand and supply curves for union services (caused by the
feminization of the work force, a change in industrial composition, regional shifts,
competitive pressures, and employer resistance).

6. Unions often take actions designed to shift the demand curve for labor to the right
and/or to reduce the elasticity of demand for union labor.

7. Strikes are intended to impose financial costs on employers if they do not agree to
union offers, but because they also impose costs on employees, both sides become
more willing to make concessions as strike duration increases. The Hicks bargaining
model suggests when strikes will end, but it does not explain why strikes occur in the
first place.

8. While strikes can occur because one party mistakes the other's true position, they can
also occur because, in the context of asymmetric information, one party wants to
elicit a signal from the other about its true preferences or constraints.

9. The Ashenfelter-Johnson model of strike activity, which is tripartite in nature, views

union leaders and union members as sometimes having conflicting perspectives.
Union leaders, who have better information than their members, may pursue the twin
goals of maintaining their positions in the union and educating union members by
recommending a strike.

10. Unions' propensities to strike vary over the business cycle and have trended down
over time, but the increase in union democracy associated with the Landrum-Griffin
Act caused a one-time increase in strike activity (as predicted by the Ashenfelter-
Johnson model).

11. Interest arbitration is used in the public sector, where strikes are generally not
permitted, and interest arbitration can be "conventional" or "final offer."

12. The "contract zone" into which pre-arbitration offers will fall can be widened by both
the uncertainty about what an arbitrator might decide and the parties' aversion to the
risk of an adverse outcome. However, it is not clear whether a wider contract zone
increases or reduces the chances a dispute will go to arbitration.

13. One would like to measure the effects of unions on wages in the absolute (comparing
wages in a world with unions to wages in a world without); however, measuring this
absolute wage advantage is impossible. We can only measure union wages relative to
nonunion wages, but this is not a good measure of the absolute effect because the
presence of unions alters the nonunion wage also.

14. If the nonunion labor market is in equilibrium, the presence of unions should lower
the nonunion wage below what it would have been otherwise and cause the relative
wage effect to overstate the absolute effect of unions.

15. If employers raise wages in the nonunion sector to keep unions out (thus creating
unemployment in that sector), the relative wage effect is smaller than the absolute
union wage effect.

16. The presence of wait unemployment in the union sector will inhibit the growth in
labor supply in the nonunion sector and could even cause the supply curve there to
shift left. Whether wait unemployment causes the relative wage effect to be greater or
smaller than the absolute wage effect depends on whether the supply curve of labor to
the nonunion sector shifts to the right or left.

17. In an overall sense, unions appear to raise the wages of their members above the
nonunion wage by something like 10 to 20 percent. These wage effects are larger in
the private than the public sector, larger in the United States than elsewhere, and
largest among unskilled (and minority) workers.

18. Union effects on employee benefits as a percentage of total compensation tend to be

positive; thus, the total compensation effects of unions may be greater than the
relative wage effects.

19. The compensation advantages enjoyed by union members; however, may be in part a
compensating wage differential for the more structured, more hazardous and less
flexible work settings in unionized firms.

20. Empirical evidence tends to suggest that, in the United States, unionization reduces
employment or employment growth, has an ambiguous effect on productivity, and a
negative effect on profits.

21. The traditional view of union effects stresses the social loss caused by inequality in
wages (and marginal productivities) among workers of comparable skill. The
traditional view also stresses the losses associated with union staffing requirements,
restrictive work practices, and strikes.

22. The alternative view of unions is that they provide a method of collective "voice" that
tends to reduce turnover and encourage firms to provide specific training to
employees. Improved communications between labor and management may also
increase productivity and worker motivation directly, but employer resistance to

unions and the degree to which stock market prices are depressed when union
organization drives are initiated suggest a widespread belief that unions reduce

23. (Appendix) When collective bargaining impasses can be decided by arbitration, the
range of possible nonarbitrated settlements is widened by greater uncertainty about
how the arbitrator will decide (if an agreement is not reached) and by greater risk
aversion on the part of the parties.

Answers to Even-Numbered Review Questions

2. Some collective bargaining agreements contain "union standards" clauses that

prohibit the employer from farming out work normally done in the plant to other
firms that pay less than the union wage.
a. What is the union's rationale for seeking a union standards clause?
b. Under what conditions will a union standards clause most likely be sought by a
labor union?

Answer. (a) The union's goal for seeking the union standards clause is to remove
incentives for the employer to substitute cheaper nonunion labor for more expensive
union labor.

(b) The ultimate goal of the union, of course, is to raise wages while preserving
employment (or at least not having to undergo large employment declines). A union
standards clause will be more attractive to a union when the firm can more easily
substitute outside factors of production for those it employs and when these substitution
effects are large. Analogously, the union standards clause will be more successful in
preserving employment of union members if the firm finds it more difficult or expensive
to substitute capital for labor within the firm. Likewise, if the supply of capital to the firm
is relatively inelastic, any tendency to substitute capital for labor will be met with a rising
price of capital, which will mitigate the amount of capital/labor substitution that might
otherwise accompany a union standards clause. Moreover, a union standards clause will
be more successful in preserving employment, and therefore more sought-after, if the
product demand curve is relatively inelastic (so that scale effects are small).

4. It has been observed that unions in the capital-intensive steel industry were able to
negotiate higher-than-average wage increases during the very period in which steel
output in the United States was declining. Using economic theory, how can this
pattern be explained?

Answer. When output is contracting, one would think that the shrinking demand for
workers (owing to the scale effect) would reduce a union’s ability to negotiate wage
increases without harming job opportunities for its members. However, when output is
contracting, employers are not adding to their capital stock or opening new plants, so
technological improvements that substitute capital for labor are not easily made. The

reduced ability to substitute capital for labor serves to strengthen the union’s hand,
because even if wages rise the substitution of capital for labor may not take place.

6. In the mid 1980s the teachers' union of a large American city was given a choice: it
could accept a 10 percent cut in the salaries paid to teachers and suffer no employment
losses, or it could keep salaries constant and accept a 10 percent cut in employment
levels (and a corresponding ten percent increase in class sizes). Given that union
leaders are elected by the membership, please answer the following:

a. Predict and explain the union's decision, assuming that its collective bargaining
agreement with the city specifies that any layoffs will occur among those
teachers most recently hired.

b. Explain whether the decision in (a) would have been different if the collective
bargaining agreement had specified that all layoffs would occur on a random
basis, independent of seniority, teaching field, or any other teacher

Answer. (a) If layoffs occur only among the most recently hired, it is possible for each
union worker to calculate whether he or she is among the group to be laid off. Therefore,
90 percent of all union members will know that they will not be laid off, and 10 percent
know for sure that they will be. Because the union leaders who had to make this decision
are elected by majority rule, it is a safe bet that they would act in the interests of the
majority (who would not be laid off) and would therefore choose to maintain current
salaries and accept layoffs of 10 percent.

(b) If layoffs are to be made randomly, then each union member has a 10 percent chance
of being laid off if salaries are held constant. Alternatively, members could accept a 10
percent cut in their salaries with the assurance that they would not be laid off. In either
case, in advance of the decision that must be made by the union, each worker would face
a 10 percent decrease in expected wages. In this case, it is much more likely that the
union would choose to accept salary cuts than cuts in employment. A simple political
model of union decision making could not predict for sure that unions would accept
salary cuts in this case, although they probably would if their workers were risk averse.
However, compared to the case in which layoffs are made among those most recently
hired, the chances of the union's deciding to accept employment cuts are much lower.

8. Is the following statement true, false, or uncertain? “The empirical studies indicating
that unions raise the wages of their members by 10 to 20 percent relative to the wages
of comparable nonunion workers imply that unions have a negative effect on national
output.” Explain your answer.

Answer. The traditional view is that when workers of comparable potential are in jobs
with different productivities, society’s output could be increased if some moved from the
low- to the high-productivity jobs. However, if unions solve problems created by market
imperfections (workers can use “voice” rather than “exit,” for example), they may
enhance society’s total output.

Answers to Even-Numbered Problems

2. The Brain Surgeon's Brotherhood faces an own-wage elasticity of demand for its
labor that equals -0.1. The Dog Catcher's International faces an own-wage elasticity
of demand for its labor that equals -3.0. Suppose that leaders in both unions push for
a 20 percent wage increase, but have no power to directly set employment levels.
Why might members of the Dog Catcher's International be more wary of the targeted
wage increase?

Answer: Own-wage elasticity of demand = %(quantity demanded)/%(wage). In the

case of the Dog Catchers this implies -3 = %(quantity demanded)/20%, thus
employment will fall by 60% if they boost wages by 20 percent. In the case of the Brain
Surgeons this implies -0.1 = %(quantity demanded)/20%, thus employment will fall by
only 2 percent if they boost wages by 20 percent.

Suggested Essay Questions

1. American unions often try to win public support for boycotting goods made in less
developed countries by workers who work very long hours at low pay in unhealthy
working conditions. (a) If successful, will these efforts unambiguously help the
targeted foreign workers? Explain fully. (b) Will they unambiguously help the
union’s American workers? Explain fully.

Answer. (a) If manufacturers in these less developed countries are induced to raise
wages and improve working conditions, output price will tend to rise and only those
consumers willing to pay the higher prices will remain in the market. Boycotting goods
made by workers in “sweatshop” conditions, then, tends to reduce demand (because of
both scale and substitution effects) for the services of low-wage labor, and these workers
may end up losing their manufacturing jobs and having to go back to farming or some
other job they previously felt was inferior, given their preferences and opportunities.

(b) If manufactured items from abroad become more expensive, we would expect that
consumers would tend to substitute American goods (now relatively cheaper) for foreign-
made goods. Thus, foreign and American workers may be considered substitutes in
production, and if this substitution effect is dominant, the will be gross substitutes. If so,
when the costs of foreign labor rise, the demand for American labor may rise. However,
we must also consider the scale effects associated with higher costs overseas. Some
American workers may have jobs (in packaging, selling, distributing) that depend on the
scale of output produced both in the U.S. and abroad. Clearly, the scale of output will
fall owing to the boycott, and the costs of producing the items in question will rise; some
workers, therefore, may be gross complements with foreign workers, and these workers
will be worse off if the boycott lasts and is successful.

2. A certain country has very centralized collective bargaining, under which wage
bargains are applied nationally. This country is thinking about adopting a bargaining

structure that is more decentralized, so that wage bargains will be made at the
individual plant or firm level. How would you expect decentralization to affect
wages and employment? Explain why.

Answer. Union power to raise wages is affected by the elasticity of demand for its
workers; where elasticity is higher, given wage increases will result in greater loss of
employment. A national union may perceive the product demand (and hence labor
demand) of employers to be relatively inelastic, because if it is able to raise wages in all
sectors, consumers have limited incentives for substituting goods from one sector for
goods from another. All employers will face higher wages and higher costs – and thus
have higher prices – although the degree to which their output prices will rise depends on
the share of labor in total cost and their ability to substitute for labor in the production
process. When bargaining is very decentralized, however, the unions involved in wage
setting will perceive the output demand elasticity to be relatively elastic, and they may
moderate their demands as a result. Thus, at a broad level of analysis, we might expect
wage increases to be larger, and employment losses to be larger, with centralized

At a more disaggregated level, however, we would expect that local unions facing labor
demand elasticities that are lower than the national aggregate to bargain for greater wage
increases than they got under centralized bargaining, while those with elasticities less
than the national aggregate would bargain for less. Thus, wage increases will be greater
than what they would have been under centralized bargaining in some sectors, and less in
other. Overall, a greater ability to tailor local demands to local elasticities should result
in a reduced level of job loss associated with unionization.

This chapter is intended to accomplish two purposes: to analyze changes in earnings
inequality after the 1980s and to review major concepts of economic theory introduced in
prior chapters. It begins with a section on measuring inequality and then moves to one
that describes changes in the 1980s and early 1990s along various dimensions: the
occupational distribution, relative wages, hours of work, and earnings dispersion within
narrowly-defined human capital groups.

The underlying causes of growing inequality are then grouped into supply factors,
institutional changes, and demand-side influences. Empirical studies are surveyed, with
the conclusion that demand factors were dominant in the 1980s (especially computer-
related technological changes that affected the mix of productive factors, leading to
increased relative demand for educated workers). The appendix discusses the derivation
of Lorenz curves and Gini-coefficients.

List of Major Concepts

1. Earnings inequality is a function of the dispersion of the earnings distribution, and

this dispersion can be measured in various ways, which differ in the ease with which
they can be completed and widely understood.

2. The most widely-used measures involve ranking the population and analyzing
earnings by percentile (comparing either shares of the total received by a group or the
earnings levels at percentile boundaries).

3. Earnings distributions for both men and women became "stretched" in the 1980s,
more because wages of those in the upper end grew relative to others' than to a
movement of jobs from the middle of the distribution to either end.

4. The most notable change in earnings for both men and women was the increase in the
relative earnings of more-educated workers; for men this increase occurred mainly
because earnings of the less-educated fell in real terms, while for women it was
associated mainly with the increased real earnings of more-educated women.

5. The returns to experience rose modestly, but only for the less-educated.

6. Changes in the relative hours of work played no role in the growth of inequality.

7. Earnings also became more dispersed within human-capital groups.

8. Growing disparities could result from labor supply, institutional, or labor demand

9. The fact that skilled employment grew faster than unskilled employment in the 1980s
appears to rule out supply changes as the dominant factor underlying the growth of

10. Such institutional factors as a "frozen" minimum wage and declining unionization
could have played a role in the growing inequality, but they apparently played a
minor one.

11. Both product demand changes and changes in the mix of productive factors
contributed to growing inequality, with the latter changes the dominant force in the
growth of education-related differentials.

12. The growth of contingent-pay plans might underlie the increased dispersion within
human-capital groups, but definitive work has yet to be done.

13. (Appendix) The Lorenz curve and Gini coefficient are related measures of disparities
that are based on groups' shares among the total.

Answers to Even-Numbered Review Questions

2. Assume that the "comparable worth" remedy for wage discrimination against women
will require governmental and large private employers to increase the wages they pay
to women in female-dominated jobs. The remedy will not apply to small firms. Given
what you learned earlier about wages by firm size and in female-dominated jobs,
analyze the effects of comparable worth on earnings inequality among women. (For a
review of relevant concepts, see Chapter 12.)

Answer: The comparable worth remedy will have contradictory effects on the dispersion
of earnings if it is applied only to large private or governmental employers. The reason is
that the employers to which the comparable worth remedy will apply are the highest-
paying employers, but the jobs to which it will apply are among the lowest-paying jobs.
Raising the wages of the lowest-paying jobs will tend to equalize the distribution of
earnings, but because the remedy applies only to the highest-paying firms, there may be
offsetting tendencies. In particular, there may be downward pressure on wages in female-
dominated jobs in the small business sector as employment shifts from sectors
experiencing large wage increases to "uncovered" sectors. Thus, it is possible that the
lowest-paying jobs with the lowest-paying employers will actually experience reductions
in wages that could widen the dispersion of earnings. In summary, women in the very
lowest part of the income distribution might experience wage reductions, while those
slightly above them will experience wage increases; the more dominant of these two
tendencies is difficult to forecast.

4. Proposals to tax health and other employee benefits, which are not now subject to the
income tax, have been made in recent years. Assuming that more highly paid workers
have higher employee benefits, analyze the effects on earnings inequality if these tax
proposals are adopted. (For a review of relevant concepts, see Chapter 8.)

Answer: If employee benefits are subjected to the income tax, the receipt of
compensation in the form of "in-kind" or deferred benefits becomes less attractive.
Workers now receiving a substantial fraction of their overall compensation in the form of
benefits may decide that they would prefer to have reduced benefits and receive a higher
fraction of their pay in cash earnings. Assuming overall (pre-tax) compensation levels are
not changed by this new law, its main effect will be to shift the compensation of highly-
paid workers away from benefits and toward even higher levels of cash earnings. If by
the "earnings distribution" we mean (as we did in the text) cash earnings, then the effect
of this benefits tax will be to widen the dispersion of earnings in society. Widening the
dispersion of earnings, however, is offset by a narrowing of the dispersion of employee
benefits received, with the result that the overall distribution of total pre-tax
compensation might remain essentially unchanged. (The distribution of post-tax
compensation, however, is more equal than before, owing to the increased taxes paid by
highly-compensated workers.)

6. Discuss the role of geographic mobility in decreasing or increasing the dispersion of

earnings. (For a review of the relevant concepts, see Chapter 10.)

Answer: Geographic mobility is, at least in part, a response to earnings inequality across
geographic areas. People tend to move from areas with low opportunities (wages) to
places where they believe they can improve their earnings, at least in the long-run. Thus,
geographic mobility should tend to equalize earnings across areas -- driving up wages in
low-wage areas (as workers move out) and driving down wages in high-wage areas.
While geographic mobility may do little to shrink the gap between the wages of highly-
educated workers and those of high school dropouts say it does serve to reduce the
disparity of earnings within human-capital groups.

Answers to Even-Numbered Problems

2. Suppose that the wage distribution for a small town is given below.

Sector Number of Workers Wage

A 50 $10 per hour

B 25 $5 per hour

C 25 $5 per hour

Then a minimum wage law is passed that doesn't affect the market in high-wage
sector A, but boosts wages to $7 per hour in sector B, the covered sector, while
reducing employment to 20. Displaced workers in sector B move into sector C,
where wages fall to $4.50 per hour as employment grows to 30. Has wage inequality
risen or fallen? Explain.

Answer: There are several ways to measure inequality. On the one hand, the share of
income going to the bottom half of the income distribution has risen (from $250/$750
= .333 to $275/$775 = .355), thus wage inequality has fallen by this measure. On the
other hand, the 80th-20th percentile wage ratio has risen (from $10/$5 = 2 to $10/$4.50 =
2.22). There is no unambiguous answer to this question.

Suggested Essay Questions

1. Would a higher minimum wage promote more wage equality? Discuss thoroughly.

Answer. As discussed in Chapter 4, one possible effect of a higher minimum wage is for
the “spillover effect” to cause lower wages in the uncovered sector. In this case, wages
rise for some low-wage workers, but fall for others – and the effects on wage inequality
are ambiguous. (Of course, there are different ways of measuring wage inequality, and
different results could be obtained by different measures. For example, if workers
covered by a new minimum wage were at the 20th percentile of the wage distribution,
while those who are uncovered are at the 10th, an 80-20 measure would show
improvement, but a 90-10 measure would indicate greater inequality.)

2. Countries X and Y both have agricultural and industrial sectors. Historically, they
have levied tariffs on each other’s exports, which have had the effect of reducing
trade between the two countries. X and Y now agree to drop these tariffs. Will
greater trade between the two countries lead to increased real incomes in each country?
Will greater trade lead to increased real wage equality within each country?

Answer. As discussed in Chapter 4 and its appendix, moving from restricted to free trade
will have effects similar to technological change (in that the principle of comparative
advantage will lead to more efficient production of goods and services). Technological
change will increase real incomes in both countries, which can now take advantage of
cheaper goods and services. However, technological change does not imply anything
about changes in wage equality. Comparative advantage may call for more of one kind
of worker in a given country to be used, and less of another, depending on the ratio of
marginal product to wages for the two kinds of workers in X and Y (clearly, these
sectoral shifts do not depend on wage levels alone).

This chapter defines the unemployment rate and discusses its strengths and weaknesses
as a measure of economic welfare. Further, it emphasizes that while the unemployment
rate is a stock concept, there is an underlying set of flows into and out of that stock. Thus,
different demographic groups may have different unemployment rates because of
different propensities to quit or be laid off from jobs.

Next, we discuss four general types of unemployment: frictional, structural, demand

deficient, and seasonal unemployment. In the section on frictional unemployment we
discuss search theory and the effects of unemployment benefits on job search. When
analyzing structural unemployment, occupational and geographical imbalances,
government policies, and efficiency wages (including the "wage curve") are discussed.
The section on demand deficient unemployment analyzes wage rigidity, the financing of
unemployment benefits, and policies and adjustments found in Europe.

The chapter also treats the issue of "full employment." The structure of unemployment
rates across demographic groups is presented, and we then discuss effects of the
changing age, race, and sex composition of the labor force on the full employment rate of

List of Major Concepts

1. The unemployment rate is the number of nonemployed people seeking work divided
by the number of people who are in the labor force. While this measure has a number
of serious drawbacks, it remains a useful indicator of labor market conditions.

2. While one can think of a stock of unemployed persons, this stock is constantly
changing, focusing only on the stock masks the highly dynamic nature of labor
markets. Unemployment rates are affected by layoffs and quits, new hires and recalls,
retirees from and new entrants into the labor market. As these flows change relative
to each other, the unemployment rate changes in predictable ways.

3. Frictional unemployment occurs because labor market information is imperfect. It

takes time even in the best of markets for unemployed workers and employers with
job vacancies to find each other. Government efforts to improve the rapidity of the
job-matching process could reduce the extent of frictional unemployment.

4. A period of unemployment can be looked upon as a period when job search can be
undertaken, and the extent to which someone will spend time searching for work
depends upon the expected benefits and costs of continuing search.

5. Since the receipt of unemployment insurance benefits reduces the costs of extra time
spent searching for work, the more generous these benefits are, the longer
unemployed workers will tend to search for jobs. Thus, more generous

unemployment insurance benefits may be expected to increase the duration of

6. Structural unemployment occurs when flows of workers into and out of particular
labor markets -- defined by skill or geography -- are impeded. Government efforts to
subsidize mobility and/or job training would serve to reduce the incidence of
structural unemployment.

7. Structural unemployment is also associated with firms' decisions to pay efficiency

wages; in such cases, some workers without jobs "wait" for jobs in the high-paying
sector and do not seek jobs in lower-paying firms.

8. Efficiency wages may underlie the "wage curve" found in almost all countries studied;
this curve plots a negative relationship between the regions' wage rates and their
unemployment rates.

9. Demand-deficient unemployment occurs when the aggregate demand for labor

declines in the face of downward inflexibility of nominal or real wages.

10. There are several reasons for the downward inflexibility of money wages, including
the preference of unions for layoffs over wage cuts, the asymmetry of information
between employers and employees, firm-specific training, risk aversion of older
workers, concerns about status, and implicit contractual agreements that after an
initial period, when the risk of layoff is high, yearly earnings of workers will stabilize.

11. The government's tax treatment of most unemployment insurance benefits and its
failure to perfectly experience-rate the unemployment insurance charges levied on
employers lead to a higher layoff rate than would otherwise prevail.

12. In Europe, cyclical fluctuations in labor demand are more likely to be manifest in
adjustments in hours, not employment levels. One reason for the greater adjustments
in hours of work may lie in the wider use of partial unemployment compensation
benefits in Europe.

13. Seasonal unemployment may be associated with changes in the weather or with
model changeovers, but it is also affected by downward rigidity of money wages and
the government's unemployment insurance program.

14. The full-employment rate of unemployment (i.e., unemployment in "normal" times)

appears to have changed in recent years, in part due to the changing proportion of
women, teenagers, and ethnic minorities in the labor force.

Answers to Even-Numbered Review Questions

2. Government officials find it useful to measure the nation's "economic health." The
unemployment rate is currently used as a major indicator of the relative strength of
labor supply and demand. Do you think the unemployment rate is a useful indicator
of labor market tightness?

Answer. There are many problems with using the aggregate unemployment rate as an
indicator of labor market tightness. It provides no information about the number of
individuals who are not actively searching for work because they were unsuccessful in
the past. It tells us nothing about whether the employed are working at jobs
commensurate with their skill levels. It does not distinguish between unemployment of a
skilled adult and unemployment of an unskilled new entrant into the labor force. It tells
us nothing about the number of workers who are employed. Finally, it is sensitive to the
generosity of some social programs, like the unemployment insurance system. It is
consistently measured over time, however, so that – despite its weaknesses – we can
measure changes in labor market conditions from one year to the next.

4. Is the following assertion true, false, or uncertain? "Increasing the level of

unemployment insurance benefits will prolong the average length of spells of
unemployment. Hence, a policy of raising UI benefit levels is not socially desirable."
Explain your answer.

Answer. The social desirability of raising unemployment insurance (UI) benefits depends
on a number of factors. From an efficiency perspective, the costs of higher UI benefits
are the prolonged spells of unemployment they induce (they encourage unemployed
workers to search longer). The benefits of higher UI benefits are the better job matches
they induce, as they provide unemployed workers with the resources to continue to
search for jobs more commensurate with their skill levels. If individuals find higher-
paying jobs that match their skills more closely, output will increase and the probability
that these workers will quit their jobs in the future will decline. The empirical magnitudes
of these various effects must be evaluated before one can decide if it is desirable to raise
UI benefits.

6. In the 1970s Sweden adopted several new labor market policies affecting layoffs.
Three were notable: (1) Plants that provided in-plant training instead of laying off
workers in a recession received government subsidies; (2) All workers had to be
given at least one month's notice before being laid off, and the required time in the
average plant was two to three months; (3) Laid-off workers had to be given first
option on new jobs with the former employer. What probable effects would these
policies, taken as a whole, have on wages, employment, and unemployment in the
long run?

Answer. Policy number one clearly increases the incentives of firms to retain workers
during a recession. Therefore, this policy will probably reduce unemployment associated
with demand deficiency. If the training that is subsidized is useful and geared to market
demands, then the substitution of training periods for periods of unemployment may

reduce structural unemployment as well. That is, firms could train their workers to fill
newly created jobs for which their current work force was not trained, or it could use the
subsidy to provide training to its workers that other firms might find useful. This latter
policy might be pursued by firms whose labor demand was shrinking and who might find
it more profitable to induce employees to quit rather than undergo the expense of laying
them off.

Requiring advance notice of layoff to employees raises labor costs because firms are no
longer free to lay off workers the instant marginal productivity slips below the wage rate.
While the short-run effects of this may indeed inhibit layoffs, the longer-run effect is that
the rehiring of labor when the market improves will be discouraged. That is, by raising
the costs of labor, the firm will be induced to scale down its operations and to substitute
capital for labor. The slowing down of the post-recession recall of laid-off workers and
the possibility of reduced labor demand could cause unemployment to rise (or remain)
higher than it otherwise would. Of course, this conclusion would change if wages were to
fall when this new "fringe benefit" of advance notice was mandated.

A requirement that laid-off workers be given first option on any new jobs with the former
employer might prevent such employers from hiring younger, less experienced workers,
and could cause the composition of unemployment to change. That is, the unemployment
rate for older workers might fall and the unemployment rate among younger workers
might be expected to rise as a result of this policy. By constraining firms to hire older,
displaced workers from declining departments in their expanding departments, the
requirement raises labor costs and will tend to cause employment growth to be smaller
than it would otherwise be. This, of course, could contribute to unemployment in the long

8. The present value of benefits in many pension plans are larger if a person retires
before the normal retirement age. In short, there is a large inducement for many
private sector workers to retire early. What effect will increasing the inducements to
retire early have on the unemployment rate of older men? Fully explain your answer,
making use of the assumption that retired workers withdraw from the labor force and
do not seek or obtain other jobs.

Answer. It is possible that increased inducements for early retirement will increase the
unemployment rates of older workers. The reason for this is that they may induce older
workers who are currently employed to withdraw from the labor force earlier than they
otherwise would, thereby increasing flows from employment to "out of the labor force."
When flows from employment to out of the labor force are increased, other things equal,
the unemployment rate will rise because there is a fall in the labor force without a
corresponding fall in unemployment.

It is possible, of course, that the older workers deciding to retire early are predominantly
those who have just been laid off -- perhaps permanently -- by their employers. They
may believe that, rather than continuing to search for work, they would be better-off
accepting their pension and dropping out of the labor force. Obviously, the more

generous their pension is prior to normal age of retirement, the greater is the likelihood
that they will drop out of the labor force after layoff. Thus, it is impossible to say whether
more generous early retirement benefits tend to increase or reduce the unemployment
rate of older men.

Answers to Even-Numbered Problems

3. Suppose that initially the Pennsylvania economy is in equilibrium with no

Ls = -1,000,000 + 200W and Ld = 19,000,000 – 300W, where W = annual wages and L
= number of workers. Then structural unemployment arises because the demand for
labor falls in Pennsylvania but wages there are inflexible downward and no one
moves out of state. If labor demand falls to Ld = 18,000,000 – 300W, how many
workers will be unemployed in Pennsylvania? What will be its unemployment rate?

Answer: Find the initial wage and employment level by solving as below:

-1,000,000 + 200W = 19,000,000 – 300W, so

500w = 20,000,000, and W = $40,000

With W = $40,000, L = 7,000,000. Next, find the gap between Ls and Ld at W = $40,000
after the labor demand curve shifts:

Ls = -1,000,000 + 200x40,000 = 7,000,000

Ld = 18,000,000 - 300*40,000 = 6,000,000

This gap, Ls - Ld, shows that unemployment equals 1,000,000. The unemployment rate
is (1,000,000/7,000,000)x100 = 14.3%.

Suggested Essay Questions

1. “With the growth of free trade, Mexican employers have sought to reduce union
control over internal labor markets, and they have eliminated promotion by seniority,
rules against subcontracting, and restrictions on the use of temporary workers – all in
the name of greater flexibility.” Would you expect greater employer flexibility in
hiring and assigning workers to increase or decrease unemployment in Mexico?

Answer. The level of unemployment in any country is a function of how fast workers
flow into unemployment as compared to how fast they flow out. The above changes in
Mexican employment conditions will probably tend to increase flows into unemployment,
but by allowing employers to be more flexible, they lower the costs of creating new jobs.
Hence, flows out of unemployment will tend to increase as the pace of job creation is
enhanced. Even the flows into unemployment will be reduced, however, if the changes

permit labor costs to be more flexible in a downward direction during periods of falling

2. One student of the labor market effects of free trade argues that the government
should offer “wage insurance” to workers who lose jobs because of free trade. Under
this proposal, the government would replace a substantial portion of lost earnings if,
upon re-employment, eligible workers find that their new job pays less than the one
they lost. This wage insurance would be available for up to two years after the initial
date of job loss. Would this wage insurance program reduce unemployment?

Answer. Wage insurance should cause unemployed workers to take offers they would
normally have rejected, thus increasing the flows out of unemployment and into jobs –
which, by itself, would reduce unemployment. Two factors might serve to increase the
flows into unemployment, however, which could increase the level of unemployment.
First, if workers take jobs hastily (the clock starts ticking upon job loss), they may be
poorly matched with their employers – and may quit or be fired after the insurance runs
out. Second, employers may be less reluctant to lay off workers, knowing that they have
wage insurance for two years after layoff.