Вы находитесь на странице: 1из 14

American Economic Association

A Child's Guide to Rational Expectations


Author(s): Rodney Maddock and Michael Carter
Source: Journal of Economic Literature, Vol. 20, No. 1 (Mar., 1982), pp. 39-51
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/2724658 .
Accessed: 25/06/2013 16:54
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to Journal
of Economic Literature.
http://www.jstor.org
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
Journal of Economic Literature
Vol. XX (March 1982), pp. 39-51
A C hild ' s G uid e to
Rational Expectations
By
RODNEY MA DDOC K A ND MIC HA EL C A RTER
Res earch School of Social Sciences
A us tralian National Univers ity
DRA MA TIS PERSONA E
(In ord er of s peaking)
Ernie, firs t s tud ent, is s omething of a Keynes ian.
Bert, s econd s tud ent, is more inclined to monetaris m.
Scene i Prologue
Scene ii The Id ea of Rational Expectations
Deriving the Impotence Res ults
C riticis ms
Scene iii Tes ting
Significance
C onclus ion
A ppend ix A A ggregate Supply
A ppend ix B A lgebra of the Mod el
References
Scene i. Prologue
(Two s tud ents s haring coffee in the union of
an A us tralian univers ity.)
Ernie: Did you read that rid iculous article in
C hallenge the other d ay?
Bert: Which?
Ernie: Somebod y named Bennett McC allum
was s aying that rational expectations
proved that the government could not s ta-
bilize the economy. Hang on, I' ve got it
here: "A n accurate und ers tand ing of how
expectations are formed lead s to the con-
clus ion that s hort-run s tabilization policies
are untenable." (McC allum, 1980, p. 37).
I d on' t know how they could d evelop theo-
ries like that. It' s pretty obvious that gov-
ernment policy d oes affect the economy
in the s hort run.
*
Our thanks to Neville C ain for his ins piration
for this paper and to our colleagues at A NU, notably
Malcolm G ray, A d rian Pagan and Jim Trevithick,
for their comments . We are als o grateful to Fred
G ruen and to an anonymous referee and the ed itor
for their as s is tance. A ll faults , of cours e, remain ours .
39
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
40 Journal of Economic Literature, Vol. XX (March 1982)
Bert: He d id n' t s ay they could not affect the
economy in the s hort run or even in the
long run. The key word is s tabilize.' Jus t
think about what' s happened in the las t
few years -record inflation and record un-
employment. You d on' t call that s tabiliza-
tion, d o you?
Ernie: Well, maybe they' ve been s table at
high levels but I take your point. There
d oes s eem to have been s ome break-
d own of the ways in which the govern-
ment can influence the macroeconomy.
Do thes e rational expectations blokes
think they have a mod el to explain s tag-
flation?
Bert: Yes , they d o. It' s caus ed by mis guid ed
governments following Keynes ian policies
that haven' t worked , d on' t work, and
won' t work in the future.2
Ernie: I s uppos e they ad vocate d oing nothing
and letting the ' free market' d o its wors t.
G reat! They s ound jus t like Fried man and
all thos e old -fas hioned monetaris ts . They
have always s aid inflation was jus t a mone-
tary phenomenon and macro policy
could n' t s hift the economy to higher levels
of employment.
Bert: Yes , that' s right. Mos t economis ts now
agree that the long-run Phillips curve is
vertical.3 That means that there exis ts a
natural rate of unemployment.4 G overn-
ment policy can bring about a d eparture
from that only in the s hort run and then
only by fooling people. But you can' t fool
all the people all the time. Therefore, s ys -
tematic policy is ineffective.5
Ernie: I' m not at all s ure that the long-run
Phillips curve is vertical.6 We us ed to have
about one percent unemployment; now
we s eem to be s tuck at about eight per-
cent. How can you explain that with a ver-
tical Phillips curve. "The Phillips curve is
vertical but moves around a lot"-hard ly
s eems much of a theory.7 Even if it is verti-
cal and we can' t get away from it except
by fooling people, clearly the government
can fool people. Every time it changes pol-
icy the people d on' t know about the new
policy for a while s o it takes time before
they catch up.8
Bert: But that' s jus t what rational expecta-
tions is all about! It s ugges ts that people
anticipate the effects of the new policy.
If that' s true, then the policies won' t caus e
any increas e in employment!
Ernie: How on earth are people s uppos ed to
anticipate the effects of policy? I jus t can' t
s ee it. Have they all got econometric mod -
els und er the s ink?9
Bert: (A ngrily) Now you' re jus t being s illy.
Have you read any of the bas ic litera-
ture-Lucas , Sargent, Wallace, and s o on?
Ernie: I' ve looked at s ome but it jus t s eems
unreal-too many equations . They never
d efine exactly how they think anybod y
forms thes e "rational expectations ."
Bert: Look, I' ve got to go to my macro lec-
ture. How about we meet again tomorrow,
I
Us ually d efined as minimization of the variance
around s ome fixed macroeconomic objectives (G reg-
ory C how, 1970).
2 There is a clear id eological component to much
rational-expectations work and opponents will be
tempted to d is mis s the theory on id eological
ground s . Later we s ugges t that there are merits in
the theory quite s eparate from its us e to s upport
particular propos itions about the role of govern-
ment.
3
See M. Fried man (1968) and Robert J. G ord on
(1976) for views on this . A ppend ix A d eals with the
is s ue in s ome more d etail.
4
"Natural" in the s ens e that everybod y who wants
a job at the going wage has one. This d efinition d e-
nies the pos s ibility of unemployment aris ing from
a failure of effective d emand and hence from the
"Keynes ian" problem (Ed mond Malinvaud , 1977).
There is no neces s ary connection between verti-
cal Phillips curves and a natural rate of employ-
ment.
5 This is Fried man' s propos ition that in the long
run anticipated and actual economic values mus t be
equal s o that policies that work through illus ions ,
or s ys tematically wrong anticipations , will be ineffec-
tive in that long run.
6 G ord on (1976) cons id ers a number of pos s ibilities
mainly relying on d ifferent forms of s luggis h price
ad jus tment.
7
Robert Hall (1975) makes this criticis m. A s he
interpreted the evid ence, mos t of the variation in
output came from changes in the natural rate, pro-
voking ques tions about the importance of a theory
which only explained d eviations from the natural
rate. It would be a us eful theory if it explained the
movements in the rate its elf.
8John Taylor (1975) explores the pos s ibilities for
policy while people learn the new rule. Benjamin
Fried man (1979) ad d res s es the s ame ques tion.
9John Muth (1961, p. 317) in outlining what he
meant by rational expectations anticipated this criti-
cis m.
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
Mad d ock and C arter: Rational Expectations 41
and I' ll introd uce you to the magnificient
world of rational expectations . Same time?
Scene ii. A . The Id ea
of Rational Expectations
(In the s ame place, next d ay.)
Bert: Well, are you read y to try to und er-
s tand what rational expectations is about?
Ernie: Yes . Have you got it figured out yet?
Bert: I' ve been thinking about it. Let' s go
through it s ys tematically. Firs t, we can talk
about jus t what rational expectations are.
Then we can look at the way the policy
impotence res ult is d erived . By then we
s hould have a pretty clear id ea of what
this line of res earch is all about s o we can
try to figure out how it relates to the Phil-
lips curve, monetaris m, econometric mod -
els , and all that. O.K.?
Ernie: A lright. What' s the d efinition of ra-
tional expectations ? What on earth might
irrational expectations be?
Bert: Firs t things firs t. Let' s s tart with famil-
iar ground . What would you s ay is the bas ic
behavioral as s umption of economic behav-
ior?
Ernie: Utility maximization, I s uppos e.
Bert: More or les s . I would s ay that the bas ic
as s umption about ind ivid ual behavior is
that economic agents d o the bes t they can
with what they have. This principle forms
the bas is of cons umption theory, prod uc-
tion theory, human capital theory and s o
on.
Ernie: So it' s the bas is of microeconomics .10
But what' s that got to d o with expecta-
tions ?
Bert: Everything. A t its mos t fund amental,
rational expectations theoris ts argue that
the s ame principle s hould be applied to
the formation of expectations . If you want
a d efinition, how about: rational expecta-
tions is the application of the principle
of rational behavior to the acquis ition and
proces s ing of information and to the for-
mation of expectations ."1
Ernie: A m I to infer that my utility function
and I s it d own together and rationally d e-
cid e how much information I s hould ac-
quire in ord er to form the expectations
that will help me maximize my utility? In-
cred ible!
Bert: Yes , you can attack it that way if you
like, but that' s a more general criticis m
of utility theory which we can argue about
s ome other time. A ll I' m s aying here is
that if one cons id ers economic agents to
be rational maximizers , then it' s
cons is tent12 to cons id er information gath-
ering and expectation formation as d eter-
mined by the s ame proced ure.
Ernie: O.K. So you' d ins is t upon a rational ex-
pectations pos tulate that private economic
agents gather and us e information effi-
ciently. That means you believe the mar-
ginal cos ts of gathering and us ing informa-
tion are equated to their marginal benefits .
McC allum d oes n' t agree with you. He s ays :
"Ind ivid ual agents us e all available and
relevant information "13 and it s eems to me
that Sargent and Lucas s ay the s ame. It
almos t s eems as if they think information
is a free good .14
Bert: That' s a good point. Many theoris ts
have ignored the cos ts of information us ed
in the formation of expectations . That is
one of my criticis ms of the literature. But
I think it is us eful to d is tinguis h between
rational expectations as a principle of in-
formational efficiency and rational expec-
tations as it appears in s ome of the macro-
economic literature.' 5
10
There is clearly s ome tens ion in macroeconom-
ics between its empirical behavioral as pects (e.g.,
the cons umption function) and its d erivation of in-
s ights from a microeconomic bas is (e.g., permanent
income hypothes is ). The micro found ations of macro-
economics literature, for example G eoffrey Harcourt
(1977) attempts to res olve this conflict but, s o far,
not very s ucces s fully.
11 This is not the approach us ually ad opted by ra-
tional expectations theoris ts (fn. 15). It is , however,
clos er to the us ual economic method ology and s eems
preferable.
12 That is , cons is tent with the method ological ap-
proach of explaining all behavior in terms of utility
maximization.
13 McC allum (1980, p. 38). In fact, Ernie has quoted
McC allum out of context. He goes on to ad mit that
information cos ts are neglected for s implicity.
14 Ed gar Feige and Douglas Pierce (1976) cons id er
the implications of cos tly information for rational
expectations .
15 The d is tinction s eems important for clarifying
id eas within macroeconomics . The all-information
approach ad opted by Sargent et al. s hould id eally
be given another name, for example "Muth expecta-
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
42 Journal of Economic Literature. Vol. XX (March 1982)
Ernie: The term ' rational' is quite confus ing
in the context and you are right that the
d is tinction between the two things is im-
portant. But what d ifference d oes rational
expectations make to ind ivid uals ? C an you
give examples ?
Bert: The example mos t often us ed in the lit-
erature involves the allocation of time be-
tween labor and leis ure.16 In d ecid ing how
many hours to work this period , an ind ivid -
ual mus t take account of expected future
wages and not jus t the pres ent wage. For
example, if you expect the real wage to
be $10 per hour this week, and $1 next
week, then it makes s ens e to work as much
as pos s ible this week, and have s ome time
off next week. Therefore the number of
hours worked in any period , that is , the
labor s upply, will d epend not only on the
current real wage but on expected future
real wages . A rational expectation of real
wages will take into account all available
information, includ ing the effects of gov-
ernment policy.
Ernie: But my old man works 40 hours every
week-he d oes n' t have much choice.
Bert: But your old man' s bos s d oes . When he
is d ecid ing whether to hire more people
or lay them off, he need s to take into ac-
count future prices and wages . His expec-
tations s hould be bas ed on all the available
information. This includ es , among other
things , the impact of future government
policy.
Ernie: O.K. I s ee how the level of employment
might d epend upon expectations and how
' good ' expectations are better than ' bad '
ones . But I can' t s ee why that means that
there is no room for government policy.
B. Deriving the Impotence Res ult
Bert: Without realizing it you' ve jus t mad e
a very important d is tinction. The relation-
s hip between the level of employment and
expectations is logically quite s eparate
from beliefs about how expectations are
formed . The conclus ion that there is no
s cope for government policy-the impo-
tence res ult-d epend s crucially upon im-
pos ing a s pecial as s umption about expec-
tations -rational expectations -upon a
s pecial type of macroeconomic mod el.
Ernie: Well I think I und ers tand the meaning
of rational expectations . What types of
macro mod els d o rational expectations
theoris ts us e?
Prices
Supply
\
\
Demand ,
Demand o
Yn Income = Output
Diagram 1
Bert: (Drawing a d iagram.) Mos t of them
work with the id ea that the levels of out-
put and prices are d etermined by the in-
ters ection of an aggregate d emand and ag-
gregate s upply function. The aggregate
s upply curve is taken to be vertical, s o that
output cannot d eviate from Yn as a d irect
res ult of any change in the level of d e-
mand . Thus government policies d es igned
to change the level of aggregate d emand
are not likely to be effective. The level
Yn is the output as s ociated with equilib-
rium in the labor market at the natural
rate of unemployment s o we can call Yn
the natural rate of output or income for
the economy.17
tions ," s ince the ad jective "rational" is normally re-
s erved in economics to d es cribe the outcome of a
utility maximization proces s . J. J. Sijben writes :
"Muth' s view implies that economic agents build up
their expectations as if they are fully informed of
the proces s which ultimately generates the real out-
come of the variable concerned " (1980, p. 66).
Pus hed further, McC allum follows the line that all
mod els are "unrealis tic," which s eems to lead him
to the pos ition that theories s tand or fall on their
pred ictions .
16 Rational expectations in labor s upply d ecis ions
have fairly obvious corollaries on capital inves tment
d ecis ions (Robert Lucas , 1975, for example).
17
A ppend ix A d eals with the problems of the verti-
cal aggregate s upply curve in more d etail. It s hould
be noted that the mod els are us ually expres s ed in
logarithms s o that the real d ebate concerns rates
of change rather than levels . The d is tinction is ne-
glected here.
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
Mad d ock and C arter: Rational Expectations 43
C ons id er the pos s ibility that the govern-
ment takes action that may, at firs t blus h,
be s uppos ed to increas e output. For exam-
ple, let it act to increas e nominal income
and aggregate money d emand . Money
wage rates will tend to ris e, and if workers
regard this as equivalent to an increas e
in real wages , employment will increas e
and output will temporarily ris e to a level
higher than
Y..
But if prod uction is carried
on s ubject to d iminis hing returns to labor,
prices will ris e relative to nominal wages ,
and real wages will fall. When workers re-
alize this , employment will fall back to its
original pos ition, and output will return
to Yn. A t this point, nominal wage rates
and prices are higher (the nominal d e-
mand curve cros s es the vertical s upply
curve at a higher level), but output and
employment are back where they s tarted .
Since the aggregate s upply curve had not
s hifted , the pos s ibility of increas ing em-
ployment and output aris es only as long
as people confus e nominal changes in
wages (for example) with real changes .
This means that government policy will
only increas e the level of income in
real terms if it is able to fool people into
confus ing nominal changes with real
ones .18
Ernie: That' s rid iculous ! The ' natural' rate of
unemployment d epend s intimately upon
all s orts of government policies -for exam-
ple tax laws , minimum wage laws , immi-
gration policy, s chool-leaving age, etc., etc.
Do you really mean that the government
can' t change aggregate s upply by increas -
ing the inves tment allowance? Or by go-
ing to war, for that matter?
Bert: You' re right, you' re right! I s hould have
been more careful. C learly government
policy can alter the natural rate of unem-
ployment or, if you like, the pos ition of
the aggregate curve. What I s hould have
s aid is that the only way in which govern-
ment policy can bring about d eviations
from the natural rate of unemployment
is by ind ucing private agents to have mis -
taken expectations . Let' s write d own a
s imple mod el.
(Bert' s s cribbling is attached as A ppend ix B.
For thos e who like mathematical d es criptions
it s hould make the d is cus s ion clearer but is
not a neces s ary ad junct.)
Ernie: That makes your pos ition clearer. The
actual aggregate s upply function implies
that d eviations of actual output from the
natural rate are d irectly proportional to
d eviations of actual prices from expected
prices . Since people with rational expecta-
tions never make mis takes about policy
rules , policy will never fool them, and out-
put will never d eviate from its natural rate
as a res ult of any policy rule.
Bert: That' s the id ea but you' ve put it too
s trongly. If government policies are ran-
d om, they will be effective although not
neces s arily d es irable. It' s the s ys tematic
component of policy that the theory s ug-
ges ts will be ineffective.
Ernie: I' m not too s ure about the neutrality-
of-money propos ition19 generally but will
let it rid e for now. You explain the res t
of the argument-then I' ll put my objec-
tions one by one.
Bert: Now the point of rational expectations
is that people won' t be s urpris ed by any
s ys tematic policy. A ny government that
relies upon a policy rule-one that has a
fixed growth of the money s upply, or one
s ys tematically related to income or unem-
ployment-will never caus e any d eviation
from the natural rate.20 A rand om policy
will affect real output. But any policy rule
that is s ys tematically related to economic
cond itions , for example one d es igned with
s tabilization in mind , will be perfectly an-
ticipated , and therefore have no effect on
output or employment. In other word s , to
have real effects . monetary Dolicv mus t be
"8A s Thomas Sargent and Neil Wallace s ugges t
"it mus t s omehow trick the public" (1976 p. 177).
The argument is more complex with capital in the
mod el as may be clear from A ppend ix A .
19
The id ea that changes in money s upply d o not
influence people' s preferred hours of work, portfolio
hold ings , etc. A gain, this is cons id ered in A ppend ix
A .
20
Sargent and Wallace (1976, pp. 177-78), put this
argument in almos t the s ame form. Expectations can
be wrong but not s ys tematically wrong (i.e., bias ed ),
hence there is no s cope for s ys tematic policy.
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
44 Journal of Economic Literature, Vol. XX (March 1982)
completely unpred ictable. A ny s ys tematic
policy will be impotent.21
Ernie: C an we put it this way? Rational expec-
tations are bas ed on all available informa-
tion. The available information s et in-
clud es the government policy rule.
Therefore, a rational expectation of in-
flation, for example, will includ e the an-
ticipated effects of government policy s o
that the policy will have no effect on out-
put.22
Bert: Yes , that pretty well s ums it up.
C . C riticis ms
Ernie: Now that I think I und ers tand what
you' re on about, can I tell you what I think
is wrong with the mod el?
Bert: O.K.
Ernie: Firs t, I d on' t like your s upply curve.
There are lots of criticis ms one could
make, but the mos t important in the con-
text of the mod el is that you as s ume an
extreme form of the neutrality of money.23
Perfectly anticipated inflation has no real
effects in your mod el. That' s clearly
wrong. Buiter (1980) put the s tand ard ar-
gument in terms of the portfolio read jus t-
ments required becaus e inflation changes
the real rate of return on thos e financial
as s ets which have a zero nominal return.
Pers onally, I think the d is tortions intro-
d uced by the progres s ive tax s tructure in
an inflationary s ituation are far more im-
portant empirically.
Bert: Yes , but all mod els are approximations .
Ernie: True, but not all approximations are
good approximations ! Here' s another
problem. If expectations are rational, then
expectation errors s hould be rand omly d is -
tributed over time. A s traightforward im-
plication of that for this mod el is that the
level of output (or unemployment) is un-
correlated over time. Yet everybod y
knows that the G NP and unemployment
s eries have a high d egree of s erial cor-
relation.24 We tend to go through a s eries
of years in which unemployment is below
the ' natural rate,' and then a s eries of years
in which it is above the ' natural rate.' It
d oes n' t s eem to be d is tributed very ran-
d omly. C ompare the s ixties and s eventies
in A us tralia-it' s the old s tory of bus ines s
cycle expans ion and contraction.
Bert: I can' t d eny the s erial correlation in the
unemployment or income s eries . Mos t ra-
tional expectations mod els includ e lagged
income or lagged unemployment as expla-
natory variables in the s upply function.25
This d oes make the mod els fit the d ata
better, but there is no good theoretical
jus tification for it. Lucas is the only one I
know who really ad d res s es the is s ue.26 He
relies on the well-known ' fact' that all peo-
ple live on is land s . A t the end of each trad -
ing period people choos e a new is land at
rand om. Since they d on' t know the his tory
of their new is land , they can' t d is tinguis h
immed iately between real and nominal
effects .27
Ernie: Thes e is land mod els s eem appropriate
to a s ociety in which the fas tes t form of
communication is a floating coconut.
Has n' t Lucas ever heard of rad io and the
telephone?28
Bert: I have to agree with you. I s aid that
the explanations for pers is tence weren' t
21
G ord on (1976) makes this clear. See es pecially
pp. 200-01.
22This follows the us ual s olution method followed
by rational expectations mod els . See Lucas (1973)
for an example.
23
Some criticis ms are d is cus s ed in A ppend ix A .
24
This was Hall' s criticis m (1975), and is als o put
by G ord on to Sargent (1973, p. 478).
25
Lucas (1973) introd uced the lagged term with
a footnote explaining that not all d eviation from the
natural rate of unemployment could be accounted
for by the error in expectations terms .
26 Lucas (1975) attempts a s ys tematic explanation
for the s erial correlation in terms of information lags .
See Rod ney Mad d ock (1979) for a d is cus s ion of the
importance of pers is tence for the rational expecta-
tion program.
27 "The id ea behind this is land abs traction is not,
of cours e, to gain ins ight into maritime affairs , or
to comment on the aimles s nes s of life. It is intend ed
s imply to capture in a tractable way the fact that
economic activity offers agents a s ucces s ion of am-
biguous , unanticipated opportunities which cannot
be expected to s tay fixed while more information
is collected . It s eems s afe and , for my purpos es , s ens i-
ble to abs tract here from the fact that in reality this
s ituation can be s lightly mitigated by the purchas e
of ad d itional information" Lucas (1975, p. 1120).
28 Lucas d oes actually mention the problem in the
quotation in the previous footnote, but makes noth-
ing of it.
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
Mad d ock and C arter: Rational Expectations 45
very convincing, es pecially when the gov-
ernment regularly publis hes lots of s tatis -
tics . The news papers carry s tock exchange
prices every d ay. The information s eems
to be es s entially free.
There is another line of argument
though. If prices change s ud d enly, firms
can increas e their prod uction les s than
their s ales by us ing up s ome of their inven-
tories . If there is no price s hock in the
next period prod uction would then be
rais ed to build inventories back to their
original level. Thus there would be an in-
creas e in prod uction to meet the original
s tock and for as long thereafter as the re-
s tocking took.
Ernie: But that implies there s hould be a
s trong relations hip between inventory cy-
cles and output cycles and that' s not really
true, is it?
Bert: Well, the relation is far from perfect.
I was really jus t s ugges ting that in an econ-
omy characterized by d urable good s it
s hould n' t be too d ifficult to accept that ad -
jus tments of various s orts will have effects
that pers is t.29 We really d on' t have a good
explanation for pers is tence (s erial correl4-
tion). I willingly conced e that point.
What' s next?
Ernie: O.K. Even if all that information is
freely available, you as s ume that all the
agents know the correct mod el of the
economy. How . . .
Bert: No, I d on' t. Well, not me really. I mean
that rational-expectations people d on' t
neces s arily s ay that everybod y knows the
correct mod el of the economy. They s ug-
ges t that s ome arbitrage proces s takes
place whereby the people who have the
correct mod el d ominate the outcome.30 If
there are mis apprehens ions , then well-
informed agents can make profits at the
expens e of the ill-informed . This will inevi-
tably lead the s ys tem to converge to the
rational expectations equilibrium. A s your
old mate, John Maynard Keynes , s aid :
* . .actions bas ed on inaccurate anticipations
will not long s urvive experiences of a con-
trary character, s o that the facts will s oon
overrid e anticipation except where they
agree. [1930, p. 160].
Ernie: G ranted there is a role for arbitrage.
But how d o we know that expectations
of the experts will converge on the true
value? G ive me any rational expectations
mod el and I think I can s how you a reas on-
able ad jus tment proces s that will not con-
verge to the rational expectations equilib-
rium.
Bert: A nd I can probably s how you one that
can. Unles s the theoris ts s pecify an ad jus t-
ment mechanis m, we can' t really argue
about this point. Rational expectations the-
oris ts haven' t ad d res s ed this problem.31
Ernie: That' s a big gap in your theory. But let
me read to you what Robert Shiller s ays :
Even if a mod el d oes eventually converge
on a rational expectations equilibrium, it may
take s uch a long time to d o s o that, s ince
the s tructure of the economy changes occa-
s ionally, the economy is never clos e to a ra-
tional expectations equilibrium. [1978, p. 39].
To recalculate a quarterly econometric
mod el after a change in policy rule might
take 20 quarters . To es timate the effect
of policy bas ed on the new es timates might
take another 20 quarters . Thus , even if the
proces s converges , each s tage in the con-
vergence to the new equilibrium could
take five years -by which time we may
all be d ead !
Bert: But if the G overnment' s objective is to
s tabilize the economy, then it wants to
29Thes e is s ues are rais ed in a penetrating d is cus -
s ion of the problem of pers is tence by G ord on (1981).
30For example, Muth (1961) argued that econo-
mis ts could s ell the information profitably if expecta-
tions were not rational. Since he wrote, many have
d one s o. This s ugges ts that market forces would tend
to d rive d ecis ions to thos e rationally bas ed .
31
Shiller (1978, p. 38) focus es upon the is s ue of
convergence. There s eem to be two s eparate is s ues
involved . Since rational expectations for this period
d epend upon es timates about the future while the
future d epend s in part upon pres ent expectations ,
there need be no unique rational expectation for
the current period . In many mod els , method s of ad -
jus ting expectations (i.e., forecas ts ) of the future will
either converge on a rational-expectations s olution
or explod e. The implicit argument of protagonis ts
s eems to be that s ince we d o not obs erve prices ex-
plod ing off to infinity we need only cons id er con-
verging cas es . This type of counter-factual reas oning
is s omewhat d ubious . The d ynamics of expectation
formation might s till be explos ive but s ome other
fact or-e.g., policy action-act to cons train the ex-
plos ive tend ency.
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
46 Journal of Economic Literature, Vol. XX (March 1982)
make s ure that private expectations are
rational, and hence it will inform the pub-
lic of any new policy rule.32
Ernie: That d oes n' t get you out of hot water.
Firs t, the learning problem d oes n' t con-
cern the policy rule alone. A gents als o
have to learn the s tructure of economy,
which is s ubject to change. Econometric
mod ellers d on' t have an outs tand ing rec-
ord of s ucces s , d o they? Second , why d o
you as s ume that the goverment' s objective
is to s tabilize the economy? It s eems to
me that the government' s real objective
is to remain in power-you know, the po-
litical bus ines s cycle id ea.33 A nd if that is
their objective it may be in their interes t
to hid e information and fool the public.
If that is the cas e the voters can hard ly
be expected to believe the s ignals the gov-
ernment is s end ing out and the whole
macroeconomic proces s d egenerates into
a gues s ing game.
Bert: But you mus t agree that mos t macro-
economics d oes as s ume the objective of
s tabilization. This literature falls into that
trad ition which is concerned with govern-
ment policy rules d es igned to achieve
macroeconomic s tabilization.34
Ernie: Let' s go and get a beer.35
Scene iii. (In the union bar)
A . Tes ting
Bert: Now the tes ting is a bit tricky. It' s a
pretty young res earch program and there
are no well-accepted tes ting proced ures
as yet. The principal d ifficulty is that we
are really tes ting a joint hypothes is -the
economic mod el and the expectations
mechanis m. That makes it d ifficult to d e-
cid e jus t where the res pons ibility for fail-
ures of tes ts really lie.
Ernie: But why can' t you jus t tes t the expecta-
tions mechanis m d irectly? A s k people
what they expect, and s ee if they are right?
Bert: Stephen Turnovs ky (1970) and James
Pes and o (1975) and a couple of other peo-
ple have d one that, but the res ults have
been inconclus ive. Economis ts trad ition-
ally d on' t like s urveys , anyway.
Ernie: Well, how have rational expectations
protagonis ts tried to tes t the theory?
Bert: Bas ically they have taken two d ifferent
approaches . Have a look at the s upply
function again. The natural rate hypothe-
s is will only allow non-rand om d eviations
if there are expectations errors of s ome
s ort. Und er the rational expectations hy-
pothes is there are no expectations errors -
or at mos t, only rand om ones . Thus , d evia-
tions from the natural rate of output mus t
be rand om.36 In particular, d eviations can-
not be s ys tematically related to any other
explanatory variable, for example the
(lagged ) money s upply or the wage rate.
A nd s o the firs t type of tes t is es s entially
to s ee whether d eviations from the natural
rate are s ys tematically related to any other
variables . This tes t has been applied in a
number of d ifferent ways es pecially by
Sargent (1973; 1976). In thos e papers the
joint hypothes is -natural rate and rational
expectations -was rejected in a number
of cas es , jointly rejected becaus e they
were jointly tes ted . With s lightly d ifferent
s pecifications , they weren' t. Sargent con-
clud es that rational expectations is ' not ob-
s cenely at variance with the d ata.' 37
Ernie: Remarkable res ilience, eh!
Bert: Yes , what' s more, his next paper was
entitled ' The Obs ervational Equivalence
of Natural and Unnatural Rate Theories
of Macroeconomics .' 38 This initiated the
s econd approach to tes ting rational expec-
tations . It was bas ed on the id ea that what
rational expectations mod els ad d , com-
pared with other mod els , is that price ex-
32Sargent and Wallace (1976, pp. 181-83) argue
this point.
33The nature of the problem when the govern-
ment' s objectives vary over time d oes not s eem to
have been well explored . C learly, rational expecta-
tions forces economis ts to think more about the pre-
cis e nature of learning.
34
Following the trad ition of Dutch economis ts Jan
Tinbergen (1952) and Henri Theil (1958).
35 Following a s ound A us tralian trad ition.
36A ctually, s ome allowance in the tes ts is mad e
for pers is tence by the inclus ion of lagged values of
the d epend ent variable.
37The theory s ugges ts that no extra information
would s ignificantly contribute to the pred iction. The
evid ence would thus appear to fals ify the theory.
Sargent (1976), ins tead , went on to try an alternative
type of tes t. See es pecially p. 233.
38Sargent
(1976) and Sargent and Wallace (1975)
s tart to d evelop this id ea.
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
Mad d ock and C arter: Rational Expectations 47
pectations take into account the policy
rule the government is us ing. If that rule
d oes not change, ord inary mod els and ra-
tional expectations mod els will fit any d ata
s et equally well, though probably with d if-
ferent parameters . What it means is that
you can only d is tinguis h between rational
expectations macro mod els and ord inary
ones if the policy rule has changed .39 It
s eems a reas onable id ea to me.
Ernie: Maybe.
Bert: Well, anyway, it gave Sargent and Salih
Neftci (1978) an id ea for a new type of
tes t. Firs t, they es timated the govern-
ment' s policy rule by regres s ing the
money s upply on pas t levels of income.
They then analys ed the res ults to s ee if
there had been any s ignificant changes in
the relations hip, that is , to s ee if the policy
rule had changed .40 They found changes
in 1929 and 1964. They then looked at
s ome ord inary macroeconomic mod els to
s ee whether the parameters had changed
at about the s ame time that the policy rule
changed . In each cas e they found s ome
evid ence that it had .4'
Ernie: Well, that s eems like a reas onable s ort
of approach. Really it d epend s on the ra-
tional expectations id ea-that people
change their behavior as policy changes -
rather than on the natural rate propos i-
tion. I' m inclined to agree that people take
into account what the government is try-
ing to d o when they plan for the future
but the extreme form of the natural rate
of unemployment is an over-s implification
of a complex world . Of cours e, the tes t
is a bit tricky. It' s not really clear to me
that that' s the way to es timate the govern-
ment policy rule and then, even if it is ,
there are probably other non-rational ex-
pectations mechanis ms which would s ug-
ges t a change in the econometric s tructure
as a res ult of a change in policy rule.
There' s no real alternative hypothes is in-
volved in the tes t!
Bert: Everyone agrees it' s not a very s trong
tes t; it' s titled ' A Little Bit of Evid ence
. . .
'
but it is s ugges tive and d oes focus
upon the relation between expectations
and the policy rule and not upon the natu-
ral rate. It gets away from the d ogmatic
form of the natural rate, impotence of pol-
icy area, and focus es upon the pos itive
contribution of the rational expectations
id ea.
B. Significance
Ernie: O.K., we' ve covered the mod el and the
evid ence, s uch as it is . What' s all the fus s
about?
Bert: Well, it really nails the Phillips curve.
Much pos t-war s tabilization policy has
been bas ed on the id ea that there is a
trad e-off between unemployment and in-
flation that the government can exploit by
influencing aggregate d emand -the s o-
called Phillips curve. Fried man largely un-
d ermined that with the natural rate id ea.
He s aid that policy only worked by fooling
people and that in the long-run they could
not be fooled . This s till left the way open
for effective s hort-run policy. If people
have rational expectations they won' t be
fooled . If people have rational expecta-
tions they won' t be fooled by s ys tematic
policy even in the s hort run s o there is
no s cope for s hort-run policy either. This
explains why the Phillips curve became
uns table the moment policy makers tried
to exploit it.
Ernie: But the explanation d epend s really
heavily, as you' ve alread y agreed , upon
the particular macroeconomic mod el you
have s et up.42 The monetaris ts have
39 Since expectations are us ually unobs ervable
they are eliminated from econometric mod el to be
es timated by introd ucing an equation about their
relation to known variables . The parameters of this
equation then become embed d ed in the actual re-
d uced form equations which are es timated . A n
"equation" for rational expectations incorporates the
parameters of the policy rule being us ed by the gov-
ernment. Und er the as s umption of rational expecta-
tions thes e parameters of the policy rule become
embed d ed in the red uced form (i.e., es timating)
equations of the economy. Thus they s ugges t that
s tructure of the economy, as meas ured by us ual
econometric mod els , will appear to change when-
ever the policy rule changes .
40
Us ing Ml there was a policy change, but with
M2 none appeared to have taken place.
41
The tes ts were non-parametric.
42
A lan Pres ton and A d rian Pagan (Forthcoming,
C h. 10) explore the way in which the impotence
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
48 Journal of Economic Literature, Vol. XX (March 1982)
grabbed this id ea43 to s upport their trad i-
tional pos ition that active government pol-
icy is not d es irable.44 It' s largely been their
baby until now. The reas on is not hard
to s ee. The impotence res ult s trongly s up-
ports their id eological pos ition that the
government s hould keep its hand s off the
economy.45
Bert: That' s rather hars h. Monetaris ts are no
more id eological than other economis ts .
What would be your ' un-id eological' view
of rational expectations ?
Ernie: I' ve been wond ering how a future his -
torian of thought might as s es s it. I think
rational expectations theory will be s een
as a very important d evelopment in eco-
nomics , but not becaus e of the impotence
res ult. Rational expectations are important
in any s ituation in which market behavior
is influenced by expectations . Take the
cas e of an aggregate d emand d eficiency
in a Keynes ian mod el. The us ual argument
is that monetary expans ion will work
through affecting interes t rates s o that
grad ually the economy s hifts to a higher
output level. With rational expectations
the s hift to the new level would be ex-
tremely rapid . If bus ines s men und ers tand
the economic implications of expans ionary
government policy, they can expand their
output in anticipation of thos e effects
rather than waiting for the ris e in d emand
to be obvious in the market. In that cas e,
far from policy being impotent, rational
expectations may make policy more effec-
tive.
This example, by the way, illus trates the
fact that mos t of the rational expectations
literature has a particular economic mod el
built in, one in which all markets clear
ins tantaneous ly; unemployment is , there-
fore, voluntary, hence ' natural' ; and
money is neces s arily neutral. But if that
mod el is not applicable, policy need not
be impotent, and , as s aid , rational expec-
tations may make it more rapid ly effec-
tive.
What' s more rational expectations can
be applied in microeconomic s ituations .46
C obweb mod els of d ynamic behavior in
commod ity markets d epend upon ' irra-
tionality of expectations ' -the id ea that
next year' s prices will be the s ame as this
year' s or even a s imple extrapolation of
it. C learly, prod ucers can d o better than
that, and if they have rational expectations
the market is likely to approach its equilib-
rium quite quickly.
Bert: You' re right! The id ea can be applied
in a wid e variety of mod els but what about
your general conclus ions about policy for-
mation?
Ernie: Yes , I agree that there are important
implications for policy d es ign. Private eco-
nomic agents are intelligent d ecis ion-mak-
ers and can be expected to take the effects
of government policy changes into ac-
count in d ecid ing their behavior. This
means that the policy-maker mus t antici-
pate the effect of policy on private expec-
tations and the cons equent changes in be-
havior. In practical terms it means that
we need to know a lot more about the
availability and us e of information by pri-
vate d ecis ion-makers .47 Thus , the focus of
the theory of policy s hould be on expecta-
res ult d epend s on particular s pecifications of the
macroeconomic mod el being cons id ered . They d e-
velop a more general mod el that includ es the us ual
rational expectations mod els as particular cas es .
43
Lucas and Leonard A . Rapping us ed ad aptive
expectations (1969), then rational d is tributed lag ex-
pectations (1969) before Lucas firs t introd uced ra-
tional expectations (1972). Where ad aptive expecta-
tions as s umed that people jus t s imply ad apted to pas t
errors , rational d is tributed lag expectations were the
very bes t econometrically pred icted es timates of
prices d erived from analys is of all pas t price informa-
tion.
44
A ny s table und ers tand able rule would have no
effect if people were exactly able to pred ict it. By
that tes t a "d is cretionary" rule and a fixed money
growth rate rule would be equally impotent. If s to-
chas tic effects of d is cretionary rules are allowed for
however, thes e cannot be pred icted and would intro-
d uce fluctations into the s ys tem. See Sargent and
Wallace (1976) for a d is cus s ion of the is s ues .
45
McC allum' s (1980) popularization of their pos i-
tion carries the inference that the res ults are quite
robus t, but that d oes not appear to be the cas e. See
Pres ton and Pagan (Forthcoming).
46The s eminal article by Muth (1961) d eals with
microeconomic market s ituations .
47Sargent and Wallace (1975, p. 251) provid e a
s tart in this d irection by mod elling a cas e where
government has an information ad vantage over pri-
vate actors .
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
Mad d ock and C arter: Rational Expectations 49
tions and information and on their role
in d etermining behavior.
Bert: I think you can go further. People us ed
to think that the only reas on s tabilization
policy d id n' t work well was that policy-
makers d id n' t have enough knowled ge
about the s tructure of the economy. Ra-
tional expectations has taught us that the
problem may not be jus t one of the abs o-
lute knowled ge of the authorities but
rather of how much more or les s they
know than the public d oes -a problem of
relative knowled ge. If this is true, the
problem will always be with us .
C . C onclus ion
Ernie: Well, I s tarted off inclined toward s G or-
d on' s view that rational expectations is an
example of a recent d evelopment in eco-
nomics "in which theory proceed s with
impeccable logic from unrealis tic as s ump-
tions to conclus ions that contrad ict the his -
torical record " (1976, p. 5). But now I s ee
that that' s a bit too hars h.
Mos t of the res earch on rational expecta-
tions has exhibited great technical compe-
tence, ' impeccable logic,' and cons id era-
ble ingenuity. This has contributed in no
s mall meas ure to its apparent s ucces s , and
to the confus ion and uncertainty which ra-
tional expectations have arous ed in the
res t of the economics profes s ion. The fun-
d amental s implicity of the id eas involved
has become obs cured by overly rigorous
d evelopment, and es pecially by the un-
convincing res ort to extraneous cons truc-
tions , s uch as the ' is land s ' mentioned
above.
Und oubted ly, it is the impotence of pol-
icy res ults that has arous ed mos t attention.
Yet thes e res ults d epend very heavily on
a particular type of macroeconomic mod el
us ually embod ying a s trong form of the
natural rate hypothes is . If you s tart with
' clas s ical' mod els in which policy can have
no real effects , it is hard ly s urpris ing that
you get res ults in which policy is impotent.
Becaus e of this the novelty of rational ex-
pectations has become bund led up with
tired and worn notions of the way in which
the world works . It is vitally important to
unbund le thes e id eas .
The rational expectations hypothes is , in
its elf, s hould not be provocative to econo-
mis ts . It merely brings expectations within
the s cope of ind ivid ual maximizing behav-
ior. Expectations us ed to be hand led
within economic mod els on an ad hoc ba-
s is . Rational expectations provid es a way
of incorporating expectations which is con-
s is tent with the orthod ox economic theori-
zing.
The d evelopment of rational expecta-
tions theory will make a more s ignificant
contribution to economics in the impetus
it gives to res earch on the vital areas of
learning and expectations formation. It
brings to the fore ques tions about the
availability and us e of information. Ins tead
of being the finale of the monetaris t' s cas e
agains t policy intervention, it s hould be
s een as the prologue for a revitalized the-
ory of expectations , information and pol-
icy.
Bert: I gues s you' re right. Let' s go and get
another beer.
A PPENDIX A : A ggregate Supply
The und erlying ins piration for rational ex-
pectations macro mod els is d erived from the
notion of general equilibrium. With price flexi-
bility, for given end owments and s kills , a cond i-
tion of general equilibrium requires equilib-
rium in the labor markets . In s uch a world all
unemployment is voluntary, everybod y who
wants a job has one. Every ind ivid ual has his
labor hours and as s ets allocated accord ing to
s ome pers onal optimum. The remaining unem-
ployment can be termed the "natural" rate
of unemployment and the level of output
termed the "natural" level of output.
A bs tracting from inter-ind us try s hifts in pro-
d uction, the only way output can change is
through a change in employment. To increas e
or d ecreas e the level of output government
policy mus t alter the equilibrium in the labor
markets . But if the natural rate of unemploy-
ment repres ents an optimal pos ition for pri-
vate actors , how can government policy affect
it?
The mod els rational expectations theoris ts
us ually work with s ugges t that this is pos s ible
only if the government is able to fool people.
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
50 Journal of Economic Literature, Vol. XX (March 1982)
If people confus e nominal wage changes for
real ones , they might reallocate their portfolios
and their hours of work, and thus increas e out-
put. While allowing for this pos s ibility the mod -
els s ugges t that s uch a change would not be
d es irable for the worker (repres enting a s ub-
optimal d ecis ion) and would be avoid ed if they
had rational expectations . They s ugges t that
the labor s upply d ecis ion is mad e in real terms
s o that labor market equilibrium is ind epen-
d ent of prices which, in turn, is taken to imply
that output is ind epend ent of prices . This re-
s ult is pres ented in a vertical aggregate s upply
curve.
A lternative macro-economic theories s ug-
ges t that the optimal allocation d ecis ions of pri-
vate actors will be affected by changes in
prices , but not jus t becaus e people are fooled .
If this were true, increas es in aggregate d e-
mand could increas e output and employment
even with rational expectations . One argu-
ment for the propos ition s ugges ts that people
d on' t hold money in their as s et portfolios s im-
ply for trans action purpos es . If prices go up,
the d es irability of hold ing s uch money goes
d own, changing people' s private allocation d e-
cis ions , and perhaps the rate of capital forma-
tion or number of hours worked . Thus it might
be s aid that the rational expectations mod els
as s ume that the only motive for hold ing money
is the trans actions motive.
A PPENDIX B: A lgebra of the Mod el
Supply: yt- -=
a(pt-pt*) + ut [1]
Demand : Yt =
-bpt + eXt [2]
Expectations : Pt*
=
E[ptlIt-1]
[3]
where Yt= income
y= income level corres pond ing
to the natural rate of unem-
ployment
Pt= prices
Pt*= price expectations
xt= government policy ins tru-
ment, e.g., money s upply
It-, = all information available at
time t- 1
ut
=
rand om error term; Eut
= 0
E =
expectations operator.
Equating d emand and s upply, we obtain the
following red uced form equation:
Pt = a + b (apt + c [4]
Now, by the rational expectations as s umption
(eqn. 3),
pt*
=
E[ptjIt-1]
1
us ing (4)
=
E[ (apt*+ext- y--u)]
=
a + b (aEp* +
cExt -
E-- Eud )
But Ept*
=
pt*, Ey= y EutO=
s o
Pt+b(aPt*
+
cExtY-) [5]
Subtracting [5] from [4]
1
Pt -Pt* =
a+b
[c(xt -Ext) -Ut] [6]
Subs tituting [6] in equation [1]
a
Yt Y =
a + b
[c(xt-Ext)-ut] + ut
a+ b
=acb (Xt -Ext) + b [7]
a +b a+ b Ut
That is , the d eviation of output from the ' natu-
ral' level y-d epend s only on the uns ys tematic
component of government policy (x - Ext). To
s ee this , as s ume that the government us es the
following policy rule:
xt= kxt-1 +
lyt-,
+ mpt-i -
nPt-2 + vt [8]
where vt is a rand om variable, Evt= 0.
Then,
Ext
=
kxt_- +
lyt-,
+ mpt-i -
nPt-2 [9]
Subtracting [9] from [8]
Xt- Ext
= vt
Putting this in [7]
- ac
+
b
Yt Y
a+ b a+ b`
1
=
a+
b (acvt
+
but)
Deviations of Yt from
-
are thus entirely ran-
d om. This implies that s ys tematic government
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions
Mad d ock and C arter: Rational Expectations 51
policy is impotent (in this mod el), s ince the
s ys tematic component of any policy will be in-
corporated in Ext, and therefore be cancelled
out in forming xt - Ext.
REFERENC ES
BUITER, WILLEM. "The Macroeconomics of Dr. Pan-
glos s : A C ritical Survey of the New C las s ical Mac-
roeconomics ," Econ. J, March 1980,
90(1), pp. 34-
50.
C How, G REG ORY. "Optimal Stochas tic C ontrol of
Linear Economic Sys tems ," J Money, C red it,
Banking, A ug. 1970, 2(3), pp. 291-302.
DEC A NIO, STEPHEN. "Rational Expectations and
Learning from Experience," Quart. J. Econ., 1979,
93(1), pp. 47-57.
FEIG E, EDG A R A ND PEA RC E, DOUG LA S. "Eco-
nomically Rational Price Expectations ," I. Polit.
Econ., 1976, 84(3), pp. 499-522.
FRIEDMA N, BENJA MIN. "Optimal Expectations and
the Extreme Information A s s umptions of ' Rational
Expectations ' Macromod els ," J Monet. Econ., Jan.
1979, 5(1), pp. 23-41.
FRIEDMA N, MILTON. "The Role of Monetary Policy,"
A mer. Econ. Rev., March 1968, 58(1), pp. 1-17.
G ORDON, ROBERT A . "Rigor and Relevance in a
C hanging Ins titutional Setting," A mer. Econ. Rev.,
1976, 66(1), pp. 1-14.
G ORDON, ROBERT J. "Recent Developments in the
Theory of Inflation and Unemployment," J Monet.
Econ., A pril 1976, 2(2), pp. 185-220.
. "Output Fluctuations and G rad ual Price A d -
jus tment," J Econ. Lit., June 1981, pp. 493-530.
HA LL, ROBERT E. "The Rigid ity of Wages and the
Pers is tence of Unemployment," Brookings Pap.
Econ. A ct., 1975, 2, pp. 301-49.
HA RC OURT, G EOFFREY, ed . The microeconomic
found ations of macroeconomics . Lond on: Macmil-
lan, 1977.
KEYNES, JOHN. A treatis e on money. Vols I and II.
New York: Harcourt Brace, 1930.
LUC A S, ROBERT E., JR. "Tes ting the Natural Rate
Hypothes is ," in The econometrics of price d etermi-
nation, C onference. Ed ited by OTTo EC KSTEIN.
Was hington, D.C .: Fed eral Res erve Board , 1972,
pp. 50-59.
. "Some International Evid ence on Output-
Inflation Trad eoffs ," A mer. Econ. Rev., June 1973,
63(3), pp. 326-34.
. "A n Equilibrium Mod el of the Bus ines s C y-
cle," J Polit. Econ., Dec. 1975, 83(6), pp. 1113-
44.
. "Econometric Policy Evaluation: A C ri-
tique," J Monet. Econ., Supplementary Series
1976, 1, pp. 19-46.
LUC A S, ROBERT E., JR. A ND RA PPING , LEONA RD A .
"Price Expectations and the Phillips C urve,"
A mer. Econ. Rev., June 1969a, 59(3), pp. 342-50.
. "Real Wages , Employment and Inflation,"
J Polit. Econ., Sept./Oct. 1969b, 77(5), pp. 721-
54.
MA DDOC K, RODNEY. Rational expectations , political
bus ines s cycles and the cours e of macroeconomic
theory. Unpublis hed Ph.D. Thes is . Duke Univer-
s ity, 1979.
MA LINVA UD, EDMOND. The theory of unemploy-
ment recons id ered . New York: Wiley. Oxford :
Blackwell, 1977.
MC C A LLUM, BENNETT T. "The Significance of Ra-
tional Expectations Theory," C hallenge. Jan./Feb.
1980, pp. 37-43.
MUTH, JOHN. "Rational Expectations and the Theory
of Price Movements ," Econometrica, July 1961,
29(3), pp. 315-35.
NEFTC I, SA LIH A ND SA RG ENT, THOMA S. "A Little
Bit of Evid ence on the Natural Rate Hypothes is
from the U.S.," J Monet. Econ., 1978, 4(1), pp.
315-19.
PESA NDO, JA MES. "A Note on the Rationality of the
Livings ton Price Expectations ," J. Polit. Econ.,
A ug. 1975, 83(4), pp. 849-58.
PRESTON, A LA N A ND PA G A N, A DRIA N. The theory
of economic policy: Statics and d ynamics . New
York: C ambrid ge Univers ity Pres s , forthcoming.
SA RG ENT, THOMA S. "Rational Expectations , the Real
Rate of Interes t, and the Natural Rate of Unem-
ployment," Brookings Pap. Econ. A ct., 1973, 2, pp.
429-72.
. "A C las s ical Macroeconomic Mod el for the
United States ," J. Polit. Econ., A pril 1976a, 84(2),
pp. 207-37.
_
. "The Obs ervational Equivalence of Natural
and Unnatural Rate Theories of Macroeconomics ,"
I. Polit. Econ., June 1976b, 84(3), pp. 631-40.
A ND WA LLA C E, NEIL. "' Rational' Expecta-
tions , the Optimal Monetary Ins trument and the
Optimal Money Supply Rule,"J Polit. Econ., 1975,
83(2), pp. 241-55.
. "Rational Expectations and the Theory of
Economic Policy," J Monet. Econ., A pril 1976,
2(2), pp. 169-83.
SHILLER, ROBERT J. "Rational Expectations and the
Dynamic Structure of Macroeconomic Mod els : A
C ritical Review," J Monet. Econ., Jan. 1978, 4(1),
pp. 1-44.
SIJBEN, J. J. Rational expectations and monetary pol-
icy. A lphen aan d en Rijn, The Netherland s : Sijthoff
and Noord hoff, 1980.
TA YLOR, JOHN. "Monetary Policy d uring the Trans i-
tion to Rational Expectations ," J. Polit. Econ., 1975,
83(5), pp. 1009-21.
THEIL, HENRI. Economic forecas ts and policy. A m-
s terd am: North-Holland , 1958.
TINBERG EN, JA N. On the theory of economic policy.
A ms terd am: North-Holland , 1952.
TURNOVSKY, STEPHEN J. "Empirical Evid ence on
the Formation of Price Expectations ," J. A mer.
Statis t. A s s oc., Dec. 1970, 65(332), pp. 1441-
54.
This content downloaded from 190.223.64.239 on Tue, 25 Jun 2013 16:54:30 PM
All use subject to JSTOR Terms and Conditions