. 96
( 126 .)


Test Yourself Question 1 at the end of the chapter asks you to demonstrate that a vertical aggregate supply

curve leads to a vertical Phillips curve.

Copyright 2009 Cengage Learning, Inc. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.
Licensed to:
Part 3
328 Fiscal and Monetary Policy

real wages will decline by 2 percent. More generally, real wages will fall if workers under-
estimate inflation at all. The effects of inflation on real wages will be somewhere in
between those shown in Tables 1 and 2.4 So firms will retain some incentive to raise pro-
duction as the price level rises, which means that the aggregate supply curve will retain
some upward slope. We thus conclude that
The short-run aggregate supply curve is vertical when inflation is predicted accurately
but upward-sloping when inflation is underestimated. Thus, only unexpectedly high infla-
tion will raise output, because only unexpected inflation reduces real wages.5 Similarly,
only an unexpected decline in inflation will lead to a recession.
Because people often fail to anticipate changes in inflation correctly, this analysis seems
to leave our earlier discussion of the Phillips curve almost intact for practical purposes.
Indeed, most economists nowadays believe that the Phillips curve slopes downward in
the short run but is vertical in the long run.

However, an influential minority of economists disagrees. This group, believers in the
hypothesis of rational expectations, insists that the Phillips curve is vertical even in the
Rational expectations
are forecasts that, while not short run. To understand their point of view, we must first explain what rational expecta-
necessarily correct, are the tions are. Then we will see why rational expectations have such radical implications for
best that can be made given
the trade-off between inflation and unemployment.
the available data. Rational
expectations, therefore,
What Are Rational Expectations?
cannot err systematically. If
expectations are rational,
In many economic contexts, people must formulate expectations about what the future
forecasting errors are pure
random numbers. will bring. For example, those who invest in the stock market need to forecast the future
prices of the stocks they buy and sell. Likewise, as we have just discussed, workers and
businesses may want to forecast future prices before agreeing on a money wage. Rational
expectations is a controversial hypothesis about how such forecasts are made.
As used by economists, a forecast (an “expectation”) of a future variable is considered
rational if the forecaster makes optimal use of all relevant information that is available at
the time of the forecast. Let us elaborate on the two italicized words in this definition, us-
ing as an example a hypothetical stock market investor who has rational expectations.
First, proponents of rational expectations recognize that information is limited. An in-
vestor interested in Google stock would like to know how much profit the company will
make in the coming years. Armed with such information, she could predict the future
price of Google stock more accurately. But that information is simply unavailable. The in-
vestor™s forecast of the future price of Google shares is not “irrational” just because she
cannot foresee the future. On the other hand, if Google stock normally goes down on
Fridays and up on Mondays, she should be aware of this fact.
Next, we have the word optimal. As used by economists, it means using proper statisti-
cal inference to process all the relevant information that is available before making a fore-
cast. In brief, to have rational expectations, your forecasts do not have to be correct, but
they cannot have systematic errors that you could avoid by applying better statistical
methods. This requirement, although exacting, is not quite as outlandish as it may seem.
A good billiards player makes expert use of the laws of physics even without understand-
ing the theory. Similarly, an experienced stock market investor may make good use of
information even without formal training in statistics.

To make sure you understand why, construct a version of Table 2 based on the assumption that workers expect

4 percent inflation (and hence set next year™s wage at $10.40 per hour), regardless of what the actual rate of infla-
tion is. If you create this table correctly, it will show that higher inflation leads to lower real wages, as in Table 1.
To see this point, compare Tables 1 and 2.

Copyright 2009 Cengage Learning, Inc. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.
Licensed to:

Chapter 16 329
The Trade-Off between Inflation and Unemployment

Rational Expectations and the Trade-Off
Let us now see how some economists have used the hypothesis of rational expectations to
deny any trade-off between inflation and unemployment”even in the short run.
Although they recognize that inflation cannot always be predicted accurately, propo-
nents of rational expectations insist that workers will not make systematic errors. Remem-
ber that our argument leading to a sloping short-run Phillips curve tacitly assumed that
workers are slow to recognize changes. They thus underestimate inflation when it is rising
and overestimate it when it is falling. Many observers see such systematic errors as a realis-
tic description of human behavior. But advocates of rational expectations disagree, claim-
ing that it is fundamentally illogical. Workers, they argue, will always make the best
possible forecast of inflation, using all the latest data and the best available economic
models. Such forecasts will sometimes be too high and sometimes too low, but they will
not err systematically in one direction or the other. Consequently:
If expectations are rational, the difference between the actual rate of inflation and the
expected rate of inflation (the forecasting error) must be a pure random number, that is:
Inflation 2 Expected inflation 5 A random number
Now recall that the argument in the previous section concluded that employment is
affected by inflation only to the extent that inflation differs from what was expected. But,
under rational expectations, no predictable change in inflation can make the expected rate of
inflation deviate from the actual rate of inflation. The difference between the two is simply
a random number. Hence, according to the rational expectations hypothesis, unemploy-
ment will always remain at the natural rate”except for random, and therefore totally
unpredictable, gyrations due to forecasting errors. Thus:
If expectations are rational, inflation can be reduced without a period of high unem-
ployment because the short-run Phillips curve, like the long-run Phillips curve, will be
According to the rational expectations view, the government™s ability to manipulate ag-
gregate demand gives it no ability to influence real output and unemployment because the
aggregate supply curve is vertical even in the short run. (To see why, experiment by moving
an aggregate demand curve when the aggregate supply curve is vertical, as in Figure 10(a)
on page 327.) The government™s manipulations of aggregate demand are planned ahead
and are therefore predictable, and any predictable change in aggregate demand will change
the expected rate of inflation. It will therefore leave real wages unaffected.
The government can influence output only by making unexpected changes in aggregate
demand. But unexpected changes are not easy to engineer if expectations are rational, be-
cause people will understand what policy makers are up to. For example, if the authori-
ties typically react to high inflation by reducing aggregate demand, people will soon come
to anticipate this reaction. And anticipated reductions in aggregate demand do not cause
unexpected changes in inflation.

An Evaluation
Believers in rational expectations are optimistic about reducing inflation without losing
any output, even in the short run. But are they right?
As a piece of pure logic, the rational expectations argument is impeccable. But as is
common in the world of economic policy, controversy arises over how well the theoretical
idea applies in practice. Although the theory has attracted many adherents, the evidence
to date leads most economists to reject the extreme rational expectations position in favor
of the view that a trade-off between inflation and unemployment does exist in the short
run. Here are some of the reasons.

Contracts May Embody Outdated Expectations Many contracts for labor and other
raw materials cover such long periods of time that the expectations on which they were

Copyright 2009 Cengage Learning, Inc. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.
Licensed to:
Part 3
330 Fiscal and Monetary Policy

based, although perhaps rational at the time, may appear “irrational” from today™s point of
view. For example, some three-year labor contracts were drawn up in 1996, when inflation
had been running near 3 percent for years. It might have been rational then to expect the
1999 price level to be about 9 percent higher than the 1996 price level, and to have set wages
for 1999 accordingly. By 1997, however, inflation had fallen to below 2 percent, and such an
expectation would have been plainly irrational. But it might already have been written into
contracts. If so, real wages wound up higher than intended, giving firms an incentive to re-
duce output and therefore employment”even though no one behaved irrationally.

Expectations May Adjust Slowly Many people believe that inflationary expectations
do not adapt as quickly to changes in the economic environment as the rational expecta-
tions theory assumes. If, for example, the government embarks on an anti-inflation policy,
workers may continue to expect high inflation for a while. Thus, they may continue to in-
sist on rapid money wage increases. Then, if inflation actually slows down, real wages
will rise faster than anyone expected, and unemployment will result. Such behavior may
not be strictly rational, but it may be realistic.

When Do Workers Receive Compensation for Inflation? Some observers ques-
tion whether wage agreements typically compensate workers for expected inflation in ad-
vance, as assumed by the rational expectations theory. More typically, they argue, wages
catch up to actual inflation after the fact. If so, real wages will be eroded by inflation for a
while, as in the conventional view.

What the Facts Show The facts have not been kind to the rational expectations hy-
pothesis. The theory suggests that unemployment should hover around the natural rate
most of the time, with random gyrations in one direction or the other. Yet this is not what
the data show. The theory also predicts that preannounced (and thus expected) anti-
inflation programs should be relatively painless. Yet, in practice, fighting inflation has
proved very costly in virtually every country. Finally, many direct tests of the rationality
of expectations have cast doubt on the hypothesis. For example, survey data on people™s
expectations rarely meet the exacting requirements of rationality.
But all these problems with rational expectations should not obscure a basic truth. In
the long run, the rational expectations view should be more or less correct because people
will not cling to incorrect expectations indefinitely. As Abraham Lincoln pointed out with
characteristic wisdom, you cannot fool all the people all the time.

This chapter has now taught us some of the reasons why economists disagree about the
proper conduct of stabilization policy. It also helps us understand some of the related
political debates.
Should the government take strong actions to prevent or reduce inflation? You will say
yes if you believe that (1) inflation is more costly than unemployment, (2) the short-run
Phillips curve is steep, (3) expectations react quickly, and (4) the economy™s self-correcting
mechanism works smoothly and rapidly. These views on the economy tend to be held by
believers in rational expectations.
But you will say no if you believe that (1) unemployment is more costly than inflation,
(2) the short-run Phillips curve is flat, (3) expectations react sluggishly, and (4) the self-
correcting mechanism is slow and unreliable. These views are held by many Keynesian
economists, so it is not surprising that they often oppose using recession to fight
The tables turn, however, when the question becomes whether to use demand manage-
ment to bring a recession to a rapid end. The Keynesian view of the world”that unem-
ployment is costly, that the short-run Phillips curve is flat, that expectations adjust slowly,
and that the self-correcting mechanism is unreliable”leads to the conclusion that the

Copyright 2009 Cengage Learning, Inc. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.
Licensed to:

Chapter 16 331
The Trade-Off between Inflation and Unemployment

benefits of fighting unemployment are high and the costs are low. Keynesians are there-
fore eager to fight recessions. The rational expectations positions on these four issues are
precisely the reverse, and so are the policy conclusions.

We have seen that policy makers face an unavoidable trade-off. If they stimulate aggre-
gate demand to reduce unemployment, they will aggravate inflation. If they restrict
aggregate demand to fight inflation, they will cause higher unemployment.
But wait. Early in the chapter we learned that when inflation comes from the supply
side, inflation and unemployment are positively correlated: They go up or down together.
Does this mean that monetary and fiscal policy makers can escape the trade-off between
inflation and unemployment? Unfortunately not.
Shifts of the aggregate supply curve can cause inflation and unemployment to rise or
fall together, and thus can destroy the statistical Phillips curve relationship. Neverthe-
less, anything that monetary and fiscal policy can do will make unemployment and in-
flation move in opposite directions because monetary and fiscal policies influence only
the aggregate demand curve, not the aggregate supply curve.
Thus, no matter what the source of inflation, and no matter what happens to the
Phillips curve, the monetary and fiscal policy authorities still face a disagreeable trade-
off between inflation and unemployment. Many policy makers have failed to under-
stand this principle, which is one of the Ideas we hope you will remember well Beyond BEYOND THE
the Final Exam.
Naturally, the unpleasant nature of this trade-off has led both economists and public
officials to search for a way out of the dilemma. We conclude this chapter by considering
some of these ideas”none of which is a panacea.

One highly desirable approach”if only we knew how to do it”would be to reduce the
natural rate of unemployment. Then we could enjoy lower unemployment without higher
inflation. The question is: How?
The most promising approaches have to do with education, training, and job placement.
The data clearly show that more educated workers are unemployed less frequently than
less educated ones are. Vocational training and retraining programs, if successful, help un-
employed workers with obsolete skills acquire abilities that are currently in demand. By so
doing, they both raise employment and help alleviate upward pressures on wages in jobs
where qualified workers are in short supply. Government and private job placement and
counseling services play a similar role. Such programs try to match workers to jobs better
by funneling information from prospective employers to prospective employees.
These ideas sound sensible and promising, but two big problems arise in implementa-
tion. First, training and placement programs sometimes look better on paper than in prac-
tice. In some cases, people are trained for jobs that do not exist by the time they finish their


. 96
( 126 .)