C H A P T E R 2 7 Money and Inflation 653
the aggregate supply curve will keep on shifting leftward. As a result, policymakers are forced to accommodate the cost push by continuing to shift the aggregate demand curve to the right to eliminate the unemployment that develops. The accommodating, activist policy with its high employment target has the hidden cost or disadvantage that it may well lead to inflation.7
The main advantage of a nonaccommodating, nonactivist policy, in which policymakers do not try to shift the aggregate demand curve in response to the cost push, is that it will prevent inflation. As depicted in Figure 4, the result of an upward push on wages in the face of a nonaccommodating, nonactivist policy will be a period of unemployment above the natural rate level, which will eventually shift the aggregate supply curve and the price level back to their initial positions. The main criticism of this nonactivist policy is that the economy will suffer protracted periods of unemployment when the aggregate supply curve shifts leftward. Workers, however, would probably not push for higher wages to begin with if they knew that policy would be nonaccommodating, because their wage gains will lead to a protracted period of unemployment. A nonaccommodating, nonactivist policy may have not only the advantage of preventing inflation but also the hidden benefit of discouraging leftward shifts in the aggregate supply curve that lead to excessive unemployment.
In conclusion, if workersÕ opinions about whether policy is accommodating or nonaccommodating matter to the wage-setting process, the case for a nonactivist policy is much stronger.
Do Expectations About Policy Matter to the Wage-Setting Process? The answer to this question is crucial to deciding whether activist or nonactivist policy is preferred and so has become a major topic of current research for economists, but the evidence is not yet conclusive. We can ask, however, whether expectations about policy do affect peopleÕs behavior in other contexts. This information will help us know if expectations regarding whether policy is accommodating are important to the wage-setting process.
As any good negotiator knows, convincing your opponent that you will be nonaccommodating is crucial to getting a good deal. If you are bargaining with a car dealer over price, for example, you must convince him that you can just as easily walk away from the deal and buy a car from a dealer on the other side of town. This principle also applies to conducting foreign policyÑit is to your advantage to convince your opponent that you will go to war (be nonaccommodating) if your demands are not met. Similarly, if your opponent thinks that you will be accommodating, he will almost certainly take advantage of you (for an example, see Box 1). Finally, anyone who has dealt with a two-year-old child knows that the more you give in (pursue an accommodating policy), the more demanding the child becomes. PeopleÕs expectations about policy do affect their behavior. Consequently, it is quite plausible that expectations about policy also affect the wage-setting process.8
7The issue that is being described here is the time-consistency problem described in Chapter 21.
8A recent development in monetary theory, new classical macroeconomics, strongly suggests that expectations about policy are crucial to the wage-setting process and the movements of the aggregate supply curve. We will explore why new classical macroeconomics comes to this conclusion in Chapter 28, when we discuss the implications of the rational expectations hypothesis, which states that expectations are formed using all available information, including expectations about policy.
654 P A R T V I Monetary Theory
Box 1
Perils of Accommodating Policy
The Terrorism Dilemma. A major dilemma con- |
The terrorism dilemma illustrates the principle |
fronting our foreign policy in recent years is whether to |
that opponents are more likely to take advantage of |
cave in to the demands of terrorists when they are hold- |
you in the future if you accommodate them now. |
ing American hostages. Because our hearts go out to the |
Recognition of this principle, which demonstrates the |
hostages and their families, we might be tempted to |
perils of accommodating policy, explains why gov- |
pursue an accommodating policy of giving in to the ter- |
ernments in countries such as the United States and |
rorists to bring the hostages safely back home. |
Israel have been reluctant to give in to terrorist |
However, pursuing this accommodating policy is likely |
demands even though it has sometimes resulted in |
to encourage terrorists to take hostages in the future. |
the death of hostages. |
Rules Versus
Discretion:
Conclusions
The following conclusions can be generated from our analysis: Activists believe in the use of discretionary policy to eliminate excessive unemployment whenever it develops, because they view the wage and price adjustment process as sluggish and unresponsive to expectations about policy. Nonactivists, by contrast, believe that a discretionary policy that reacts to excessive unemployment is counter-productive, because wage and price adjustment is rapid and because expectations about policy can matter to the wage-setting process. Nonactivists thus advocate the use of a policy rule to keep the aggregate demand curve from fluctuating away from the trend rate of growth of the natural rate level of output. Monetarists, who adhere to the nonactivist position and who also see money as the sole source of fluctuations in the aggregate demand curve, in the past advocated a policy rule whereby the Federal Reserve keeps the money supply growing at a constant rate. This monetarist rule is referred to as a constant-money-growth-rate rule. Because of the misbehavior of velocity of M1 and M2, monetarists such as Bennett McCallum and Alan Meltzer of Carnegie-Mellon University have advocated a rule for the growth of the monetary base that is adjusted for past velocity changes.
As our analysis indicates, an important element for the success of a nonaccommodating policy rule is that it be credible: The public must believe that policymakers will be tough and not accede to a cost push by shifting the aggregate demand curve to the right to eliminate unemployment. In other words, government policymakers need credibility as inflation-fighters in the eyes of the public. Otherwise, workers will be more likely to push for higher wages, which will shift the aggregate supply curve leftward after the economy reaches full employment at a point such as point 2 in Figure 11 and will lead to unemployment or inflation (or both). Alternatively, a credible, nonaccommodating policy rule has the benefit that it makes a cost push less likely and thus helps prevent inflation and potential increases in unemployment. The following application suggests that recent historical experience is consistent with the importance of credibility to successful policymaking.
656 P A R T V I Monetary Theory
Key Terms
accommodating policy, p. 640 |
demand-pull inflation, p. 639 |
printing money, p. 644 |
constant-money-growth-rate rule, |
government budget constraint, p. 643 |
Ricardian equivalence, p. 645 |
p. 654 |
monetizing the debt, p. 644 |
|
|
|
cost-push inflation, p. 639 |
|
|
QUIZ Questions and Problems
Questions marked with an asterisk are answered at the end of the book in an appendix, ÒAnswers to Selected Questions and Problems.Ó
1.ÒThere are frequently years when the inflation rate is high and yet money growth is quite low. Therefore, the statement that inflation is a monetary phenomenon cannot be correct.Ó Comment.
*2. Why do economists focus on historical episodes of hyperinflation to decide whether inflation is a monetary phenomenon?
3.ÒSince increases in government spending raise the aggregate demand curve in Keynesian analysis, fiscal policy by itself can be the source of inflation.Ó Is this statement true, false, or uncertain? Explain your answer.
*4. ÒA cost-push inflation occurs as a result of workersÕ attempts to push up their wages. Therefore, inflation does not have to be a monetary phenomenon.Ó Is this statement true, false, or uncertain? Explain your answer.
5.ÒBecause government policymakers do not consider inflation desirable, their policies cannot be the source of inflation.Ó Is this statement true, false, or uncertain? Explain your answer.
*6. ÒA budget deficit that is only temporary cannot be the source of inflation.Ó Is this statement true, false, or uncertain? Explain your answer.
7.How can the FedÕs desire to prevent high interest rates lead to inflation?
*8. ÒIf the data and recognition lags could be reduced, activist policy would more likely be beneficial to the
economy.Ó Is this statement true, false, or uncertain? Explain your answer.
9.ÒThe more sluggish wage and price adjustment is, the more variable output and the price level are when an activist policy is pursued.Ó Is this statement true, false, or uncertain? Explain your answer.
*10. ÒIf the public believes that the monetary authorities will pursue an accommodating policy, a cost-push inflation is more likely to develop.Ó Is this statement true, false, or uncertain? Explain your answer.
11.Why are activist policies to eliminate unemployment more likely to lead to inflation than nonactivist policies?
*12. ÒThe less important expectations about policy are to movements of the aggregate supply curve, the stronger the case is for activist policy to eliminate unemployment.Ó Is this statement true, false, or uncertain? Explain your answer.
13.If the economyÕs self-correcting mechanism works slowly, should the government necessarily pursue an activist policy to eliminate unemployment?
*14. ÒTo prevent inflation, the Fed should follow Teddy RooseveltÕs advice: ÔSpeak softly and carry a big stick.ÕÓ What would the FedÕs Òbig stickÓ be? What is the statement trying to say?
15.In a speech early in the Iraq-Kuwait crisis in 1990, President George Bush stated that although his heart went out to the hostages held by Saddam Hussein, he would not let this hostage-taking deter the United States from insisting on the withdrawal of Iraq from Kuwait. Do you think that BushÕs position made sense? Explain why or why not.
C h a p t e r
28 Rational Expectations: Implications for Policy
After World War II, economists, armed with Keynesian models (such as the ISL M model) that described how government policies could be used to manipulate employment and output, felt that activist policies could reduce the severity of business cycle fluctuations without creating inflation. In the 1960s and 1970s, these economists got their chance to put their policies into practice (see Chapter 27), but the results were not what they had anticipated. The economic record for that period is not a happy one: Inflation accelerated, the rate often climbing above 10%, while unemployment figures deteriorated from those of the 1950s.1
In the 1970s and 1980s, economists, including Robert Lucas of the University of Chicago and Thomas Sargent, now at New York University, used the rational expectations theory discussed in Chapter 7 to examine why activist policies appear to have performed so poorly. Their analysis cast doubt on whether macroeconomic models can be used to evaluate the potential effects of policy and on whether policy can be effective when the public expects that it will be implemented. Because the analysis of Lucas and Sargent has such strong implications for the way policy should be conducted, it has been labeled the rational expectations revolution.2
This chapter examines the analysis behind the rational expectations revolution. We start first with the Lucas critique, which indicates that because expectations are important in economic behavior, it may be quite difficult to predict what the outcome of an activist policy will be. We then discuss the effect of rational expectations on the aggregate demand and supply analysis developed in Chapter 25 by exploring three models that incorporate expectations in different ways.
A comparison of all three models indicates that the existence of rational expectations makes activist policies less likely to be successful and raises the issue of credibility as an important element affecting policy outcomes. With rational expectations, an essential ingredient to a successful anti-inflation policy is the credibility of the policy in the eyes of the public. The rational expectations revolution is now at the center of many of the current debates in monetary theory that have major implications for how monetary and fiscal policy should be conducted.
1Some of the deterioration can be attributed to supply shocks in 1973Ð1975 and 1978Ð1980.
2Other economists who have been active in promoting the rational expectations revolution are Robert Barro of Harvard University, Bennett McCallum of Carnegie-Mellon University, Edward Prescott of the University of Minnesota, and Neil Wallace of Pennsylvania State University.
C H A P T E R 2 8 Rational Expectations: Implications for Policy 659
The Lucas Critique of Policy Evaluation
http://cepa.newschool.edu /het/profiles/lucas.htm
A brief biography of Robert Lucas, including a list of his publications.
In his famous paper ÒEconometric Policy Evaluation: A Critique,Ó Robert Lucas presented an argument that had devastating implications for the usefulness of conventional econometric models (models whose equations are estimated with statistical procedures) for evaluating policy.3 Economists developed these models for two purposes: to forecast economic activity and to evaluate the effects of different policies. Although LucasÕs critique had nothing to say about the usefulness of these models as forecasting tools, he argued that they could not be relied on to evaluate the potential impact of particular policies on the economy.
Econometric
Policy Evaluation
Example: The Term
Structure
of Interest Rates
To understand LucasÕs argument, we must first understand econometric policy evaluation: how econometric models are used to evaluate policy. For example, we can examine how the Federal Reserve uses its econometric model in making decisions about the future course of monetary policy. The model contains equations that describe the relationships among hundreds of variables. These relationships are assumed to remain constant and are estimated using past data. LetÕs say that the Fed wants to know the effect on unemployment and inflation of a decrease in the fed funds rate from 5% to 4%. It feeds the new, lower fed funds rate into a computer that contains the model, and the model then provides an answer about how much unemployment will fall as a result of the lower fed funds rate and how much the inflation rate will rise. Other possible policies, such as a rise in the fed funds rate by one percentage point, might also be fed into the model. After a series of these policies have been tried out, the policymakers at the Fed can see which policies produce the most desirable outcome for unemployment and inflation.
LucasÕs challenge to this procedure for evaluating policies is based on a simple principle of rational expectations theory: The way in which expectations are formed (the relationship of expectations to past information) changes when the behavior of forecasted variables changes. So when policy changes, the relationship between expectations and past information will change, and because expectations affect economic behavior, the relationships in the econometric model will change. The econometric model, which has been estimated with past data, is then no longer the correct model for evaluating the response to this policy change and may consequently prove highly misleading.
The best way to understand LucasÕs argument is to look at a concrete example involving only one equation typically found in econometric models: the term structure equation. The equation relates the long-term interest rate to current and past values of the short-term interest rate. It is one of the most important equations in Keynesian econometric models because the long-term interest rate, not the short-term rate, is the one believed to have an impact on aggregate demand.
In Chapter 6, we learned that the long-term interest rate is related to an average of expected future short-term interest rates. Suppose that in the past, when the shortterm rate rose, it quickly fell back down again; that is, any increase was temporary. Because rational expectations theory suggests that any rise in the short-term interest rate is expected to be only temporary, a rise should have only a minimal effect on the
3Carnegie-Rochester Conference Series on Public Policy 1 (1976): 19Ð46.
660 P A R T V I Monetary Theory
average of expected future short-term rates. It will cause the long-term interest rate to rise by a negligible amount. The term structure relationship estimated using past data will then show only a weak effect on the long-term interest rate of changes in the short-term rate.
Suppose the Fed wants to evaluate what will happen to the economy if it pursues a policy that is likely to raise the short-term interest rate from a current level of 5% to a permanently higher level of 8%. The term structure equation that has been estimated using past data will indicate that there will be just a small change in the longterm interest rate. However, if the public recognizes that the short-term rate is rising to a permanently higher level, rational expectations theory indicates that people will no longer expect a rise in the short-term rate to be temporary. Instead, when they see the interest rate rise to 8%, they will expect the average of future short-term interest rates to rise substantially, and so the long-term interest rate will rise greatly, not minimally as the estimated term structure equation suggests. You can see that evaluating the likely outcome of the change in Fed policy with an econometric model can be highly misleading.
The term structure example also demonstrates another aspect of the Lucas critique. The effects of a particular policy depend critically on the publicÕs expectations about the policy. If the public expects the rise in the short-term interest rate to be merely temporary, the response of long-term interest rates, as we have seen, will be negligible. If, however, the public expects the rise to be more permanent, the response of long-term rates will be far greater. The Lucas critique points out not only that conventional econometric models cannot be used for policy evaluation, but also that the publicÕs expectations about a policy will influence the response to that policy.
The term structure equation discussed here is only one of many equations in econometric models to which the Lucas critique applies. In fact, Lucas uses the examples of consumption and investment equations in his paper. One attractive feature of the term structure example is that it deals with expectations in a financial market, a sector of the economy for which the theory and empirical evidence supporting rational expectations are very strong. The Lucas critique should also apply, however, to sectors of the economy for which rational expectations theory is more controversial, because the basic principle of the Lucas critique is not that expectations are always rational but rather that the formation of expectations changes when the behavior of a forecasted variable changes. This less stringent principle is supported by the evidence in sectors of the economy other than financial markets.
New Classical Macroeconomic Model
We now turn to the implications of rational expectations for the aggregate demand and supply analysis we studied in Chapter 25. The first model we examine that views expectations as rational is the new classical macroeconomic model developed by Robert Lucas and Thomas Sargent, among others. In the new classical model, all wages and prices are completely flexible with respect to expected changes in the price level; that is, a rise in the expected price level results in an immediate and equal rise in wages and prices because workers try to keep their real wages from falling when they expect the price level to rise.
This view of how wages and prices are set indicates that a rise in the expected price level causes an immediate leftward shift in the aggregate supply curve, which
C H A P T E R 2 8 Rational Expectations: Implications for Policy 661
Effects of
Unanticipated and
Anticipated Policy
leaves real wages unchanged and aggregate output at the natural rate (full-employment) level if expectations are realized. This model then suggests that anticipated policy has no effect on aggregate output and unemployment; only unanticipated policy has an effect.
First, let us look at the short-run response to an unanticipated (unexpected) policy such as an unexpected increase in the money supply.
In Figure 1, the aggregate supply curve AS1 is drawn for an expected price level P1. The initial aggregate demand curve AD1 intersects AS1 at point 1, where the realized price level is at the expected price level P1 and aggregate output is at the natural rate level Yn. Because point 1 is also on the long-run aggregate supply curve at Yn, there is no tendency for the aggregate supply to shift. The economy remains in longrun equilibrium.
Suppose the Fed suddenly decides the unemployment rate is too high and so makes a large bond purchase that is unexpected by the public. The money supply increases, and the aggregate demand curve shifts rightward to AD2. Because this shift is unexpected, the expected price level remains at P1 and the aggregate supply curve remains at AS1. Equilibrium is now at point 2 , the intersection of AD2 and AS1. Aggregate output increases above the natural rate level to Y2 and the realized price level increases to P2 .
If, by contrast, the public expects that the Fed will make these open market purchases in order to lower unemployment because they have seen it done in the past, the expansionary policy will be anticipated. The outcome of such anticipated expansionary policy is illustrated in Figure 2. Because expectations are rational, workers and firms recognize that an expansionary policy will shift the aggregate demand curve
F I G U R E 1 Short-Run Response
to Unanticipated Expansionary Policy
in the New Classical Model
Initially, the economy is at point 1 at the intersection of AD1 and AS1 (expected price level P1) . An expansionary policy shifts the aggregate demand curve to AD2, but because this is unexpected, the aggregate supply curve remains fixed at AS1. Equilibrium now occurs at point 2 Ñaggregate output has increased above the natural rate level to Y2 , and the price level has increased to P2 .
Aggregate |
|
|
Price Level, P |
|
AS1 |
|
|
|
|
(expected |
|
|
price level = P1) |
P2 |
|
2 |
P1 |
1 |
AD2 |
|
|
AD1 |
|
Yn |
Y2 |
662 P A R T V I Monetary Theory
F I G U R E 2 Short-Run Response
to Anticipated Expansionary Policy in
the New Classical Model
The expansionary policy shifts the aggregate demand curve rightward to AD2 , but because this policy is expected, the aggregate supply curve shifts leftward to AS2. The economy moves to point 2, where aggregate output is still at the natural rate level but the price level has increased to P2.
Aggregate |
AS2 |
Price Level, P |
(expected price level = P2) |
|
AS1 |
|
(expected |
|
price level = P1) |
P2 |
2 |
AD1
to the right and will expect the aggregate price level to rise to P2. Workers will demand higher wages so that their real earnings will remain the same when the price level rises. The aggregate supply curve then shifts leftward to AS2 and intersects AD2 at point 2, an equilibrium point where aggregate output is at the natural rate level Yn and the price level has risen to P2.
The new classical macroeconomic model demonstrates that aggregate output does not increase as a result of anticipated expansionary policy and that the economy immediately moves to a point of long-run equilibrium (point 2) where aggregate output is at the natural rate level. Although Figure 2 suggests why this occurs, we have not yet proved why an anticipated expansionary policy shifts the aggregate supply curve to exactly AS2 (corresponding to an expected price level of P2) and hence why aggregate output necessarily remains at the natural rate level. The proof is somewhat difficult and is dealt with in Box 1.
The new classical model has the word classical associated with it because when policy is anticipated, the new classical model has a property that is associated with the classical economists of the nineteenth and early twentieth centuries: Aggregate output remains at the natural rate level. Yet the new classical model allows aggregate output to fluctuate away from the natural rate level as a result of unanticipated movements in the aggregate demand curve. The conclusion from the new classical model is a striking one: Anticipated policy has no effect on the business cycle; only unanticipated policy matters.4
4Note that the new classical view in which anticipated policy has no effect on the business cycle does not imply that anticipated policy has no effect on the overall health of the economy. For example, the new classical analysis does not rule out possible effects of anticipated policy on the natural rate of output Yn, which can benefit the public.