Printer Friendly

Inflation dynamics.

Policy toward inflation is a continual struggle. The Federal Reserve can reduce inflation by slowing money growth, but at a large cost: when inflation falls, as in the early 1980s, output falls as well. Recently, fears of a deeper recession have prevented the Fed from pushing toward zero inflation. In addition, even when the Fed pays the price of disinflation, the gains may not be permanent. Inflation rises again when there is an adverse macroeconomic shock, such as wartime spending (in the late 1960s) or a rise in oil prices (in the 1970s and, to some degree, in 1990). Finally, inflation is often unstable: periods of high average inflation exhibit considerable variability. The resulting uncertainty adds to the economic costs of inflation.

Could policymakers keep inflation low and stable without large output losses? It is difficult to address this question, because economists do not agree on the behavior of inflation. Indeed, many economic models suggest that this behavior should not occur--for example, disinflations should not cause the recessions that we observe in actual experience. This article describes research that seeks a better understanding of inflation, with the ultimate goal of designing better policies.

The Costs of Disinflation

Classical economics postulates that money is neutral: changes in nominal variables, such as inflation, do not affect real output. Unfortunately, this idea is refuted by history: Romer and Romer show that efforts to reduce inflation almost inevitably produce recessions.(1) The conventional explanation for this fact centers on slow adjustment of nominal wages and prices. In particular, following Taylor's influential work,(2) "New Keynesians" argue that the staggered timing of price adjustments across firms creates inertia in the price level, Since prices adjust slowly to a monetary contraction, the effects fall largely on output.

I argue that this view is incorrect.(3) In particular, it confuses levels with growth rates. Taylor shows that, with staggered adjustment, the price level adjusts slowly to a decrease in the money stock. (Intuitively, no group of firms wants to go first in substantially cutting its prices.) In modern economies, however, a monetary contraction is a slowdown in the growth of money, not an absolute decrease in the level of the money stock. I show that staggered adjustment in theory should not impede the response of inflation to such a policy. If disinflation is "credible"--if it is announced in advance, and price setters believe the announcement--then inflation should fall quickly without lower output. (Indeed, I can show that, in theory, disinflation raises output.) This theoretical analysis is unappealing because it does not fit the U.S. experience of costly disinflations. It shows that the high cost of disinflation is probably not the result of staggered price adjustments.

In another paper, I present further negative results, and also some positive ones.(4) The main alternative to New Keynesian theories of disinflation is Sargent's "New Classical" view.(5) According to Sargent, disinflations are costly because they are often not credible: price setters do not believe announcements of slower money growth. When money growth does fall, price setters are surprised, and these negative surprises reduce output.

In my view, this argument alone--like staggered adjustment alone--does not explain the effects of disinflation. If price setters have rational expectations, negative monetary surprises must be balanced by positive surprises in other episodes. Intuitively, positive surprises occur when tough-talking policymakers renege on promised disinflation. Theoretically, positive surprises should raise output just as negative surprises reduce it; announced disinflations should not reduce output on average. In reality, disinflations often cause recessions, but we do not observe significant booms when policymakers renege.

On the positive side, while neither New Keynesian nor New Classical theories are satisfactory, a combination of the two may be. My paper adds the Classical assumption of imperfect credibility to a Keynesian model of staggered price adjustment. In this case, announcements of disinflation lead to lower output not only when policy proves tighter than expected, but on average. That is, in contrast to the pure Classical case, recessions are not matched by equally large booms.

A brief explanation is that firms' fears that the Fed will renege prevent them from greatly reducing their rate of price increase. Thus the absence of credibility produces the inflation inertia that is missing in the pure Keynesian case. These results are attractive because they are empirically plausible, and because they suggest that Keynesian and Classical theories are complementary.

The Genesis of Inflation

So far I have asked why it is costly to reduce inflation. But how does high inflation arise in the first place? In the postwar United States, major increases in inflation have followed macroeconomic shocks, such as overheated fiscal policy in the late 1960s and increases in oil prices in the 1970s. The idea that such shocks are inflationary is part of conventional wisdom. However, as with the costs of disinflation, conventional wisdom has weak foundations.

The difficulty here is in explaining why the inflationary effects of shocks are persistent. In textbook models, an oil shock shifts the aggregate supply curve to the left, raising the price level. This increase in prices is a temporary increase in inflation. If the initial shock were the end of the story, inflation would return to its previous level after prices adjusted. In practice, however, oil shocks produce a persistent rise in inflation--one that lasts until the Fed creates a recession to disinflate. In other words, the shock triggers not a one-time rise in prices, but a series of increases that lasts indefinitely.

A common explanation for persistence is that an initial rise in inflation raises expectations of future inflation. The Fed accommodates these expectations by keeping inflation high, because unexpectedly low inflation would reduce output. Unfortunately, this explanation begs the question of why individuals expect inflation to stay high. Again, the direct effect of an oil shock is temporary. The expectation that inflation will continue may be self-fulfilling, but so is the expectation that it will not: if expected inflation stays low, the Fed need not inflate to avoid a recession. Why doesn't inflation end with the direct effect of the shock?

I propose a solution to this puzzle by building on the game-theoretic analysis of monetary policy initiated by Kydland and Prescott (1979).(6) As in Barro (1986), the public's inflation expectations depend on their beliefs about how strongly policymakers dislike inflation. Policymakers have an incentive to keep inflation low to convince the public that they hate inflation.

In my paper, however, a low-inflation regime can be upset by an adverse supply shock. When the shock occurs, maintaining low inflation requires anti-accommodative monetary policy, and hence higher unemployment. If a policymaker cares about unemployment as well as inflation, he allows inflation to rise even though he previously kept inflation low. This behavior reveals that the policymaker's aversion to inflation is limited, and thus raises the public's inflation expectations. Once expected inflation rises, the policymaker must raise actual inflation to avoid higher unemployment. In other words, while the direct effect of a supply shock is temporary, a policymaker's response provides information about his tastes that influences future expectations.

The Level and Variability of Inflation

Many authors have argued that a higher level of inflation produces greater uncertainty about future inflation. The costs of this uncertainty, such as greater risk in nominal contracts, are a major reason that economists dislike inflation. However, while there is a widespread belief in an inflation-uncertainty link, there is no consensus about why this relationship exists. I propose an explanation using a game-theoretic model with heterogeneous policymakers.(7)

There are both "conservative" policymakers whose only goal is to keep inflation low, and "liberals" who care about both inflation and unemployment. The relationship between inflation and uncertainty depends on the degree of policy consensus. When inflation is low, both conservatives and liberals try to maintain the status quo--nobody thinks it is desirable to raise inflation intentionally. In contrast, high inflation creates a dilemma: policymakers would like to disinflate, but fear the recession that would probably result. Conservatives are willing to pay the cost of disinflation, but liberals are not. The public does not know which faction will control policy in a given period. Thus, when inflation is high, the public is uncertain of inflation's future course.

Stephen Cecchetti and I present empirical evidence on this point.(8) Engle argues that high inflation is not associated with particularly high levels of uncertainty in the postwar United States.(9) We show, however, that this result holds only for short-run uncertainty, measured by forecast errors for next quarter's inflation. Uncertainty about inflation over the next few years is strongly related to the current level.

1 C. D. Romer and D. H. Romer, "Does Monetary Policy Matter? A New Test in the Spirit of Friedman and Schwartz," NBER Reprint No. 1353, February 1990, and in NBER Macroeconomics Annual, 1989, Volume 4, O. J. Blanchard and S. Fischer, eds. Cambridge, MA: MIT Press, 1989, pp. 121-170.

2 J. B. Taylor, "Aggregate Dynamics and Staggered Contracts," Journal of Political Economy 88 (1980), pp. 1-23.

3 L. M. Ball, "Credible Disinflation with Staggered Price Setting," NBER Working Paper No. 3555, December 1990.

4 L. M. Ball, "Disinflation with Imperfect Credibility," NBER Working Paper No. 3983, February 1992.

5 T. J. Sargent, "Stopping Moderate Inflations: The Methods of Poincare and Thatcher," in Inflation, Debt, and Indexation, R. Dornbusch and M. H. Simonsen, eds. Cambridge, MA: MIT Press, 1983, pp. 54-96.

6 L. M. Ball, "Time-Consistent Policy and Persistent Changes in Inflation," NBER Working Paper No. 3529, December 1990.

7 L. M. Ball, "Why Does High Inflation Raise Inflation Uncertainty?" NBER Working Paper No. 3224, January 1990, and Journal of Monetary Economics (June 1992), pp. 371-388.

8 L. M. Ball and S. G. Cecchetti, "Inflation and Uncertainty at Short and Long Horizons," NBER Reprint No. 1522, February 1991, and Brookings Papers on Economic Activity 1 (1990), pp. 215-245 and 253-254.

9 R.F. Engle, "Estimates of the Variance of U.S. Inflation Based upon the ARCH Model," Journal of Money, Credit and Banking 15 (1983), pp. 286-301.
COPYRIGHT 1992 National Bureau of Economic Research, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Ball, Laurence M.
Publication:NBER Reporter
Date:Sep 22, 1992
Previous Article:Public economics.
Next Article:Product variety, trade, and growth.

Related Articles
Measuring and controlling inflation.
A test of strategic interaction in monetary policy.
Korean inflation during the U.S. military administration of 1945-48.
Inflation targeting.
Mean reversion of real exchange rates in high-inflation countries.
Persistence in international inflation rates.
Monetary Economics.
Further long memory properties of inflationary shocks.
Estimating asymmetric output cost of lowering inflation for Australia.
Inflation dynamics: combining measurement with theory.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters