71-

re

-HNZ

+ SPA

+ ITA

AUsf FRA/NOR/SWE

-HDEN

+ JAP -HCAN

+ BEL

+USA

GERtSWI

l I

0.5 1 1.5 2 2.5 3 3.5 4 Index of central-bank Independence

FIGURE 9.6 Central-bank independence and inflation!

although the relationship between central-bank independence and inflation is striking, its ultimate implications remain to be determined.

Limitations of Dynamic-Inconsistency Theories of Inflation

Theories based on dynamic inconsistency provide a simple and appealing explanation of inflation. Unfortunately, it is not clear that their explanation is important to actual inflation, particularly for the industrialized countries. There are two problems. First, the importance of forward-looking expectations to aggregate supply, which is central to the dynamic-inconsistency explanation, is not weU estabhshed. For example, Canada and New Zealand have recently taken strong measures to commit themselves to low-inflation monetary policies. New Zealand, for instance, has modified the central banks charter to make price stability the sole objective of pohcy and to provide for the dismissal of the banks governor if inflation falls outside a target range. Yet, contrary to the predictions of dynamic-inconsistency models, these measures do not appear to have had a major impact on the output-inflation relationship in these countries (DebeUe, 1994).

Second, there is a great deal of variation in inflation that the dynamic-inconsistency models have difficulty accounting for. In the United States, for example, policymakers were able to reduce inflation from about 10% at the end of the 1970s to under 5% just a few years later, and to maintain the lower inflation, without any significant change in the institutions or

1 Figure 9.6, from "Central Bank Independence and Macroeconomic Performance" by Alberto Alesina and Lawrence H. Summers, Joumal of Money, Credit, and Banldng, Vol. 25, No. 2 (May 1993), is reprinted by permission. Copyright 1993 by the Ohio State University Press. All rights reserved.

D. Romer (1993b) argues that dynamir-inconsistency models predict that more open economies will have lower inflation, and that the evidence from outside the industriahzed world strongly supports this prediction. He does not find any relation between openness and mflation among industriahzed countries, however. Similarly, Cukierman, Edwards, and Tabellim (1992) find that inflation is higher in countries that are less politically stable, and they observe that this may reflect the dimimshed importance of reputation when policymakers horizons are shorter. The variauon in mstability among industriahzed countries is small, however, and thus this variable also fails to account for much of the differences in inflation among these countries.

rules governing monetary policy. Similarly, Japan has had consistently low inflation despite the fact that its central bank is not particularly independent. Indeed, if one is not willing to interpret the correlation between central-bank independence and inflation as reflecting the effects of dynamic inconsistency and delegation, it is hard to identify any important part of either the time-series or cross-section variations in inflation in the industrialized countries that is due to dynamic-inconsistency considerations.

These weaknesses of dynamic-inconsistency theories suggest that we should consider other ways that inflation could come about. In addition, there are a variety of important issues concerning how monetary policy should be conducted that do not involve dynamic inconsistency. The next section discusses some of those issues and considers several ways that inflation could arise from other sources.

9.6 Some Macroeconomic Policy Issues

The discussion m the previous two sections makes it appear that monetary policymakers face a single problem: they must find a way of getting inflation to its optimal level. Actual policymaking is much more complicated. There are two issues. First, it is not clear what the optimal rate of inflation is; this issue is addressed in Section 9.8. Second, various kinds of disturbances are continuahy affecting the economy. This section addresses some of the issues that are raised by the presence of these shocks.

What Can Policy Accomplish? A Baseline Case

How much weight should policymakers put on stabilizing output as opposed to other objectives, such as keeping inflation low and predictable? To address this issue, it is useful to begin with a simple case. Suppose that aggregate supply relates the change in inflation linearly to the departure of the unemployment rate from the natural rate, and that it has no forward-looking element (see equations [5.36]-[5.37]):

TTt = TTt-i ~ a{Ut -u) + et, a> 0, (9.26)

"In addition, as described in Chapter 5, if policymakers allow inflation to grow without bound, the aggregate supply curve (9.26) will almost surely break down. This is not relevant to the point made here, however.

We are assuming for the moment that policymakers can control inflation perfectly, subject to (9.26).

-That is, policymakers maximize = (1 + P) t-

where f f represents supply shocks. In addition, suppose that social welfare depends on unemployment and inflation, and that the dependence on unemployment is linear:

Wt =-cut - ), c>0, f"i-)>0. (9.27)

This simple model has strong implications for policy. First, the aggregate supply curve, (9.26), implies that policy has no control over average unemployment unless policymakers are willing to accept ever-increasing (or ever-decreasing) inflation. Equation (9.26) implies that the average change in inflation is determined by average unemployment and average supply shocks. Thus altering average unemployment alters the average change in inflation. But if the average change in inflation is anything other than zero, the level of inflation grows (or falls) without bound."

This result, coupled with the assumption that social welfare is linear in unemployment, implies that policy should put essentially no weight on unemployment. Suppose that policymakers discount rate is zero, and consider the first-order condition for ? Raising by a smaU amount changes current-period social welfare by both its direct effect, -(( ) ! , and its effect via unemployment, . In addition, the increase in current inflation means (for given next-period inflation) higher unemployment next period; this contributes - to social welfare. Thus the first-order condition for TTt is simply f{TTt) = 0: policymakers should keep inflation at its optimal level and pay no attention to unemployment. This is true regardless of the importance of unemployment (that is, regardless of c), and regardless of what supply shocks are buffeting the economy. Intuitively, any change in the path of inflation that does not permanently raise inflation can only rearrange the timing of unemployment, which has no effect on welfare. And with a discount rate of zero, any policy that permanently raises inflation above the optimal level has infinite costs regardless of how small inflations costs are.

With discounting, one can show that the first-order condition for TTt is

r(7rf) = ca, . (9.28)

where p is policymakers discount rate. Thus inflation should be set at the level where the cost of a permanent increase in inflation just balances the benefit of the associated one-time decrease in unemployment. Even with discounting, there is little scope for sophisticated stabilization policy: