Central bank independence and disinflationary credibility: a missing link?(Special Issue on the Economics of Central Bank Independence)

1. Introduction

Central bank independence is rapidly becoming the worldwide norm for the organization of monetary authorities. A remarkably broad consensus of academics and policymakers, conservatives and liberals, have rallied around the idea that freeing monetary policymaking from political direction eases the attainment of price stability at little or no real economic cost.(1) The erection of legal barriers to elected government control of monetary policy has been agreed to (where they were not already in place) in virtually all the member countries of the European Union, most of the remaining members of the Organization for Economic Cooperation and Development (OECD), and a growing number of both newly industrializing economies and former communist economies in transition.

This convergence upon a particular institutional fix for inflation stems from the wide acceptance of the analysis of two stylized macroeconomic facts in the papers of Kydland and Prescott (1977) and Barro and Gordon (1983): that the long-run Phillips curve is vertical - that is, inflation has no permanent effect on real outcomes; and that governments nonetheless have an incentive to spring inflationary surprises upon the public. As a result, these papers argued, a primary cause of inflation was government's inability in the eyes of the public to commit credibly to a low inflation policy. One could remove the time-inconsistency problem by making government unable to renege upon a commitment to low inflation. In Rogoff (1985), the appointment of a conservative central banker was shown to be one means to that end. Since then, legal central bank independence has been identified with a credible commitment to price stability.(2) This credibility bonus is presumed to be the source of the widely known negative correlation between central bank independence and average inflation rates.

Despite the theoretical and policy attention paid to central bank independence, however, no previous study(3) has directly examined the post-war experience prior to recent developments for evidence of credibility effects of central bank independence. The coexistence of the Barro-Gordon/Rogoff explanation with the robust negative inflation correlation for central bank independence seems to have satisfied curiosity on this point.(4) Yet, this coexistence, while necessary to establish a causal link between central bank independence and low inflation, is insufficient to do so. If independent central banks' policies are inherently more credible, not only inflation levels but also expectations of monetary policy must differ systematically across countries with differences in central bank independence. This paper examines cross-sectionally public- and private-sector behavior in a sample of 17 OECD countries from 1950-89 for evidence of such differences.

A direct link between central bank independence and disinflationary credibility is not supported by this paper's results. Disinflation appears to be consistently more costly and no more rapid in countries with independent central banks, even holding wage indexation constant. Moreover, neither wage nor price rigidities as measured in this paper are found to vary significantly across countries with different central bank institutions. This invariance indicates that the greater price stability correlated with central bank independence does not induce the lower frequency of recontracting one might expect from a credibility-caused inflation reduction. Furthermore, the hypothesis that the elected governments' attempts to collect seignorage revenues or to manipulate macroeconomic policy for electoral gain will be less prevalent where central banks are more independent is rejected.

The paper is organized as follows. Section 2 discusses what theory predicts would be the effects on the disinflationary process and on contracting behavior of increased central bank independence if such an increase represents a greater credibility of commitment to price stability. Section 3 examines the historical record for evidence of these predicted private-sector behavior differences. Section 4 discusses how public-sector behavior should be affected by central bank independence if the independence does credibly constrain government opportunism with monetary policy, and performs a search across countries for evidence of those predicted differences. Section 5 summarizes the results and considers their implications for explanations of the negative correlation between central bank independence and inflation.

2. Predicted effects of central bank independence as a credibility enhancement

Let us start with the assumptions upon which the whole argument for central bank independence is based: individuals are (at least somewhat) forward looking and the long-run Phillips curve is vertical. Under those assumptions, if central bank independence is a mechanism through which governments increase the credibility of their commitments to price stability, then an increase in a country's central bank independence should reduce the incentive for that country's private decision-makers to alter their behavior to protect themselves from government-induced inflation surprises. Inflation rates should be lower in countries with higher central bank independence, but only as an observable result of differences in the behavior of private agents. The purpose of this section is to derive empirically testable predictions from these theorized effects.

The first logical place to look for private-sector effects of more credible monetary policymaking is in disinflations. That is to say, when a change in policy, such as a monetary tightening, is announced, what does it matter that people are more likely to believe in and respond to the announcement when a more independent central bank makes it? Of course, the role of credibility in determining the course of disinflations has been widely discussed, especially since Sargent (1982). In the flexible price, natural rate world that he describes - the world in which the Barro-Gordon (1983) analysis also is set - expected money growth has no real effect. Consequently, it is evident that a disinflationary announcement that is believed changes the price level with no cost in terms of unemployment or output.

More interestingly, credibility appears to have an effect in less classical worlds as well. In all the various New and neo-Keynesian models that rely on staggered contracts, sticky prices, or sticky wages to produce real effects of monetary policy - including Barro and Grossman (1976), Fischer (1977), Taylor (1980), Calvo (1982), and Buiter and Miller (1982) - the presence of price level inertia is no impediment to costless disinflation if the announced policy is completely believed and properly timed.(5) Fischer (1985) and Chadha et al. (1992) simulate disinflations under differing levels of credibility for specific staggered long-term wage contracting models, and both find that as the time necessary for the public's expectations of monetary policy to match actual policy shrinks, so do the costs of disinflation. Ball (1992) derives the general result that increased credibility(6) decreases the costs of disinflation, even where staggering of price and wage setting exists.

In short, recession will occur with monetary tightening only to the degree that the rate of growth of the money stock falls more than expected. Enhanced credibility of monetary policy, whether in a flexible or sticky wage-price environment, should reduce the distance between expectations and reality. Thus:

Prediction 1 Where central bank independence is greater, the cost of disinflation should be lower, ceteris paribus.

There are, of course, other factors that determine the slope of the Phillips curve or output-inflation trade-off, but to whatever extent people are able to react to policy (for example, are not bound by long-term contracts), the clearer signal that a more credible policy conveys must translate into a smaller real effect of monetary policy.(7)

Naturally, the extent of private actors' freedom to react to policy is itself a matter of choice. People can be supposed to pick the nature of their contracting arrangements in response to the situation in which they find themselves. In all the models discussed above, however, the degree of nominal rigidity in the economy is treated as parametric. While this modeling assumption is understandable for the purposes of those papers, it shapes this paper's investigation of the credibility effects of central bank independence in two ways. The first, straightforward way is that a clear set of control variables to use in isolating the credibility bonus of central bank independence suggest themselves. Uncontrolled regressions of sacrifice ratios on central bank independence will likely have significant omitted variable bias toward a smaller effect as nominal rigidity flattens the Phillips curve.

The second impact of recognizing the endogeneity of contracting behavior is to yield a new set of predictions with which to test for credibility effects of central bank independence. If we assume that people in a low-inflation and low-inflation-uncertainty environment would prefer to contract in longer nominal terms in order to economize on the costs of renegotiating, information gathering, and so on,(8) then a policymaking regime that offers greater credibility for price stability should encourage greater nominal rigidity than one that does not. Consequently:

Prediction 2 Where central bank …