Home | Business News | Browse by Publication | J | Journal of Money, Credit & Banking

Monetary policy and inflation in the 70s.

Article Excerpt
THE CAUSES of the "great" inflation of the 1970s remain the subject of debate. One of the most popular explanations has been provided by Clarida, Gall, and Gertler (2000) (henceforth CGG). CGG argue that the great inflation was caused by a well-meaning, nonopportunistic Fed that committed an honest, technical mistake. In particular, the Fed is alleged to have employed a policy rule that involved too weak of a reaction to expected inflation. This triggered--inadvertently--indeterminacies and allowed self-fulfilling inflation expectations to contribute to higher inflation. CGG claim that the empirical evidence favors their interpretation. When estimating a policy rule for the pre-Volcker period, they find that, unlike the rule for the post-Volcker era, it indeed violates the Taylor principle.

The CGG thesis has its critics. For instance, Orphanides (2004) has repeated the CGG exercise using real-time data on inflation and output as well as a real-time measure of potential output (partly constructed by the Department of Commerce). His main result is that the estimated interest rate rule does not differ significantly across the pre- and post-Volcker periods. And that it satisfies determinacy in both periods. This result seems to owe much to the assumption that the Fed in the 70s formulated monetary policy on the basis of a very large and persistent, perceived output gap. While some part of this gap came from measurement error in actual output, most of it arose from the estimate of potential output that was prepared by the Council of Economic Advisors.

An alternative test of the CGG thesis has been undertaken by Lubik and Schorfheide (2004) (henceforth L-S). They estimate a small scale, forward-looking New Keynesian model without restricting the parameters to lie in the determinacy region. They specify a prior probability distribution over parameters that places equal weight on determinate and indeterminate regions of the parameter space and compute posterior odds ratios for these regions. Their main result regarding the policy rule supports the CGG claim of indeterminacy. (1)

Does the L-S analysis provide an adequate framework for evaluating the case for indeterminacy? A possible problem with their approach--a problem they acknowledged--regards its sensitivity to model misspecification. In particular, the endogenous dynamics are richer in the indeterminacy region of the parameter space than in the determinacy region. Omitted propagation mechanisms under determinacy may thus bias the posteriors toward indeterminacy. (2) To deal with this problem they include additional sources of dynamics under determinacy (such as habit persistence and backward-price indexation) and find that they do not materially affect the results. (3)

Our view is that such rigidities are unlikely to help the New Keynesian model generate large inflation volatility and persistence under determinacy. (4) Consequently, the case for determinacy may not have received a "fair hearing" in the L-S specification. In this paper we propose an alternative, more plausible source of inflation inertia that has been identified in the literature as a possible important contributor to the great inflation. Namely, misperceptions about the state of the economy and gradual learning, an element that has been emphasized by Orphanides (2004). We examine whether the CGG thesis survives against this alternative specification.

Naturally, we are not the first ones to study the contribution of this type of misperceptions to the great inflation. There exists a large literature that has looked at this issue (5) and that claims that misperceptions may have indeed been the driving force behind the inflation of the 70s. Nevertheless, we find the message from this literature rather incomplete for three reasons.

First, much of this literature has been conducted in the context of backward-looking models (where all or some of the agents do not employ rational expectations). Second, with rare exceptions, the learning mechanism utilized is not rational. Typically, the monetary authorities are assumed to make use of plausible but arbitrary forecasting rules. Both of these features introduce extra degrees of freedom and are controversial. And third, and more importantly, these works restrict the estimation to the determinacy region of the parameter space, so they are not in a position to differentiate among competing explanations (e.g., misperceptions against indeterminacy) and select the one most consistent with the data.

Our analysis addresses all three problems. Relying on the empirical approach of L-S, we also estimate a version of the standard, rational expectations, New Keynesian model that includes misperceptions about the true state of the economy and learning. Learning is assumed to be rational (it is based on the Kalman filter). We find that the case of indeterminacy in the pre-Volcker period is overwhelming, independent of whether the alternative determinate model contains misperceptions.

We also evaluate the case for indeterminacy using model specifications that closely follow the Orphanides (2004) reasoning. In particular, we estimate a determinate version where the signals available to the agents/monetary authority are real-time inflation and the real-time output gap (the series used by Orphanides 2004) and another version where the monetary policy rule involves a reaction to these signals only. (6) The case for indeterminacy remains overwhelming independent of the competing determinant model. We consider these results as providing strong support to the CGG thesis. (7)

The rest of the paper is organized as follows. Section 1 describes the model. Section 2 presents the solution method and the estimation strategy. Section 3 discusses our results. A last section concludes.

1. THE MODEL

We consider the...