Ebook Monetary Policy, Inflation Expectations and the Price Puzzle
Structural vector auto regressions (SVARs) are widely used for measuring and understanding the effects of monetary policy innovations on the aggregate economy. While most results in the VAR literature are consistent with economic intuition and macroeconomic theory, the typically found positive reaction of the price level on impact to a monetary policy shock is a fact that most monetary models have difficulty explaining. This anomaly, first noted by Sims (1992) and labelled ‘the price puzzle’ by Eichenbaum (1992), casts serious doubts on the ability of correctly identifying a monetary policy shock. If the central bank monitors and responds to a larger information set than that of the VAR, what is referred to as a policy shock is actually a combination of a genuine policy shock and some endogenous policy reactions.
Sims (1992) argues that the central bank may have more information about future inflation than a simple VAR could adequately capture. The result of this omission is that a policy tightening in anticipation of future inflation would be incorrectly interpreted by the econometrician as a policy shock. As long as monetary policy only partially offsets inflationary pressures, the VAR would deliver a spurious correlation between a tightening of policy and a rise in inflation, namely the price puzzle. Sims (1992) observes that the inclusion of a commodity price index in the VAR appears to capture enough additional information about future inflation as to possibly solve the puzzle.
This paper offers a theoretically and empirically consistent explanation for the price puzzle using a small scale DSGE model and structural VARs. Earlier contributions have shown, using zero restrictions, that the price puzzle has been a distinctive feature of US data only before the appointment of Paul Volcker as Fed Chairman in 1979 (see Hanson, 2004). In Castelnuovo and Surico (2006), we confirm this result for a number of alternative, augmented VAR specifications. In this paper, we show that the price puzzle emerges in the pre-1979 period also when the monetary policy shock is identified using the sign restrictions implied by a standard sticky price model.
A number of contributions to the empirical literature on monetary policy have shown that a shift in the conduct of US monetary policy occurred in 1979 (Clarida, Galí, and Gertler, 2000, Boivin and Giannoni, 2006, Lubik and Schorfheide, 2004, Cogley and Sargent, 2005, among others). We therefore investigate the correlation between the empirical result of this literature about monetary policy and the empirical finding about the price puzzle. Using a sticky price model of the U.S. economy as data generating process, we show that structural VARs on artificial data, based on either zero restrictions or model-consistent sign restrictions, are capable of reproducing the price puzzle only when the central bank does not raise the interest rate sufficiently in response to inflation. The DSGE model, in contrast, does not generate, on impact, a positive response of the price level to a monetary policy shock, not even when monetary policy is passive. A contribution of the paper is to show that the price puzzle can actually be a spurious correlation induced by the omission in the VAR of a variable capturing the persistence of expected inflation, which is remarkably higher under the passive regime. And, the policy induced omitted variable problem is found to account quantitatively for the apparently puzzling response of inflation to a policy shock observed on actual data. Interestingly, our results show that the arguments in Sims (1992) are supported in the context of a structural model only when monetary policy is passive and thus multiple equilibria arise. To the best of our knowledge, this paper is the first attempt of rationalizing the price puzzle using a structural model.
The paper is organized as follows. Section 2 presents a re-examination of the empirical evidence using estimated SVARs in output, inflation and the nominal interest rate. The following part describes the sticky price model used for the theoretical investigation. In Section 4, the dynamic responses of the theoretical model to a monetary policy shock are compared to the impulse responses of the structural VARs estimated on artificial data. The latter are shown to be systematically above the former under indeterminacy only, and to reproduce the sign and magnitude of the price puzzle observed in the pre-1979 period. Section 5 offers a new interpretation of the price puzzle and shows that augmenting the SVAR on actual data with the Federal Reserve’s inflation forecasts reduces significantly the omitted variable problem that would otherwise emerge.
Download
PDF Ebook Monetary Policy, Inflation Expectations and the Price Puzzle
Posted in :
