Academic journal article Journal of Money, Credit & Banking

On the Derivation of Monetary Policy Shocks: Should We Throw the VAR out with the Bath Water?

Academic journal article Journal of Money, Credit & Banking

On the Derivation of Monetary Policy Shocks: Should We Throw the VAR out with the Bath Water?

Article excerpt

THERE HAS BEEN A GREAT DEAL of recent interest in identifying monetary policy shocks and in estimating their effects on various macroeconomic variables. Christiano and Eichenbaum (1992) and Leeper and Gordon (1992), for example, focused on the "liquidity effects" of monetary policy shocks, the immediate reaction of economic variables to unexpected changes in the stance of monetary policy. More recently, Bernanke and Blinder (1992), Gordon and Leeper (1995), Strongin (1995), Christiano, Eichenbaum, and Evans (1996), Brunner (1994), and Bernanke and Mihov (1995) have explored alternative ways of identifying monetary policy shocks and have traced out the effects of these shocks on various sectors of the macroeconomy.

Importantly, much of this research was conducted using vector-autoregressive (VAR) models, a methodology popularized by Sims (1980) and used widely and extensively by economists to study the dynamic behavior of economic variables. The appeal of VAR models is likely due to several attractive features relative to other econometric modeling approaches, including a minimum number of identifying restrictions, relatively few exogenous variables, and ease of implementation. Still, the use of a VAR model requires a few strong assumptions about the availability of information to economic agents, some of which are also common to other more-overidentified econometric models. This paper considers an alternative approach that addresses some possible shortcomings of the VAR approach, while maintaining many of its appealing features.

The estimation of a structural VAR model generally requires two steps. First, a vector of economic variables (dated at time t) is regressed on several lags of itself. The set of lagged variables (dated t-1 and earlier) is assumed to be a good proxy for the information set that is available to economic agents at the beginning of period t. Thus, the residuals from these regressions are interpreted as innovations--new information about the economic variables that became available during period t. In the second step of estimation, the innovations are regressed on themselves, using one of several statistical procedures. The second-stage regressions are often given a structural or behavioral interpretation. (For example, innovations to the Federal Reserve's policy instrument are regressed on innovations to variables believed to be in the Fed's reaction function.) Thus, the residuals from the second-stage regressions are often viewed as structural shocks--the unexpected component of a behavioral relationship.

This paper is concerned with two implicit assumptions that are made in estimating the structural VAR model that may not accord well with reality, particularly for the case of deriving monetary policy shocks. First, there is the assumption that all economic agents, at the beginning of period t, have access to all economic data dated t-1 and earlier. In fact, many economic data are not publicly available until subsequent periods. This is especially problematic for the derivation of monetary policy shocks, since the Fed likely reacts in period t to new information about economic variables that describe economic activity in period t and are observable during period t and to new information about those variables that describe activity in earlier periods but become observable only in period t. Second, there is the assumption that the information set for all economic agents consists of only those variables used in the VAR model. In actuality, the correct information set probably contains many additional economic variables that are not used in the VAR analysis.

The main objectives of this paper are to present an alternative approach to the derivation of monetary policy shocks that addresses the above criticisms and to determine whether these concerns are important for the identification and analysis of monetary policy shocks. This approach follows naturally from Rudebusch (1996) and Sims (1996). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.