Abstract:
This paper considers an alternative econometric approach to the VAR methodology for identifying
and estimating the effects of monetary policy shocks. The alternative approach incorporates available
measures of market participants' expectations of economic variables in order to calculate economic
innovations to those variables. In general, expectations measures should provide important additional
information relative to a standard VAR analysis, since market participants presumably use a much
richer information set than that assumed in a typical VAR model. The resulting innovations are easily
incorporated in a VAR-like framework.
The empirical results are quite surprising. First, when expectations are incorporated, the variance of
all innovations is reduced substantially. Second, innovations to the federal funds rate derived using
the alternative approach are only somewhat correlated with their VAR counterparts, while innovations
to other economic variables are essentially uncorrelated. Still, monetary policy shocks derived using
both approaches are still somewhat correlated, however, since innovations to prices and economic
activity explain only a small fraction of innovations to the federal funds rate. As a consequence, the
impulse responses of economic variables to the two sets of monetary policy shocks
have remarkably similar properties.
|
