What do the VARs mean? Measuring the output effects of monetary policy

Abstract

VARs describe the history of output and other variables following monetary shocks. To measure the effects of monetary shocks, one must add economic identifying assumptions. I specify the relative effects of anticipated and unanticipated money, and I calculate how VAR-based measures of the effect of money on output change as one varies this assumption. The anticipated/unanticipated assumption influences measured output effects as much or more than the variable selection and shock orthogonalization assumptions on which the VAR literature focuses. Assuming that anticipated monetary policy can have some effect on output results in much shorter, smaller, and perhaps more believable estimates of the output response to monetary shocks.