Abstract
Few propositions in macroeconomics are less controversial than long-run money neutrality, yet clear and robust empirical support has not been found in time series studies. Bernanke and Mihov (1998) are comparatively successful in this hunt, but their output response to monetary policy shocks remains stubbornly persistent. This paper argues that the omission of a measure of output gap from the VAR estimated by Bernanke and Mihov lies at the heart of this "excessive" persistence. In the theoretical framework of a New Keynesian model similar to that of Svensson (1997) and Clarida, Gali and Gertler (1999), I prove that this omission induces persistence overestimation under relatively mild assumptions. The inclusion of a proxy for the output gap in the VAR is then shown to drastically increase the evidence for long-run money neutrality on US data, as predicted by the theoretical analysis.