Narrative and VAR approaches to monetary policy: Common identification problems

Romer and Romer Romer, C.D., Romer, D.H. 1989. Does monetary policy matter? A new test in the spirit of Friedman and Schwartz. In: Blanchard, O.J., Fischer, S. (Eds.), NBER Macroeconomics Annual 1989. MIT Press, Cambridge, MA, pp. 121–170; Romer, C.D., Romer, D.H., 1994. Monetary policy matters. Journal of Monetary Economics 34, 75–88] adopted a narrative approach to address the identification problems in time series models of monetary policy. Based on Federal Reserve documents, the Romers created a dummy variable equal to one in periods when the Federal Reserve contracted in response to perceived inflationary pressures. This paper shows: (1) the dummy variable is predictable from past macroeconomic variables, reflecting the endogenous response of policy to the economy; (2) unpredictable changes in the dummy do not generate dynamic responses that look like the effects of monetary policy. The identification problems that plague time series models also afflict the narrative approach.