To Respond or Not to Respond : Measures of the Output Gap in Theory and in Practice

This paper analyzes the implications of responding to either the model-based New Keynesian output gap or to its estimates, and in particular, a Hodrick-Prescott filtered output gap or a linearly detrended output gap. Responding to these estimates instead of to the “true” unobserved output gap generates long-lasting business cycles and lower welfare. Furthermore, correlations between the estimates and the theoretical output gap depend on the stochastic structure of the shocks affecting the economy. In particular, productivity shocks generate a negative such correlation. Hence, the output gap estimates may provide poor guidance to monetary policy