Optimal Policy Restrictions on Observable Outcomes

Abstract

We study the restrictions implied by optimal policy DSGE models for the volatility of observable endogenous variables. Our approach uses a parametric family of singular models to discriminate which volatility sample outcomes have zero probability of being generated by an optimal policy. Thus the set of volatility outcomes generated by the model is not of measure zero even if there are no random deviations from optimal policymaking. This methodology is applied to a new Keynesian business cycle model widely used in the optimal monetary policy literature, and its implications for the assessment of US monetary policy performance over the 1984-2005 period are discussed.