Monetary Policy with Unobserved Potential Output

This paper studies some effects of real time information for the implementation of monetary policy. We consider an economy in which several sources of uncertainty, such as measurement errors and imperfectly observable states, do not allow the policy maker to identify the true state of the economy. Optimal policy thus requires the policy maker to jointly solve a filtering and an optimization problem. We focus in particular on the case of a non-observable measure of potential output and analyze the consequences of this assumption for the macroeconomy (policy, output and inflation). The paper provides a benchmark model to assess the claim that conditioning policy on a potential output measure using real time data may be at the root of a biased policy stance, as recently suggested by Orphanides. More generally, it offers a rigorous framework to analyze the effects of imperfect information and to assess the role of macroeconomic indicators in alleviating information problems. JEL Classification Numbers: E5