Optimal Monetary Policy with Uncertainty Aversion

This paper studies optimal monetary policy when a central bank faces uncertainty about the production technology. Our model contains nominal wage and price rigidity and is sufficiently rich to capture the empirical response of aggregate variables to a monetary shock. In this economy, introducing uncertainty of the form developed in Hansen and Sargent (2003) attenuates the central bank’s optimal response to exogenous shocks. This constitutes an empirical success relative to existing studies in which optimal policy, with or without uncertainty aversion, is more aggressive towards inflation relative to policy observed in U.S. data. To frame our findings within existing research, we present a simple model which demonstrates how adding uncertainty, depending upon the specific formulation, can lead to more or less aggressive optimal reactions to shocks. We develop a ‘non-attenuation principle’ to explain why uncertainty averse decision makers typically amplify their response to shocks as uncertainty increases. We use this principle to interpret existing findings in which robustness concerns amplify policy, including Sargent (1999) and Stock (1999). Also, we distinguish the concept of robustness to misspecification– which concerns Knightian uncertainty–from the well-known Brainard result—which concerns risk.