Optimal policy under model uncertainty: A structural-bayesian estimation approach

In this paper we propose a novel methodology to analyze optimal policies under model uncertainty in micro-founded macroeconomic models. As an application we assess the relevant sources of uncertainty for the optimal conduct of monetary policy within (parameter uncertainty) and across models (specification uncertainty) using EU 13 data. Parameter uncertainty matters only if the zero bound on interest rates is explicitly taken into account. In any case, optimal monetary policy is highly sensitive with respect to specification uncertainty implying substantial welfare gains of a robust-optimal rule that incorporates this risk.