On Monetary Policy, Model Uncertainty, and Credibility

This paper studies the design of optimal time-consistent monetary policy in an economy where the planner and a representative household are faced with model uncertainty: While they are able to construct and agree on a reference model (probability distribution) governing the evolution of the exogenous state of the economy, a representative household has fragile beliefs and is averse to model uncertainty. In such environments, management of households' expectations becomes an active channel of optimal policymaking per se. A central banker who respects the fact that private sector models are imperfect and designs her optimal policy accordingly may be able not only to mitigate a fundamental time-inconsistency problem but also to sustain higher welfare. Interestingly, in some cases the resulting welfare is even higher than in models where both a central banker and a representative household are assumed to know the true model, i.e., to have rational expectations.