Optimal disinflation under learning

We model transitional dynamics that emerge after the adoption of a new monetary policy rule. We assume that private agents learn about the new policy via Bayesian updating, and we study how learning affects the nature of the transition and the choice of a new rule. Temporarily explosive dynamics can emerge when there is substantial disagreement between actual and perceived policies. These dynamics make the transition highly volatile and dominate expected loss. The emergence of temporarily explosive paths depends more on uncertainty about policy-feedback parameters than about the long-run inflation target. For that reason, the central bank can at least achieve low average inflation. Its ability to move feedback parameters away from initial beliefs, however, is more constrained.