Model uncertainty has the potential to change importantly how monetary policy should be conducted, making it an issue that central banks cannot ignore. In this paper, I use a standard new Keynesian business cycle model to analyze the behavior of a central bank that conducts policy with discretion while fearing that its model is misspecified. My main results are as follows. First, policy performance can be improved if the discretionary central bank implements a robust policy. This important result is obtained because the central bank’s desire for robustness directs it to assertively stabilize inflation, thereby mitigating the stabilization bias associated with discretionary policymaking. In effect, a fear of model uncertainty can act similarly to a commitment mechanism. Second, exploiting the connection between robust control and uncertainty aversion, I show that the central bank’s fear of model misspecification leads it to forecast future outcomes under the belief that inflation (in particular) will be persistent and have large unconditional variance, raising the probability of extreme outcomes. Private agents, however, anticipating the policy response, make decisions under the belief that in‡ation will be more closely stabilized, that is, more tightly distributed, than under rational expectations. Third, as a technical contribution, I show how to solve an important class of linear-quadratic robust Markov-perfect Stackelberg problems.