This paper estimates a dynamic stochastic equilibrium model in which agents use a Bayesian rule to learn about the state of monetary policy. Monetary policy follows a nominal interest rate rule that is subject to regime shifts. The following results are obtained. First, the author's policy regime estimates are consistent with the popular view that policy was marked by a shift to a high-inflation regime in the early 1970s, which ended with Volcker's stabilization policy at the beginning of the 1980s. Second, while Bayesian posterior odds favor the "full-information" version of the model in which agents know the policy regime, the fall of inflation and interest rates in the disinflation episode in the early 1980s is better captured by the delayed response of the "learning" specification. Third, the author examines the magnitude of the expectations-formation effect of monetary policy interventions in the "learning" specification by comparing impulse responses to a version of the model in which agents ignore the information contained in current and past monetary policy shocks about the likelihood of a regime shift.