This paper characterizes endogenous monetary policy when policymakers are uncertain about the extent to which movements in output and inflation are due to changes in potential output or to cyclical demand and cost shocks. We refer to this informational limitation as the "information problem" (IP). Main results of the paper are: 1. Policy is likely to be excessively loose (restrictive) for some time when there is a large decrease (increase) in potential output in comparison to a full information benchmark. This provides a partial but unified explanation for the inflation of the seventies and the price stability of the nineties. 2. Errors in forecasting potential output and the output gap are generally serially correlated. 3. A quantitative assessment, based on an empirical model of the US economy developed by Rudebusch and Svensson (1999) indicates that, during and following periods of large changes in potential output, the IP significantly affects the dynamics of inflation and output. 4. The increase in the Fed's conservativeness between the seventies and the nineties, and a more realistic appreciation of the uncertainties surrounding potential output in the second period, imply that the IP problem had a stronger impact in the seventies than in the nineties.