This article proposes a theoretical explanation for the positive pre-announcement drift empirically documented ahead of scheduled announcements, using the Federal Open Market Committee (FOMC) meetings as a main example. The framework entails a general equilibrium model of disagreement (differences of opinion), where investors interpret a costly signal differently. Investors optimally decide to stop learning when an announcement is imminent, increasing the risk premium ahead of an announcement. The model jointly rationalizes puzzling empirical evidence by generating i) an upward drift in prices just before scheduled announcements, regardless of the announcement’s content, which coexists with ii) low volatility and iii) low trading volume.