The Information Cycle and Return Seasonality

Abstract

Heston and Sadka (2008) document the return seasonality anomaly—that cross-sectional stock returns depend on their historical same calendar-month returns. We propose an information-cycle explanation for this anomaly, that firms’ seasonal information releases lead to higher returns in months with such information releases, and lower returns in months with no scheduled information releases. Using past earnings announcements and decreases in implied volatility as proxies for scheduled information events, we find indeed that event-month seasonal winners and nonevent-month seasonal losers indeed drive the seasonality anomaly. Hence, return seasonality can in fact be consistent with investors’ rational response to information uncertainty.