The Other January Effect: International, Style, and Subperiod Evidence

Abstract

We report international, style, and subperiod evidence for the Other January Effect (OJE) documented in Cooper et al. (2006). When examining the OJE in 22 countries starting as early as 1801, we find that the spread between 11-month returns following positive and negative Januarys does tend to be positive. However, the spreads are rarely statistically significant and other calendar months' returns exhibit similar subsequent return spreads. Further, when first controlling for a country's OJE tied to the U.S. January return, there is essentially no incremental OJE relation tied to the own-country's January return. Next, for U.S. industry-level portfolios, we find that there is no incremental OJE spread tied to the own-industry January returns after first controlling for the market-level January returns. For U.S. size and book-to-market portfolios, the market-level January returns are also associated with larger spreads than the own-portfolio January returns. Finally, for all the U.S. portfolios and for the major international markets, the OJE is appreciably weaker over the 1975 to 2006 post-discovery period than over the 1940 to 1974 pre-discovery period. Our evidence indicates that the OJE is primarily a U.S. market-level-based phenomenon that has diminished over time, which suggests a 'temporary anomaly' interpretation.

Keywords: Return Predictability

JEL Classification: G12, G14