Is It Overreaction? The Performance of Value and Momentum Strategies at Long Horizons

Abstract

Recent studies on momentum effects document that momentum profits revert at long horizons and interpret this as support for delayed overreaction theories of momentum. In this paper I show that these long horizon reversals are a book-to-market effect. They disappear after adjusting returns for book-to-market and taking into account the predictable changes in book-to-market experienced by extreme performers following portfolio formation. The source of momentum profits is therefore closely linked to the nature of the book-to-market effect. If book-to-market captures mispricing, the findings are still consistent with the delayed overreaction explanation. However, if it captures risk, then momentum can be an underreaction phenomenon, despite reversals in unadjusted returns. I also examine long horizon returns of value strategies and find a strong decay in the value-growth return spread. This is consistent with overreaction explanations of book-to-market effects. For risk-based explanations the result implies that firm-level expected returns have a strong tendency to revert to the cross-sectional mean. In addition to CRSP/COMPUSTAT the empirical analysis employs a large new UK data set that features balance sheet information for virtually all firms listed on the London Stock Exchange since 1955. In contrast to COMPUSTAT this data set is free of survivorship bias.