Mean-reverting expected returns and seasonality in the contrarian investment strategy

This study examines the excess returns, from a contrarian strategy, which De Bondt and Thaler (1985) attribute to investor overreaction. As in Chan (1988), evidence is found which indicates that when the time-variation of portfolio beta is considered, the returns of the contrarian strategy are not excessive, as defined by the Sharpe-Lintner CAPM. However, this study determines that the mean-reverting behavior of expected returns causes the strategy to earn profits from successful market timing. In addition, evidence is found that the CAPM beta is a better measure of risk than is firm size. The study also shows that the excess returns reported by De Bondt and Thaler are realized only in January. This finding introduces the possibility that the excess returns are due to something other than risk compensation. Strong support is found for tax-loss selling as an explanation of the January returns