The overreaction hypothesis, firm size, and stock market seasonality

Abstract

curities suffering abnormally low returns (“losers”) will subsequently experience abnormally high returns. Further, securities with abnormally high returns (“winners”) will go on to earn abnormally low returns. Such a pattern, of course, is not consistent with even the weakest form of market efficiency. Even so, De I3ondt and Thaler [1987] find that this “winnerloser” pattern exists in security returns and that it appears to concentrate in January. We wish to take a closer*look at the winnerloser pattern and investigate its relationship to firm size and stock market seasonality. Our results indicate that the predicted reversal of gains and losses does not occur universally. Losers do have a pronounced tendency to become winners, but the reverse is not true. Firm size, however, influences the results, and large-firm returns are more consistent with overreaction. Most of the winner-loser pattern does arise in January. We find further that the effect occurs mostly at the turn of the year, but that it is not explainable by tax-loss selling. Strangely, when we examine behavior by day of the week, we find that Monday and Friday exhibit a definite winner-loser pattern in opposite directions. BACKGROUND