A highly controversial issue in financial economies is whether stocks overreact. In this paper we find an economically-important overreaction effect even after adjusting for size and beta. In portfolios formed on the basis of prior five-year returns, extreme prior losers outperform extreme prior winners by 5–10% per year during the subsequent five years. Although we find a pronounced January seasonal, our evidence suggests that the overreaction effect is distinct from tax-loss selling effects. Interestingly, the overreaction effect is substantially stronger for smaller firms than for larger firms. Returns consistent with the overeaction hypothesis are also observed for short windows around quarterly earnings announcements.