Momentum and reversal: information from prior returns

Abstract

This paper studies the joint dynamics of momentum and reversal strategies. Momentum investors face uncertainty about whether past patterns of price movements will continue (momentum) or turn over (reversal), thereby increasing volatility of momentum returns and occasionally leading to momentum crashes. We argue that the forces driving reversal over momentum tend to be strong if losers’ past returns are extremely low (in the time-series) or if losers are small and illiquid (in the cross-section). We subsequently propose new risk-managed momentum strategies by taking into account behavioural divergence between momentum and reversal in an effort to boost momentum profits and reduce volatility. Empirical tests for the U.S. stock market in the sample period of 1947 to 2015 document that momentum strategies in which investors implement stop-trading rules if losers’ past returns are extremely low as well as buy-small-loser rules substantially outperform traditional momentum strategies. International stock markets and robustness checks confirm our U.S. tests. Importantly, we find that outperformance is mainly attributable to the increase in abnormal returns (i.e. alpha) from various factor models.