Despite their strong performance, momentum strategies suffer from occasional large drawdowns referred to as momentum crashes. In this paper, we argue that momentum crashes are due to increase in anchoring-motivated demand on stocks far from their previous price peaks during the market rebounds. Consistent with our hypothesis, we find that nearness to 52-week high subsumes the predictive power of momentum measure during momentum crash periods. Furthermore, we provide a revised momentum strategy that is neutral on nearness to 52-week high. The strategy is free of crashes and exhibits a normal-like distribution without sacrificing its profitability
JEL Classification: G12, G14
Keywords: momentum, momentum crash, 52-week high, anchoring bias, investor sentiment, market state
