Maxing out short-term reversals in weekly stock returns

Abstract

Subrahmanyam (1991) presents a model in which increased variance in liquidity trades reduces price efficiency when market makers are risk-averse. Motivated by this theoretical insight, we hypothesize that pent-up demand from lottery-seeking investors amplifies their overreactions to news, leading to larger short-term return reversals. Consistent with this hypothesis, we identify a significant pattern in weekly U.S. stock returns for lottery-like stocks, defined by high recent maximum daily returns (MAX). Specifically, high-MAX stocks that were past 1-week losers (or winners) exhibit notably positive (or negative) returns in the following week. Applying a short-term reversal strategy to high-MAX stocks generates an average weekly return of 1.66%, significantly outperforming the 0.65% return from the same strategy applied to low-MAX stocks. This result remains robust even after controlling for market microstructure biases and survives a series of robustness tests. Interestingly, the MAX-enhanced reversal strategy proves effective only when retail order imbalance is in the highest quintile. This result holds across both value-weighted and equal-weighted portfolios, underscoring the pivotal role of retail investors. Taken together, our findings highlight a new channel through which retail investors’ preference for lottery-like payoffs amplifies their overreactions, enhancing the profitability of short-term reversal strategies.