Stock Price Jumps and Cross-Sectional Return Predictability
George J. Jiang, University of Arizona
Tong Yao, University of Iowa

In rational continuous-time asset pricing models, compensation for risk is represented by the continuous drift of asset prices, whereas jumps are the effect of large information or liquidity shocks. The authors use this distinction to evaluate risk-based explanations of cross-sectional stock return predictability. Based on the CRSP data from 1927 to 2005, they find that individual stock price jumps tend to be idiosyncratic and predominantly positive, presenting an interesting contrast to mostly negative jumps in market portfolios. More importantly, several well-known patterns of return predictability, including the size effect, the liquidity premium, and to a moderate extent the value premium, are the result of cross-sectional differences in stock price jumps. The evidence presents a challenge to theories that attribute such return predictability to simple differences in risk premium. The authors further explore several alternative hypotheses within the rational asset pricing paradigm, such as the martingale restriction on jumps, jump risk premium effect, investor preference for skewness, and discontinuity in expected returns. However, none of them can be reconciled with the empirical evidence.

Extreme volatility affects many investment strategies, and most particularly quantitative strategies using options. For many of our members, discrete price jumps represent either their greatest risks or their greatest opportunities. The results in this paper help better identify which securities are most subject to these jumps.