Momentum Crash Management

Abstract

Momentum is the largest and most pervasive market anomaly. However, despite high mean and Sharpe ratio, momentum suffers from large negative skewness that comes from momentum crash periods. These crashes occur in times of market stress and market rebound, thus the variables that capture these episodes, can be used as momentum predictor. I introduce three new momentum predictors and show that changes in momentum volatility have higher predictive power than the other predictors used in previous studies. Once momentum prediction has been proved, the predictors can be employed in momentum risk management. I introduce a new method of momentum risk management that has a lower transaction cost than those methods used in previous studies, both in terms of turnover and price impact.

Keywords: Momentum, crashes, risk management, skewness, transaction cost

JEL Classification: G11, G12, G14, G17