Stock Vulnerability and Resilience

Abstract

The authors propose a parsimonious yet flexible statistical method for predicting the relative vulnerability or resilience of individual stocks to market drawdowns. The authors’ approach compares a stock’s unique circumstances—as reflected in popular factor attributes—to the circumstances of stocks that have proven vulnerable or resilient to previous market drawdowns. Unlike other approaches, the authors’ method allows the influence of each factor attribute to vary across stocks in a nonlinear, conditional way. The authors test their explicit method for predicting stock vulnerability and resilience out of sample using the five largest market drawdowns since the global financial crisis. The nonlinear composite scores the authors derive are reliably better predictors of cross-sectional return than any of the individual factor attributes or an ex post linear combination of factor attributes.