Re-Evaluation of the Low-Risk Anomaly in Finance via Matching

Abstract

The long-term success of low-risk stocks over high-risk stocks runs contrary to the basic finance principle that risk is compensated with higher expected returns. Our paper examines this low-risk anomaly using Coarsened Exact Matching to balance high and low-risk stock portfolios on industry, company size, and trading volume. After matching, we find that the low-risk anomaly still exists but has a more muted effect than in previous studies, specially when beta is used as a measure of risk. We also find moments in which the low-risk anomaly does not hold, most notably during the dot-com bubble. To our knowledge, we are the first to apply matching techniques to the study of the low-risk anomaly, and our findings complicate various previous explanations of this phenomenon.

Keywords: low volatility, beta, market efficiency, matching, capital asset pricing model

JEL Classification: C19, G12