The Limits to Arbitrage Revisited: The Low-Risk Anomaly

Abstract

We show that over a long study period (1963-2010), the existence and trading efficacy of the well-known low-volatility stock anomaly are more limited than widely believed. For example, we find that the anomalous returns are not found within equal weighted long-short (low minus high risk) portfolios. Within value-weighted portfolios, the existence of alpha is largely eliminated when omitting low priced (less than $5) stocks. Furthermore, extracting any excess returns associated with a long-short portfolio is meaningfully hampered by high transaction costs reflecting the finding that the abnormal returns are concentrated among low liquidity and smaller stocks. Adding to the challenge, the anomalous excess returns quickly reverse requiring traders to rebalance frequently in attempting to extract profits, thus amplifying liquidity needs. Our findings are unchanged for various approaches to measuring the low-volatility anomaly.