A Low-volatility Factor for Standard Models of Stock Returns?

Given the body of research on the outperformance of low-risk stocks, should the equity asset pricing community add a low-volatility factor in standard models of stock returns? In their June 2025 paper entitled “Factoring in the Low-Volatility Factor”, Amar Soebhag, Guido Baltussen and Pim van Vliet investigate adding a low-volatility factor to standard models via four scenarios:

1. Gross (frictionless) returns for long-minus-short portfolios for all factors as conventionally done in prior factor model research.
2. Gross returns for market-hedged long and short legs as separate aspects of all factors.
3. Net returns approximated from estimated bid-ask spreads and shorting
   fees for separate market-hedged long and short legs of all factors.
4. Net returns for only the long legs of all factors.

They compute stock volatilities based on a rolling window of 252 trading days for low-volatility factor calculations. They compare models by weighting their respective factors at each rebalancing to achieve maximum test period Sharpe ratio. Using firm/stock data for U.S. common stocks with positive book-to-market ratios to construct long-minus-short and long or short factor returns for well-known asset pricing models, and estimated trading frictions and shorting costs, during January 1970 through December 2023, they find that:

(more…)