Improving Equity Fund Alpha Estimates with a Second Size Factor

Abstract

Practitioners and researchers seek to accurately estimate the value added by active equity fund managers. The authors hypothesize that the asset pricing models used to study equity funds may better capture nonlinearity in stock returns across market capitalizations by replacing the commonly used single size factor with two new size factors. This extension is important for explaining equity mutual fund returns because active fund holdings are weighted toward mid- and small-cap stocks to a greater extent than holdings of cap-weighted market indexes. In tests designed to minimize data mining issues, two size factors explain equity fund returns better than do the Fama–French single-size and style factors. Augmented Fama–French models explain over 25% of unexplained variance and yield superior adjusted R-squares for more than 75% of equity funds in the CRSP mutual fund database. Applied tests supplement these broad statistical analyses and confirm the technique’s value for practice. Also of benefit, the two proposed factor return series are readily available on the Internet to researchers and practitioners alike.