Does Style-Shifting Activity Predict Performance? Evidence from Hybrid Mutual Funds

Abstract

This study applies an innovative return-based approach to determine the style-shifting activity of mutual funds. Based on daily returns, we measure style-shifting activity as inter-quarterly changes in the style exposures of a fund. In order to test the robustness of style-shifting activity we relate it to the popular activity measures “tracking error” and “R-squared”. In the main part of the paper, we compare the ability of these three activity measures to predict fund performance. Our empirical study covers 520 U.S. hybrid mutual funds from 10/1998 to 12/2009 and shows that i) on average smaller funds tend to be more active style shifters, ii) a high degree of style-shifting relates to higher expense and turnover ratios, and iii) current style-shifting activity alone is not a significant predictor of fund performance. However, iv) style-shifting activity relates positively and significantly to future performance for currently outperforming funds and v) performance persistence is strongest for the most active style shifters. Finally, we vi) orthogonalize tracking error and R-squared on style-shifting activity and find that additional activity detected by the two alternative measures is not positively associated with future fund performance.