Does style-shifting activity predict performance? Evidence from equity mutual funds

This study introduces an innovative approach to measuring the "style-shifting activity" (SSA) of mutual funds using daily returns. Applying our new measure to a comprehensive sample of 2631 active US equity mutual funds, we show (i) that SSA predicts future performance, especially for current outperformers, and (ii) that SSA adds new information previously not captured by alternative return-based activity measures such as tracking error or R-squared. Comparing the three measures, we show that SSA captures activity very selectively, which makes it a stable and reliable predictor of future performance. Tracking error and R-squared, however, seem to additionally capture some unobserved fund characteristics, as the direction and power of their predictions depend heavily on the consideration of time- and fund-fixed effects. Moreover, investment strategies based on past SSA and past performance earn up to 2.4% (3.6%) p.a. risk-adjusted net (gross) returns which is economically and statistically significant. (C) 2015 The Board of Trustees of the University of Illinois. Published by Elsevier B.V. All rights reserved.