A Horizon Based Decomposition of Mutual Fund Performance Using Transaction Data
49 Pages Posted: 18 Nov 2019
Date Written: November 7, 2019
This paper uses transaction-level data to decompose active mutual fund performance based on the past length of funds' holdings. We find that the majority of value added is related to holdings that have been in the portfolio for more than one year. Our decomposition is markedly different depending on fund size: on average trades of large (small) funds start to add value only after 7 months (10 days), arguably because of the large price impact that large funds have. Only small funds profit from short-term anomalies suggesting that funds actively adapt their strategies to the price impact of their trades and specialize in different investment horizons. Finally, the dispersion of fund turnover is large and persistent. Trades and holdings of high-turnover funds add a substantial amount of value within a year but destroy value beyond a year, whereas trades and holdings of low-turnover funds mainly add value beyond a year.
Keywords: mutual funds, investment horizon, transaction data, daily holdings, skill
JEL Classification: G23
Suggested Citation: Suggested Citation