Volume 15, No. 1, 2017
Joshua Livnat, Gavin Smith and Martin Tarlie
One of the most crucial decisions for investors and plan sponsors is the selection of funds among the thousands of available alternatives.We find that regardless of the initial criterion used to rank funds based on past performance, more diversified top funds outperform concentrated top funds in the subsequent year. This better performance is attributed to the more consistent returns of funds with diversified holdings. We also find that when the initial criterion is based on a manager’s skill, as measured by a positive intercept in a regression of past fund returns on the five Fama–French factors, active share is a useful tool to predict future winner funds among top skill managers. However, diversified top funds provide slightly higher returns and less severe drawdowns than funds with high active share and top manager skill.
To account for the problem that high information ratio can be associated with low return but even lower tracking error, we introduce the Modified Information Ratio (IR) measure. This measure adjusts the conventional IR to account for an investor’s desired alpha. The Modified IR measure and conventional IR behave similarly with respect to diversification. We find that top Modified IR funds that are also diversified—winner funds—have significantly
better future 12-month returns than top Modified IR funds that are concentrated.