Fund Analysis and Selection: An Approach Based on Distances and Similarities between Performance Measures

Abstract

In this paper, we propose to study the distances and similarities between key performance measures based on clustering data-mining techniques. Based on 15 performance measures related to 211 investment funds and calculated on three sub-periods - 6 months, 1 year and 3 years - clustering techniques enable the observation of interesting and surprising groupings. Major findings of this study include the following. First, each obtained cluster can be used to complete and improve classical fund performance analysis. Second, similarities between the traditional performance measures confirm the superiority of the Sharpe ratio. Third, distances between the main measures reveal the three main dimensions of performance formed by the Sharpe ratio, the Information ratio and Beta coefficients. From a practical perspective, these results confirm not only the relevance of a selective approach in choosing measures but also the need for an integrated approach to performance analysis and the selection of funds. To this end, we find that ranking based on a combination of Jensen’s alpha, the Sharpe ratio and Absolute Performance over a 3-year span dominate clearly all other rankings.

Keywords: mutual funds, fund selection, performance measurement, asset management

JEL Classification: C15, G11, G23