Multiple Testing with Application to Mutual Fund Data

• Activity Number: 301 - Application of Asymmetric Classification and Multiple Testing
• Type: Invited
• Date/Time: Wednesday, August 11, 2021 : 3:30 PM to 5:20 PM
• Sponsor: WNAR
• Author(s): Xu Han* and Lijia Wang and Xin Tong
• Companies: Temple University and University of Southern California and University of Southern California
• Keywords: mutual fund; dependence; non normality; multiple testing; approximate empirical Bayes

Abstract

Mutual fund managers' talent is commonly evaluated by the intercept alphas of Carhart's (1997) four-factor model, and mutual funds with positive alphas are considered to be skilled. Since true alphas cannot be observed, selecting skilled mutual funds through multiple testing framework has received increasing attention from statisticians as well as nance researchers. After observing that the standardized OLS estimates of alphas across the funds possess strong dependence and non-normality structures, we propose a multiple testing procedure based on the posterior probabilities incorporating the information across all the funds in our study, which separates mutual funds into skilled, unskilled and non-selected funds. This testing procedure enjoys optimality and monotonicity. To model the distribution of the information used for the testing procedure, we consider a mixture model under dependence and propose a new method called "approximate empirical Bayes" to fit the parameters. Empirical studies show that our selected skilled funds have superior long term and short term performance compared to the unskilled and non-selected groups.