A1182
Title: Heterogeneous predictability on mutual fund alphas:A sparse clustering GMM approach
Authors: Liyuan Cui - City University of Hong Kong (Hong Kong) [presenting]
Guanhao Feng - City University of Hong Kong (Hong Kong)
Jiangshan Yang - City University of Hong Kong (Hong Kong)
Abstract: Mutual fund managers' skills, measured by risk-adjusted alphas, are predictable using fund characteristics identified via machine learning. However, alphas predictive power varies across funds and periods, with most funds exhibiting negligible alphas. To model the heterogeneous predictability, Sparse Clustering GMM (SCGMM), a nonparametric approach to uncover latent fund group structures, is introduced. SCGMM clusters funds using estimated alphas and identifies group-specific parameters tied to market predictors. The method accounts for heterogeneity in cross-sectional grouping and time variation mechanisms across predictors, without requiring prior cluster information or time variation specifications. The proposed estimator is theoretically grounded, ensuring consistency in both the grouping and estimation of time-varying parameters. Empirical analysis of U.S. data shows that only a small group of mutual funds exhibit alphas predictably driven by market predictors.
