This paper identifies the issue of “duplicate observations” in mutual fund performance analysis and introduces a novel wild bootstrap-based solution. Our method retains key database characteristics, such as fund entry/exit points (i.e., missing data) and cross-sectional information. We establish theoretical validity under unknown form of cross-sectional dependence and weak serial dependence. Simulations demonstrate the superiority of our approach. Furthermore, we propose a new method for identifying top-performing funds. Empirical results indicate that a small fraction of funds outperform the market.
社交媒体