To quantify the impact of luck on mutual fund performance, we use the False Discovery Rate (F DR) introduced by Benjamini and Hochberg (1995) in the statistical literature. The F DR measures the proportion of lucky funds among the funds with significant estimated alphas. We extend this methodology by developing a new approach which allows us to separately compute the F DR among funds with significant positive estimated alphas (called hereafter the best funds) and funds with significant negative estimated alphas (called hereafter the worst funds)
