Do Mutual Funds Deliver Alpha? A Bayesian and Bootstrap Analysis

We apply a new bootstrap statistical technique to distinguish between dasiaskillpsila and dasialuckpsila for individual funds. This methodology allows for non-normality in the idiosyncratic risk of the funds-a major issue when considering those funds which appear to be either very good or very bad performers, since these are the funds which investors are primarily interested in identifying. We find that the best funds performance cannot be explained by luck, there exists stock picking ability among a relatively small number of top performing mutual funds. Moreover, we show that Bayesian measures, which help overcome the short-sample problem inherent in mutual fund returns, lead to superior performance predictability.