The prediction of fund failure through performance diagnostics

Using an international database featuring 1624 mutual funds over 15 years, this paper analyses the joint abilities of performance measures to predict subsequent fund failure. We examine the probability of disappearance over a time window, and expected fund survival time, and study the circumstances of a fund's disappearance, its currency and domicile. By combining relevant measures, fund failure appears to a significant extent predictable, more than with single classical measures. Survivorship predictability has significant economic value. Such evidence suggests that past performance does not only influence investors' perception of fund quality, but also reflects managers' ability to sustain performance.