In measuring performance persistence, we use hedge fund style benchmarks. This allows us to identify managers with valuable skills, and also to control for option‐like features inherent in returns from hedge fund strategies. We take into account the possibility that reported asset values may be based on stale prices. We develop a statistical model that relates a hedge fund's performance to its decision to liquidate or close in order to infer the performance of a hedge fund that left the database. Although we find significant performance persistence among superior funds, we find little evidence of persistence among inferior funds.

We reconsider whether hedge funds’ time-varying risk factor exposures are predictive of superior performance. We construct an overall measure (BA) of fund managers and present evidence that top beta active managers deliver superior long-term out-of-sample performance compared to top alpha active managers. BA captures the time-varying nature of beta exposures and can be interpreted as a common factor of both systematic risk (SR) and (1 - R 2 ) measures. BA also compares favorably to extant measures of market timing, capturing the explanatory power of such measures of hedge fund performance.