What the 'Indexes' Don't Tell You about Hedge Funds

Abstract

In recent years, the number of firms offering hedge fund products, and assets allocated to hedge funds have grown at a rapid rate. This growth has been accompanied by a commensurate increase in the number of hedge fund research articles. A striking limitation of many analyses published to date is their focus on so-called indices, which in reality reflect the average returns of peer managers, categorized by investment style. We focus instead on individual manager return histories and investigate the relationship between individual manager performance and investable equity or fixed income market indexes. Our goal is to obtain reliable estimates of hedge funds' sensitivity to market returns, and to get a sense of the variability of outcomes across individual hedge fund managers of a given style. Using manager-level performance data from the Hedge Fund Research database, we refine and apply methods suggested in previous studies. Our approach has five notable features: - We correct for possible distortions in returns introduced by performance-based fees by using inferred gross-of-fees returns. - We regress individual manager returns on contemporaneous and lagged returns of broad market equity or fixed income indexes, as appropriate for the style of the manager being analyzed. - We correct for autocorrelation of residuals when appropriate. - We modify our regression models to detect differences in manager sensitivity to index returns in bull and bear markets. - We model participation-linked fees as call options and evaluate their impact on performance in a separate step. Our principal finding is that substantial variation in market-relative risk exists across managers within hedge fund style universes. Many of the style-level effects reported by other researchers were not observed in a majority of individual managers. This result suggests that informed manager selection helps investors avoid the unattractive performance characteristics identified by style-level analyses. Further, we find that the dramatic asymmetry noted by previous studies is largely a function of the volatility induced by the 1998 Russian Debt Crisis rather than some inherent asymmetry in hedge fund return patterns. For practitioners, the upshot is that designing a hedge fund strategy is an intensively bottom-up exercise. Generalizations at the style level are of little use when formulating a hedge fund strategy in the real world.