joint avec le Séminaire de Statistique et d’Econométrie
René Garcia (EDHEC) & Caio Almeida (Graduate School of Economics, Getulio Vargas Foundation)
We evaluate the performance of hedge funds through a new nonlinear risk adjust- ment of returns. The risk adjustment is such that it prices exactly the usual set of risk factors considered in the hedge fund literature. This nonlinear risk adjustment goes beyond the usual linear regression methodology used in many hedge fund per- formance papers, including nonlinear exposures based on option-like features. The approach proposed in this paper overcomes two important limitations of the linear methodology : it captures the nonlinear exposure of a hedge fund strategy to several risk factors, and it is not limited to nonlinear shapes resembling standard option payoff patterns. We apply this methodology to various hedge fund indices as well as individual hedge funds, considering a set of risk factors including equities, bonds, credit, currencies and commodities. The main message that emerges from our anal- ysis on the performance of hedge fund strategies is that exposure to higher-moment risks on the various factors matters. Analyzing the performance of HFRI indices on primary strategies and sub-classes of primary strategies, we report sizable differences in performance between the linear and the nonlinear risk adjustment. Most often the nonlinear risk adjustment reduces the performance but for some sub-classes it en- hances their performance. We also show how to conduct a risk analysis and provide an example where a change in a single risk factor can affect the average performance of funds when more robust risk adjustment is applied.