Hedge fund performance evaluation is essential for investors to determine which funds create value (or alphas) and allocate funds efficiently. However, this analysis is challenging because investors must choose among the large number of asset pricing models proposed to explain hedge fund returns (see, e.g., Agarwal, Mullarly, and Naik (2015) for a review).

In this project, we formally compare the performance of hedge funds across several models extensively used in previous work. We develop a novel approach to estimate the entire cross-sectional distribution of hedge fund alphas for each tested model. Comparing these distributions determines how performance evaluation varies across models. Our analysis also allows us to examine whether the estimated alphas produced by different models truly capture managerial skill or simply reflect the systematic risks taken by the funds. After explaining our econometric contributions, we will conduct an extensive empirical analysis using a large universe of hedge funds for a period ranging from January 1994 to December 2015.