By Roger Ibbotson, Peng Chen

Abstract

In this paper, we focus on two issues. First, we analyze the potential biases in reported hedge fund returns, in particular survivorship bias and backfill bias, and attempt to create an unbiased return sample. Second, we decompose these returns into their three A,B,C components: the value added by hedge funds (alphas), the systematic market exposures (betas), and the hedge fund fees (costs). We analyze the performance of a universe of about 3,500 hedge funds from the TASS database from January 1995 through April 2006. Our results indicate that both survivorship and backfill biases are potentially serious problems. The equally weighted performance of the funds that existed at the end of the sample period had a compound annual return of 16.45% net of fees. Including dead funds reduced this return to 13.62%. Excluding backfill further reduced the return to 8.98%, net of fees. In this last sample, we estimate a pre-fee return of 12.72%, which we split into a fee (3.74%), an alpha (3.04%), and a beta return (5.94%). Overall, even after correcting for data biases, we find that the alphas are significantly positive and are approximately equal to the fees, meaning that excess returns were shared roughly equally between hedge fund managers and their investors.