The Delisting Bias in Hedge Fund Databases

Abstract

As is well known, hedge fund databases suffer from various types of serious biases. While many of these biases have been addressed, the delisting bias is much more difficult to control. In this article, the authors use information from three hedge fund databases to provide direct estimates of this bias. Based on the fact that funds delisted in one database often continue to report returns to another, they estimate the delisting bias is at least 35 bps per annum. Their analysis also provides estimates of frequencies and average losses for different delisting reasons. The delisting bias largely explains the puzzling differences between the performance of the direct hedge fund investments and that implied by funds of hedge funds. The authors estimate that the performance of hedge fund indices should be adjusted downward by about 50 bps to account for the delisting bias.

TOPICS: Real assets/alternative investments/private equity, statistical methods, performance measurement