This paper investigates hedge fund idiosyncratic returns from January 2000 to December 2010. Even after filtering for the exposure to standard risk factors, idiosyncratic returns are still correlated across hedge funds. This raises the question about the potential channels of the similarities in the filtered returns. I conjecture that the presence of the same prime broker, the same domicile or the same style may induce these similarities. Exploiting these hypotheses, I find that dealing with the same prime broker is significantly related to the similarity of idiosyncratic returns. These results are robust to  the use of alternative risk factors, subsamples, as well as to alternative similarity measures.

Discussant: Wenqian Huang (VU Amsterdam)