A Note on Sorting Bias Correction in Regression-Based Mutual Fund Tournament Tests

Abstract

The tournament hypothesis of Brown et al. (1996) conjectures that mutual funds with a below average performance over the first half of the year tend to increase their risk in the second half of the year. Schwarz (2012) argues that the methodologies that have been used to test this hypothesis are flawed because they are affected by a bias that results from sorting on return, which likely also sorts on risk. He argues that both the contingency and regression approaches that have been used in the literature are affected by this sorting bias. We demonstrate that simply including the standard deviation over the first half of the year in regression-based tests corrects for most of the bias and is as suitable to control for the sorting bias than the more complex Schwarz (2012) correction.