Familial support in the mutual fund performance competition

Abstract

Numerous papers have presented evidence of mutual fund families engaging in self-interested behaviour, regardless of or contrary to the interests of its investors. Such findings have provoked industry backlash. Given the significance of the mutual fund industry in managing wealth (approx. US$31 trillion of assets under management as at December 2014), governance and agency issues are highly consequential. We contribute to this debate in two ways. First, we adopt a novel approach to identifying potential intra-family cross-subsidisation by examining its potential effects on the aggregate distribution of fund returns. After presenting our argument that mutual fund families are incentivized to concentrate cross-subsidies to constituent funds with high year-to-date returns in the fourth quarter, we test predictions arising from this hypothesis. In our linear regressions, the dependent variable are the difference in trailing-twelve-month returns of (1) the 1st and 2nd-ranked fund in each style by trailing-twelve-month returns, and (2) the 1st and 2nd-ranked funds by returns during first nine months of each twelve-month period. In our logistic regression, our dependent variable indicates whether the top-ranked fund by year-to-date returns within a style stays topranked at the end of the subsequent quarter. Second, we study whether fund families utilize internal trades to cross-subsidize their “high-value” member funds (i.e., outperforming or high-fee funds). Specifically, we apply Gaspar, Massa and Matos' (2006) empirical analysis with an updated dataset and a modified methodology, and test whether “opposite trade” between a high-value fund and a low-value fund increase the difference in net-of-style returns of two funds. In particular, we introduce a new "opposite trade" measure that varies linearly with the difference in the paired funds’ returns and, thus, fits our linear regression analysis well. In our first set of tests, we do not find evidence of cross-subsidisation in the aggregate distribution of fund returns. In the second set of tests, we use a modified version of Gaspar, Massa and Matos’ (2006) methodology, and again fail to find evidence of familial cross-subsidization. However, we find evidence of an alternative response to tournament opportunities: a modification of the risk level, conditional on year-to-date investment performance, consistent with Kempf, Ruenzi and Thiele (2009).