On the Empirical Evidence of Mutual Fund Strategic Risk Taking

We reexamine empirical evidence on strategic risk-taking behavior by mutual fund managers.Several studies suggest that fund performance in the first semester of a year influences risk-taking in the second semester.However, we show that previous empirical studies implicitly assume that idiosyncratic fund returns (in a factor model) are uncorrelated across funds.We present generalized methodologies (based on both contingency tables and regression analysis) that accommodate the case of a general error structure.We show that the correlation between idiosyncratic fund returns is essential to the analysis and, when it is taken into account, the empirical evidence of strategic risk taking by fund managers disappears.