Measuring the timing ability of mutual fund managers

This paper discusses the ARCH (Autoregressive Conditional Heteroscedasticity) effect on the Treynor–Mazuy index (TM), which is used to overcome shortcomings of other indices to measure the timing ability of mutual fund managers. The assumption behind the Treynor–Mazuy index is that the variances of individual portfolios and the market index are constant. The ARCH technique developed by Engle is used to test whether the relaxation of the above assumption has material impacts on the TM index. It is found that for most mutual funds, the assumption of constant variance is not correct. Rather, under certain conditions for some specific mutual funds, such as when the time horizon is very short, the assumption of constant volatility is reasonable. Using univariate ARCH(1) to model returns of mutual funds, we got different α (selectivity ability), β2p (timing ability), and TM values compared with those calculated with the conventional TM index. Thus, we should be cautious when applying the TM index to evaluate performance of mutual funds.