Predicting returns of equity mutual funds

This paper investigates 1-year-ahead forecasts of actively managed equity mutual funds. A multifactor forecast model is developed that employs forecasts on the manager's skill, the fund's style and the expected factor returns. On the basis of a sample of German equity funds, we show that this forecast model substantially improves forecast power in relation to a nave forecast model, which just extrapolates past returns into the future. In particular, the multifactor model reduces the mean-squared error (mean absolute error) by up to 30 per cent compared to the nave model. More importantly, from the perspective of a mutual fund investor, the return of top-decile funds chosen by the multifactor model exceeds the average return of all funds by more than 200 basis points per year. © 2009 Palgrave Macmillan.