Arch Effects in Multifactor Market-Timing Models of Polish Mutual Funds

Performance measurement of investment managers is a topic of interest to practitioners and academics alike. The traditional performance evaluation literature has attempted to distinguish stock-picking ability (selectivity) from the ability to predict overall market returns (market-timing). However, the literature finds that it is not easy to separate ability into two such dichotomous categories. To overcome these problems multifactor alternative market-timing models have been proposed. The author's recent research provides evidence of strong ARCH effects in the market-timing models of Polish equity open-end mutual funds. For this reason, the main goal of this paper is to present the regression results of the new GARCH(p, q) versions of market-timing models of these funds. We estimate multifactor extensions of classical market-timing models with Fama & French's spread variables SMB and HML, and Carhart's momentum factor WML. We also include lagged values of the market factor as an additional independent variable in the regressions of the models because of the pronounced "Fisher effect" in the case of the main Warsaw Stock Exchange indexes. The market-timing and selectivity abilities of fund managers are evaluated for the period January 2003-December 2010. Our findings suggest that the GARCH(p, q) model is suitable for such applications.