Investment styles and the multifactor analysis of market timing skill

Abstract

Purpose The purpose of this paper is to present and demonstrate how the use of a multifactor model in the analysis of market timing skill can be misleading because the use of a multifactor model does not suit all investment styles equally well. If the factors of the analysis model do not span the portfolio holdings of a fund with less conventional investment strategy, the use of a multifactor model may even deteriorate the overall inference in measuring the market timing skill of a large sample of funds. Design/methodology/approach This study investigates the limitations of multifactor models in the analysis of market timing skill by applying the traditional Treynor-Mazuy and Henriksson-Merton analysis models of market timing skill using a set of “placebo” funds which are “natural” passive market timers. Findings The results of the study show that the incorporation of the Carhart four-factor model into the analysis of market timing skill considerably reduces the percentage of significant market timing results. But, as expected, the reduction of bias is not equal for different investment styles, and it works best when the factors of the analysis model are related to the investment style of the placebo portfolio. Practical implications This style-related limitation of multifactor models in the analysis of market timing skill may result in detecting funds with less conventional investment strategies as market timers since the factors used in the analysis are not likely to span their investment styles. Originality/value This study shows that the use of a multifactor model may lead to inferring passive market timers with less conventional investment styles as market timers. In addition, the findings of the study leave option replication approaches as more preferable bias corrections than multifactor extensions.