Size and book-to-market factors in a multivariate GARCH-in-mean asset pricing application

Abstract

The analysis of Fama and French (1993), is broadened to include time varying risk. This is achieved by an MGARCH-m application which extends the approach taken by Bollerslev et al. (1988) to a multiple index context. Appropriate weightings of the conditional cross-moments of returns on portfolios that make-up factor proxies are modelled as MGARCH. The mean equation of the model is designed to determine whether proxied sources of time-varying non-diversifiable risk can explain movements in excess returns on various types of portfolios. Time-variation in conditional excess return appears only to have a significant relation with time-varying conditional variance associated with a book-to-market factor.