Asset Pricing Using Block-Cholesky GARCH and Time-Varying Betas
Grassi StefanoViolante Francesco
CEIS Research Paper
Starting from the Cholesky-GARCH model, recently proposed by Darolles, Francq, and Laurent (2018), the paper introduces the Block-Cholesky GARCH (BC-GARCH). This new model adapts in a natural way to the asset pricing framework. After deriving conditions for stationarity, uniform invertibility and beta tracking, we investigate the finite sample properties of a variety of maximum likelihood estimators suited for the BC-GARCH by means of an extensive Monte Carlo experiment. Finally, we illustrate the usefulness of the BC-GARCH in two empirical applications. The first tests for the presence of beta spillovers in a bivariate system in the context of the Fama and French (1993) three factor framework. The second empirical application consists of a large scale exercise exploring the cross-sectional variation of expected returns for 40 industry portfolios.
Keywords: Cholesky decomposition; Multivariate GARCH, Asset Pricing, Time Varying Beta, Two Pass Regression.
JEL codes: C12,C22,G12,G13
Date: Thursday 11 March 2021
Revision Date: Thursday 11 March 2021