CMStatistics 2020: Start Registration
View Submission - CFE
Title: Four Australian banks and the multivariate time-varying smooth transition correlation GARCH model Authors:  Timo Terasvirta - Aarhus University (Denmark) [presenting]
Anthony Hall - Humboldt-Universitaet zu Berlin (Germany)
Annastiina Silvennoinen - Queensland University of Technology (Australia)
Abstract: Daily returns of four Australian banks, often called the Big Four, are modelled. For this purpose, we use a new multivariate volatility model that belongs to the family of conditional correlation GARCH models. The GARCH equations of this model contain a multiplicative deterministic component to describe long-run movements in volatility and, in addition, the correlations are deterministically time-varying. Parameters of the model are estimated jointly using maximum likelihood. Since the model is strongly nonlinear, we also present certain features of the estimation algorithm.