Global uncertainties, volatility return and dependence structure on thematic investing using GARCH and copula approaches

Abstract

The current study employs GARCH modeling to explore volatility persistence and information asymmetries of certain thematic indices. The results confirm that the best way to solve heteroskedasticity problems in thematic investing is by using EGARCH and GJR-GARCH models with student-t distribution. In addition, we use several Copula models to reproduce asymmetries in asymptotic tail dependence. The statistics validate that T-student Copula is the optimal fit copula model to capture the dependence structure of thematic indices returns. After a simulation of 10,000 samples, we noticed large values of single multivariate distribution, meaning that future thematic series returns could be easily forecasted.