Title: Modeling the dependence structure of VIX and SP500
Authors: Fabio Bellini - University of Milano-Bicocca (Italy)
Lorenzo Mercuri - University of Milan (Italy)
Edit Rroji - Universita' degli studi di Milano-Bicocca (Italy) [presenting]
Abstract: Copula-GARCH models are employed in the presence of multivariate time series as an alternative to standard multivariate GARCH models. Typically, univariate ARMA-GARCH models are fiited to the marginals while a dynamic copula is fitted to the innovations. We compare the fitting of different copula-GARCH models that describe the dependence structure of SP500 logreturns with the changes in the levels of Volatility Implied Index (VIX). Starting from findings in the financial literature, we require to candidate models to be able to reproduce two main features: strong negative dependence and radial asymmetry. We replicate the analysis in several subsequent time intervals in order to investigate the stability of results.