A component model for Dynamic Conditional Correlations: Disentangling interdependence from contagion

We analyze whether the crisis sourced in US is spread over the world by contagion or through interdependence. Within this work, contagion is defined as a significant increase in cross-correlations after a crisis hits a country, we assumed that correlations are not constant over time and also evolve according to a GARCH(1,1)-type structure which give rise to the use of the popular DCC model introduced by Engle (2002) and extended in Colacito et al. (2011) to disentangle the short and long run component of the total correlation of the portfolio under study. We link interdependence with long-run fluctuations in correlations and contagion is associated with the short-run correlations.