Analyzing Contagion Effect in Markets During Financial Crisis Using Stochastic Autoregressive Canonical Vine Model

Abstract

Using the multivariate dynamic framework, we study the contagion effect through the change in the dependence structure of major stock markets of the USA, China, Japan, Russia, Hong Kong and India, during the recent (2015) slowdown in China market, the financial crisis called the subprime crisis (2008) of the USA and the Russian flu (1998). The dependence structures in these markets are captured by applying the time-varying C-Vine copula model combined with the AR(1)-GJR-GARCH(1,1) (autoregressive Glosten–Jagannathan–Runkle generalized autoregressive conditional heteroscedasticity) model. At first, we apply the AR(1)-GJR-GARCH(1,1) model to specify the marginal distribution of returns from the market indices and subsequently we apply the time-varying C-Vine (Canonical Vine) copula to estimate the dependence structure among these markets in pre and post phases of the financial crisis. The empirical findings establish the existence of contagion effect in these markets during the periods of depression. Clayton copula turns out to be the best fit copula after the China slowdown in all cases except for China–Russia pair, indicating an increased lower tail dependence in the post crisis period. The subprime crisis witnessed a significant increase in Kendall’s tau and its time-varying behavior in all pairs of countries except USA–China where a static dependence structure is observed. The same phenomena are observed during the Russian flu in Russia–India, Russia–USA and Russia–Japan pairs indicating a significant contagion effect. These empirical findings have significance in risk management for banks and financial intermediaries which invest majorally in the global markets.