Extreme Risk Spillover Measurement Between Markets Based On Time-Varying Mixed-Copula Model

The United States subprime mortgage crisis broke out and quickly spread to other financial markets in2007, and eventually led to a financial crisis sweeping the world. This fact fully indicates that global financial markets is not isolated, accompany the rapid development of economic globalization and financial integration, the relation between global financial markets increasingly increased, and the performance forms of financial risk also increasingly complex of and diversification, if less in considerations about the extreme risk spillover in times of crisis will largely underestimate the risk level of financial markets, which may cause disastrous consequences. How to measure the extreme risk spillover level between financial market effectively is an urgent task to risk manage department and financial regulators.Accompany the rapid development of financial innovation,Financial risk,especially in recent years, original of some analysis method as based on linear related of analysis method, has no longer adapt to these requirements, and a new method which can solve nonlinear, asymmetric dependent relationship of Copula model is in badly need, and quickly application to financial markets research of all area in the world wide, and in use of assets pricing, financial risk regulatory, risk management and prevention, insurance pricing and became an effective tool. This paper tries to construct a time-varying mixed Copula model, and this model could select different type of time-varying Copula to describe financial variable according to different of distribution regional. This model using time-varying Gumbel Copula function to describe dependent relationship in the upper tail of joint distribution, using time-varying Rotated Gumbel Copula function to describe dependent relationship in the lower tail of joint distribution,and using the mixed Copula functions to capture the changes of dependent relationships in the other region of joint distribution. This method could overcoming the flaws of a single Copula function that only suitable for describing dependence of the joint distribution at a specified region, compared to mixed Copula functions, this model also has a time varying parameter, and the related parameters will change in different periods, especially suitable for financial variables modeling in the critical period. Therefore, time-varying mixed Copula model is better to capture samples in different areas of variable and describe dependent relationships between seemingly unrelated patterns of change, especially in the critical period is more suitable for joint distribution models to the financial variables. This paper is based on time-varying mixed Copula model, presents a new method for calculating indicators of extreme risk spillover CoVaR, and using this method to analysis the extreme risk spillover effect between the United States stock market, China’s mainland stock markets, United Kingdom stock markets and Hong Kong stock markets. By take backtesting the result shows that the extreme risk spillover indicators calculated in CoVaR between markets based on a Normal Copula, Time-Varying Normal Copula, Gumbel Copula, Time-Varying Gumbel Copula and the mixed Copula model are unable to pass the test, this means that these models underestimated the extreme risk spillover effects. But the extreme risk spillover indicators calculated in CoVaR between markets based on time-varying mixed Copula model could pass the test,so is much better than the other models in capture extreme risk spillover. The time-varying mixed Copula models in this article has the significant advantage in extreme risk spillover measurement between different financial markets, and can be used as an effective tool for risk measurement.