Dependent Structure Of The Research, The Chinese Stock Market Based On Mixed Copula Model

Economic globalization and financial integration have greatly enhanced the interdependence between the global economy and financial markets.The relationship of financial markets and financial assets are more and more complicated.It shows nonlinearity, asymmetry and tail dependence and other models. Sklar's Copula theory believe that related information between random variables can be characterized by the Copula function completely,and can be used to describe the related patterns of financial markets.Copula theory widely used in the financial sector, particularly in the financial market risk management, portfolio selection and asset pricing, etc.It becomes a powerful tool to solve financial problems.This paper gives the related model of the six indices of Shanghai and Shenzhen stock markets by selecting the appropriate function.According to financial time series data's peak and fat tail characteristics,we established a t-GARCH(1,1) model of the six indices. And through the goodness of fit test,this paper proved that using this model to simulate marginal distribution function is reasonable. Because a single Copula function has the limitations to accurately describe the dependency structure of China's stock market,this paper do convex linear combination with some Copula function.Then a mixed Copula model was established based on Clayton Copula,Gumbel Copula,Frank Copula.It can describe Shanghai and Shenzhen stock market returns related to the asymmetric structure. Mixed Copula model's parameters are more than a single Copula model,with lager alternative function space and even more complex model structure.The empirical results show that using the mixed Copula model to describe the dependency structure of China's stock market is reasonable. It is better to capture the financial markets'asymmetric, nonlinear and tail dependency patterns.Although, in practice, the function has many advantages, we must take into consideration the potential local correlation structure and the difference of the random variables on hige-dimensional modeling.In view of this,we use pair-Copula method to do the correlation analysis.Then we analyzed the correlation among Shanghai Stock Index, Shenzhen Index and the Hang Seng index with this method.This paper use Gaussian Copula and t Copula to build the binary Copula between two random variables,and carried out the goodness of fit test.Finally,it comes to a good conclusion. The empirical results show that this method has great advantages in studying China's three major stock market returns.