The Application Of Copula Model In Correlation Study Of Shanghai-Shenzhen-Hongkong Stock Markets

Nowadays, with the development of global finance, every stock market in the world have become to co-movement.In the early days of Chinese stock markets, there seems no relationship between Chinese stock markets and some main stock markets in the world, such as American, England, Japan, and so on. But along with the development of socialist market economy and the practice of reform and opening up policy, our country and international economic dependency have increased. On the other hand, with a series reforms goes on, such as the reform of non-tradable shares of finish, the formation mechanism of exchange rate reform, and the introduction of QFⅡ and QDⅡ system, many institutional defects that have restricting the development and expansion of Chinese stock markets have gradually eliminated. Therefore, co-movement between Chinese stock market and the rest of the world, which will indefinitely impact not only micro decision-making of investors but also financial regulators deeply becomes more and more obvious. Therefore, study the relationship between stock markets systematically has great significances.In chapter there of this paper, Copula function will be introduced for it’s complication. In fact, it links joint distribution function of random variables and their respective edge distribution function together. This paper firstly introduces the definition and basic related prosperities of Copula function, and then discussed several kinds of dependence, including Kendall dependence coefficient τ and tail dependence coefficient λU, λL. Subsequently, the paper carried on the detailed introduced there kinds of common used Copula function, including Gaussian-Copula, t-Copula and Archimedes-Copula. After that, this paper emphatically researches their dependence, especially the tail dependence.In empirical aspect, this paper choose three index daily closing price from December,1996to March2012daily closing price as sample, they are Shanghai composite index (SHI), Shenzhen stocks index (SZI), as well as Hengsheng index (HSI). In order to overcome the deficiency of traditional study methods, this paper uses ARMA (p, q)-GJR (m, n)-to simulate there stock markets’distribution. Based on this, time-varying correlation of SHI-HIS and SZI-HIS can be made by using Gaussian Copula-DCC (1,1) model. At the same time, in order to study probability of crash between stock markets in extreme cases, which is called the effect of risk spillover, time-varying tail correlation are also made by the application of the latest SJC-Copula model in the end of this chapter.The empirical results show that, correlation between Chinese stock markets and Hong Kong stock market has been increasing steadily in nearly seven years. Based on the existing theories, this paper tries to explain the results from the following aspects: one is QFII and QDII system, and another is regression to A-share of H share enterprises.