Shanghai And Shenzhen Stock Markets Based On Extreme Value Theory Analysis

Financial time series has peak thick tail, auto correlativity, volatility-clustering and volatility asymmetry characteristics. Therefore, based on extreme value theory and GARCH theory, this thesis used the TGARCH-POT model to establish the marginal distribution function of Shanghai Composite Index and Shenzhen Stock.Under the background of economic globalization, there is few country or region's financial markets are closed and isolated. Between different markets or different assets, there is often a relationship of mutual influence and volatility. The traditional approach is to use the t-Copula function to describe the structure of the two markets, but the symmetry of the t-copula function is not suitable for non-symmetric tail dependence variable. In the Archimedean Copula function, Gumbel Copula and Clayton Copula function has non-symmetry, so it can better describe the tail correlation of the market. In this thesis, we select an appropriate Copula function from the Gumbel Copula and Clayton Copula function to describe the relational structure of the Shanghai and Shenzhen stock markets.We selected the Shanghai Composite Index and Shenzhen stock's10years data from2002to2011to do empirical analysis, and obtained the failure number of different confidence levels in2010and2011years by comparing the Copula Monte Carlo calculated VaR value with the actual value. From which, we know (1) TGARCH-POT-Copula is more effectively than GARCH-t-Copula, it illustrates that the POT model in extreme value theoiy can improves the marginal distribution and GARCH-t-Copula obviously underestimated risk.(2) In2010and2011years, TGARCH-POT-Gumbel Copula is better than TGARCCH-POT-Clayton Copula, the main reason is that the economic recovery started to increase investors'confidence after the2008financial crisis. So in this period, Gumbel Copula function is suitable for describing the relational stucture of these two markets.