| Correlation of financial markets has been studied extensively in the last decades. Copula model not only considers the degree of correlation between financial time series, but also considers its related structures, so it has become an important tools for multivariate dependence structure in financial risk and risk management, asset pricing,protfolio construction and so on.However, the correlation between financial market or stock is not a specific type of form but changing. When the stock market in a bull market or bear market, the co-moment will be enhanced significantly between the relevant stock market. It is difficulty to use a single copula function to capture the structure of financial markets in that case. Mixture-Copula model can contain different type of Copula function. What's more, its parameter can measure the correlation between the variable and the coefficients of linear combination can capture different dependence structure. Unfortunately, copula model has been studied in binary structure in the past and the reality is often need to consider structure of multiple markets or stocks. So we adopt Mixture-Copula to anlysis the structure of multiple markets.We selected Shanghai Industrial Index, Business Index and Real Estate index as our empirical analysis's data from jan1,2002to dec,31,2011and the total number of observation is2425. We fit GARCH model as the marginal distribution for each index series.Meanwhile we select Gumbel Copula,Clayton Copula and Frank Copula in the archimede Copula to construe our Mixture-Copula model. Our result indicate that combing Monte Carlo simulation and Multivariate-GARCH-Mixture-Copula model to calculate VAR is effective.but, single structure such as t Copula, Gumbel Copula, Clayton Copula, Frank Copula may underestimate the real VAR.This means that Mixture-Copula can describe the potential structure.Meanwhile,we give the optimal combination coefficients in different confidence levels and research is valuable is risk on the prortfolio. |
PDF Full Text Request