Research On The Dependence Structure And Risk Value Of CSI 300 Industry Index Based On MSEGARCH-EVT-Vine Copula Mode

In recent years,affected by the COVID-19 epidemic and the Russia-Ukraine war,China’s stock market has fluctuated sharply,indexes of various industries have shown increasingly complex correlations,and China’s financial security has faced major challenges.Therefore,it is one of the important topics in the field of financial statistics to quantitatively describe the dependency structure between industry indexes and scientifically measure the risk of stock market.It is also the practical need for the government to maintain the national financial security and the sustainable development of economy.In view of the possible structural mutation and extreme risk of the stock market industry index,this paper combines the MSEGARCH model,POT model in the extreme value theory and Vine Copula model to build the Msegarch-Evt-Vine Copula model.The effectiveness and superiority of this model in describing the industry index dependent structure are verified.On this basis,this paper also uses the above model to carry out an empirical study on the interdependent structure and VAR of various industry indexes.Firstly,the MSEGARCH-EVT model was used to fit the edge distribution of industry indexes,which was combined with three Vine Copula models.Then,the Vine Copula model with the highest goodness of fit was selected by AIC/BIC criterion and Vuong test to describe the dependent structure of industry indexes.Then,value at risk(Va R)was used as a risk metric to test the Va R prediction accuracy of the MSEGARCH-EVT-Vine Copula model.Compared with the Va R prediction results of the existing common models(EGARCH-EVT-Vine Copula and MSEGARCH-Vine Copula),the validity and superiority of this model in describing the industry index dependent structure were measured.Finally,under the background of COVID-19 and the Russia-Ukraine war,the MSEGARCH-EVT-Vine Copula model constructed in this paper is used to carry out an empirical study on the dependent structure and VAR of various indexes in the Chinese stock market.The empirical results show that: first,the fitting accuracy of R-Vine Copula model is better than that of C-Vine and D-Vine Copula models,indicating that R-Vine Copula model is more suitable to depict the interdependent structure of industry indexes in Chinese stock market.Second,MSEGARCH-EVT-R Vine Copula model has the best performance in Va R prediction.Compared with the existing common models,this model can more accurately describe the marginal distribution of the industry index,and then more accurately describe the dependent structure of the industry index.Third,the dependency structure and Va R of the industry index changed to varying degrees before the outbreak of the epidemic,in the early period of the epidemic,in the late period of the epidemic and during the Russia-Ukraine war,indicating that the COVID-19 pandemic and the Russia-Ukraine War had a certain impact on the industry index dependency structure and VAR of China.