Applied Research Of Financial Risk Measurement Based On GARCH Cluster-Copula-CoVaR Model

The financial time series is different from the general time series,its distribution function often shows the characteristics of peak and thick tail,and the study of financial time series affects the vital interests of investors,so the risk measurement of financial time series is particularly important.Since the 1980 s,the interaction of financial markets among economic entities has been increasing,and the traditional research methods have exposed many shortcomings.Therefore,this thesis combines GARCH cluster models and Copula models,and uses CoVaR method to obtain the optimal TGARCH-Copula-CoVaR model to measure the risk of funds.Firstly,this thesis selects 1095 trading days' data of four representative open-end funds,and analyzes their logarithmic return series.It is found that the series has the characteristics of volatility clustering,skewness,peak and thick tail.Using the normal test,it is found that the series does not obey the normal distribution.Then,ARMA models and GARCH cluster models are used to model the four funds respectively,and the maximum likelihood estimation method is used to estimate the parameters of the model.It is found that the TGARCH model is better than the GARCH model to describe the different responses of funds to positive and negative shocks.The maximum likelihood estimation method,AIC criterion and BIC criterion were used to compare the ability of different Copula models to describe the relationship between funds,and the conclusion was obtained that t-Copula model performed the best.Finally,this thesis calculates funds' VaR and CoVaR,and gets the conclusion that CoVaR method is more sensitive to the risk of open-end funds than VaR method.And through the research on CoVaR of funds,it gives investment suggestions for fund investors.TGARCH-Copula-CoVaR model we choosed not only considers the different reactions of funds to positive and negative shocks,but also studies the interaction between different funds when measuring risk.The empirical study also shows that the financial risk measurement based on TGARCH-Copula-CoVaR model has a good effect on open-end funds,and can provide investors with some investment suggestions.