The Study Of Credit Risk Based On Copula Theory

Credit is the cornerstone of market economy. The damage caused by credit risk is an issue of great concern in finance. In 2008, the credit default problem attracted worldwide attention again due to the global financial crisis, which was triggered by the subprime mortgage crisis in the United States. The issues on how to measure credit risk effectively and how to prevent risk efficiently have become the essential targets for the financial institutions and corporations in the world. Further, measuring default correlation is one of the important elements in credit risk management.The conventional method of measuring default correlation is based on the linear correlation coefficient. It means that the most existing literature assume the market risk as following the normal distribution. However, credit risk is characterized by fat tail and asymmetric pattern which is proved by practice. Therefore, the deviation will be generated through the traditional measuring method. In this paper, we introduce the Copula function into the research of default correlation. The Copula function with flexible modality can be used to capture the complex part of default correlation, especially in tail dependence.The paper investigates the credit risk model, theory of copula function and the application of the Copula function in credit default correlation. We apply the latest research conclusions on the Copula function and carries out the relevant adjustment of the credit risk model as well as our national credit risk measurement. We provide the empirical analysis in financial data as well. The major achievements are summarized as follows:First of all, the thesis combines the Chi Graph, one of non-parametric test methods, with Copula model to study the correlative changes among various credit assets. Chi Graph could be applied to observe the relationship among the variables directly and would also determine whether the variables are correlative and define the characteristics of the tailed correlation. In the process of the empirical study, in order to filter the data exactly, a conditional heteroskedasticity test is carried out toward the required data. Then the output data are filted by using t-Garch tool. Finally, the portfolio credit assets would be analyzed profoundly by observing the corresponding Chi Graph as well as the characteristics of the tailed correlation of the data.In addition, the theory of semi-parametric Copula is introduced into the process of credit risk measurement. The advantage of such theory is that it can choose the proper non-parametric fitting method to fit marginal distributions, which lays a solid foundation for the parametric estimation of the Copula function. After receiving the significant marginal distribution, the thesis uses different Copula functions to carry out the estimation. Then the optimal Copula function could be picked up by taking the quantitative test with the Probability integral transformation method. Finally the exact function would be used to estimate the risk value of the portfolio credit assets and price the credit derivatives.Furthermore, the paper stops using value of shares instead of value of assets to calculate the VaR of portfolio credit assets. It proves that both of them could not be equivalent. In the theoretical system of the structured model, the total-range value of assets could be calculated based on the Value at Risk by using the B-S formula and Merton model. In the end, the estimated value of assets would be employed in the correlative research to get VAR of the portfolio credit assets.The final point is that our research analyzes the causes of 2008 global financial crisis and the current situation of the worldwide credit derivative market. The improved Copula function discussed above is applied to correct the default correlation, calculate the default time and improve the derivative pricing model. On the stage of the empirical analysis, instead of simple theoretical modeling, it uses Japan's credit default swaps market data to carry out the practical pricing analysis based on the above theoretical outcomes of this paper in this field.