Research Of Portfolio Credit Risk Measurement In Commercial Bank Based On Copula Model

Commercial banks as important elements of the entire financial system, whether it can operate healthily or not directly related to the development of the entire national economy. As one of the major risks, credit risk's measurement and management of commercial banks has always been a major task. Both China's accession to the WTO in 2001 and its financial market opening in 2007 made China's banking sector face enormous challenges. How to make China's banking sector be with international practice and its own management and operation comply with the requirements of the New Basel Accord as soon as possible is important issues of China's banking sector to resolve. The focus of this paper is to do some research on the application of Copula model in the commercial bank's credit risk measurement. Firstly, this paper analyzed the New Basel Accord and its methods of credit risk measurement requirements, at the same time, did some comparative analysis of Commercial bank credit risk measurement model which were developed by major international financial institutions to comply with the purpose and recommendations of the New Basel Accord. Through comparison and analysis, we can get the point that KMV model has an incomparable superiority whether in terms of theoretical framework or from the practical application. Moreover, it has extremely applicability under China's special financial market conditions. But its multivariate normal distribution assumption is still the defects exists in the KMV model. To solve the problem exists in the KMV model, it introduced Copula function in detail, and pointed out the superiority of its application. Finally, this article combined several Copula functions with KMV model, through Copula function to reflect the correlation between the assets and through the application of KMV model to measure the portfolio credit risk of the assets. By comparing the pattern and Data Results, we can get that t-Copula function can capture credit portfolio risk "thick tail" feature in linear Copula. To get more accurate results, four Copulas including Guassian-Copula, t-Copula, Clayton Copula and Gumbel Copula has been compared and we get the point that Clayton-Copula can capture the "left-fat-tail", Gumbel Copula can capture the "right-fat-tail", and t-Copula can capture both of the "fat-tails". Based on the above characteristics of the three Copula functions, this article then combined the three Copula function and established a mixed Copula function to describe the actual data feature.Throughχ~2 test, we get that mixed Copula function can describe the actual data well. Sowe at last get the conclusion that combining the mixed Copula function with KMV model to measure the portfolio credit risk of the assets can get a more accurate result.