An Empirical Research On Integrated Risk Of Chinese Commercial Bank Based On Vine Copula Method

In recent years, with the deeper of economic globalization and the stronger of our national strength, financial industry is diving towards internationalization and diversification financial innovation has become increasingly active. Commercial bank, as one of the most important participator of the financial industry, is facing growing risks. Risk tend to be more complicated and diversified, the correlation between different risks are increasing as well. Commercial banks gradually began to take risk measure and control as an important part of the management of bank.Due to the risk faced by commercial banks is now no longer a single risk, but a combination of a variety of risks, we must find the right measure method of integration of commercial bank risk. Generally, the traditional method is to evaluate the loss of each single risk and add them up as the integration risk.However, in fact, the risks faced by commercial bank are not completely independent, there are some correlation between them. While when using the traditional linear additivity method, we ignore the correlation between risks, which must lead to the evaluate risk deviates from the real value, resulting in decreased efficiency of banks’ risk management.Therefore, how to measure the correlation between risks becomes the key point of the study. According to previous researches, Copula function has a good effect in this type of the application. Copula function can capture the nonlinear and unsymmetrical correlation between variables, and the joint distribution of variables can be got simply by the marginal distribution of each variable. These features make it widely used in various types of risk management, asset Pricing and other fields.Clearly pointed out in the2004release of "Basel Ⅱ", the commercial banks are facing three of the most important and serious risks:credit risk, market risk and operational risk. We consider of using vine Copula function for integration of commercial bank risk modeling, which integrate the credit risk, market risk and operational risk of China’s commercial banks.This paper is divided into six chapters. The first chapter is an introduction, the second and third chapters are the theoretical part, the fourth and fifth are the empirical part, and the last chapter is the conclusion and outlook.The first chapter introduces the background of the research and the significance of the article, reviews the results of previous studies in relative fields, gives the overall structure of the article and point out a few innovations of this paper.The second chapter gives the definition and nature of the Copula function, and then describes the structure and characteristics of several commonly used two-dimensional Copula function. On this basis, we focus on a Copula function to be used in this article, vine Copula. A detailed analysis in the structure and terms of use of two different structured vine Copula are given—C vine Copula and D vine Copula.The third chapter describes the specific steps in evaluating and fitting of the three risks, and then gives the molding steps of the vine Copula and a Monte Carlo method in evaluating VaR based on vine Copula function, at last we describes the backtesting method which used to measure the effect of the estimated VaR value.In chapter four, we do the empirical study on commercial banks’three single risk measure. After introduce the source of the data, we model three single risks as we described in chapter three, and then we evaluate the marginal distribution of each risk. Fitting results suggest that the risk-benefit ratio of three sequences are subject to ARMA-GARCH distribution, the difference is that the residual of credit risk and market risk sequences obey the Student-t distribution, while the residual of operational risk obeys standard normal distribution.In chapter five, it comes to our empirical study on the integration of the risks. With the marginal distribution of each risk we’ve got previously, according to the vine Copula modeling steps described in Chapter three, we estimate the type and parameter values of Copula function. Then we estimate the VaR of integration risks through the Monte Carlo method based on vine Copula. As the comparison, we use the traditional linear additive way to integrate the three major risks and get its VaR value. Then we use the backtesting method to check the effect of two VaR, we find out that vine Copula functions is better than the traditional method and the traditional method underestimated the real integration risk.The sixth chapter is the conclusion and outlook of the article. We summarize the conclusions of the study and point out some improvement suggestions.In conclusion, on the basis of our predecessors, it bring forth some new ideas:First, we choose the vine Copula function to integrate the risks. The main idea of the vine Copula is to break down a multi-dimensional Copula into a series of paired two-dimensional Copula through "vine" structure. Comparing with the high-dimensional Gaussian Copula and high-dimensional Student-t Copula and high-dimensional Archimedean Copula function, vine Copula is more flexible and credible.Second, based on vine Copula modeling, we estimate the VaR of the integration risk. And as the comparison, we get the VaR of the traditional linear additive way. Then we use the backtesting method to check the effect of two VaR, we find out that vine Copula functions is better than the traditional method and the traditional method underestimated the real integration risk.