With the development of economic integration and globalization, national economies andfinancial markets are closely linked, once a troubled economy will involve other economies, even tolead to the edge of economic collapse. Such as the1997Asian financial crisis spread rapidly inSoutheast Asia and around the world and became a world-class to the evolution of the financialcrisis caused by the U.S. subprime mortgage crisis in2008. Visiblely, local fluctuations in financialmarkets may spread to other financial markets, and even enlarged, eventually evolved into acatastrophic financial crisis.Bank as an important participant in the world economy, it faces risks including credit risk,market risk and operational risk. Dependency structure between the various types of risk is animportant aspect of the study the size of the bank’s risk, and this is because the risk may includelinkage effects between the same risk factor, and therefore determining the risk of dependencestructure is the top priority of the bank’s risk in many integrated metrics. At present, due to thecomplex structure between the risk and the limitations of the research tools, is lack of accuracy in acertain extent, so it is necessary to introduce new technologies and methods to study theinterdependencies between risks, in order to get more effectively measure of the risk of inter-relatedstructure, to compensate for the limitations of traditional metrics need to explore more truly reflectthe risk characteristics of the various types of risk analysis tools.Many research results show that the marginal distribution of the various types of risk are heavytailed (Heavy Tailed), and this assumption can not use the normal distribution, linear correlationcoefficient for the variance does not exist thick tail distribution and does not reflect the variablescorrelation. Copula function itself has good properties, its nonlinear correlation structure can beaccurately reflected, thereby improving the accuracy of a measure of risk. As one of the importantfinancial institutions, commercial banks, is facing a variety of risks and risk nonlinear structure, theadoption the Copula function of this technology means you can provide a more accurate method forthe commercial bank’s risk comprehensive measurement.The nine commercial banks is listed by Empirical analysis, the first study is using linear factormodel to the three risk marginal distribution, ARMA-GARCH model is established marginaldistribution of credit risk and market risk; the second is by using POT model has been marginal distribution of operational risk; Finally by using Copula function, integrated the three types of risk,using Eviews and Matlab technical to obtain the Copula function of parameters, and solved ES byusing the SPLUS software. The final conclusion is the Copula function calculated results than thesimple sum of the different risk values much smaller, showing the simple sum of the different riskvalues overestimate the value of the overall risk, therefore Copula function of this technologymeans for the commercial bank risk comprehensive measure because of a more accurate method. |