The Improvement Of The Var,cvar Methods And Its Applications On The Financial Risk Measurement

At present, the financial industry has become the main engine of economic development, and the vicissitudes of the financial industry is directly related to the speed of economic progress. With the continuous development of financial innovation, the form of financial derivatives has become more diverse. On one hand, the development of the financial industry has been contributed to the financial innovation; on the other hand, the financial innovation also brings more risk. As we know, the global financial system has no support since the collapse of gold standard, so, financial risks which caused by financial innovations become more diverse and complex, so, it is more difficult for us to manage the financial risk. Therefore, in recent years, the research in financial risk management, especially in the areas of financial risk measurement become a hotspot.Through the exploration in the financial risk measurement, VaR (Value at Risk) methods have become a global consensus measurement of financial risk in recent years. Compared to the previous methods, VaR methods has many advantages.For example, it provides the same standard for different financial institutions to analysis different risk. However, the basic risk factors of the traditional VaR methods are assumed to be normal distribution, which can't reflect the real situations of the financial assets. What's more, it does not sub-additive. There are many problems which need to be solved for this method, in other words, it is not reliable.In this dissertation, we firstly point out the defects of the traditional VaR methods based on analysis of its statistical definition, modeling steps and other aspects. Then, we introduce the principle and modeling process of the conditional VaR methods and Copula-VaR methods. By this, we select a period of the futures prices and spot prices of the 1009 Shanghai copper, and obtained that the conditional VaR methods is better than the traditional methods. In the calculation of the risk of the portfolio value, we select two stocks, China Life Insurance and China Vanke, and obtain that the Copula-VaR methods is better than the traditional methods by positive analysis. In this dissertation, we improve some of the steps and methods in the modeling process of the positive analysis. In addition, we point out some defects of the research in the conditional VaR methods and Copula-VaR methods.For example, the use of linear mixed Copula function. We also discuss some feasible research ideas in this field.