Assessment Based On The Var Of Gjr-garch Models And Extreme Value Theory

The money market is undergoing fluctuation and vulnerability that it never confronted. With the escalation of modern financial theories, information technologies, financial innovations and asset transactions, financial crises strike frequently in the process of economic globalization and financial integration. Financial administrations and institutions are strengthening the management and administration of market risk which has emerged as the primary in financial risks. Market risk will be more serious with the initial development of securities business. It is of great significance to research in how to control financial market risks in China and developing risk-management theories to evaluate stock market risks.This paper starts with the background information of risk-management in money market and introduction of Value at Risk (VaR). Using two risk-management models and taking into compare, GJR-GARCH-based VaR and Extreme-Value-Theory-based VaR, the author studies the forming of VaR model based on the researches done in this area. In order to meet the assumption in traditional Extreme Value Theory that the data analyzed is independent and identically distributed, ARMA-(Asymmetric)GARCH model is introduced to link the Extreme-Value-Theory-based risk-management model and POT model. Based on the researches incompareing the two models, it shows that the VaR values yielded from the 2 models are both effecttive and of trivial differences. However, GJR-GARCH-based model fails more easily than Extreme-Value-Theory-based model in calculating VaR and the distribution of yield rates needs to be considered. In the calculation of VaR with Extreme-Value-Theory based model, only the distribution of extreme value is considered. At the end of this paper, the disadvantages of VaR models are stated and improvements aresuggested.