The Research On Exchange Risk Management Of Domestic Commercial Bank Based On Extreme Value Theory

Commercial bank's risk management has always been an important study in theoretical and practical field. In recent years, with the international community's impact of economic globalization and economic liberalization, it has introduced financial innovation continuously and increased the volatility and risk of the financial market obviously. Especially after the 2008 financial crisis, international factors such as political, military and nature etc, the entire financial situation to deepen continually. Since July 21, 2005, China began to implement market-based, with reference of a basket of currencies, managed and floating exchange rate system, it make the range of RMB exchange rate increasing, so it further increasing uncertainty of exchange-rate risk. At the same time, China's foreign exchange reserves are? increasing, which is now combined with the post-crisis era, the extreme volatility in financial evident. In this context, how to strengthen the exchange rate risk management has become a serious problem of commercial banks, and thus an accurate measure of exchange rate risk has become a priority research topic.This paper used extreme value theory researching the exchange rate risk between RMB and dollar in current situation. It is mainly researched the extreme value in the exchange rate risk measurement applications, using the rate between RMB and U.S. dollar as the study object, analyzing their statistical distribution of income, using extreme value theory to simulate the tail yields, and using Matlab and Eviews softwares to estimated parameters, then calculate the VaR and ES values. Furthermore, back testing the model to verify accuracy of it.The results show that: (1) The return series of most financial assets are fat tails distribution and the volatility appears the characteristics of aggregation and durative. Traditionally, VaR and ES, which is based on normal distribution, neglected the time deformation of volatility, so that the risk is likely to be estimated rather lowly since the information in the tail of the models is effectively unable to be captured. (2) In this paper, which associated with the theory of different variance-POT, the author put forward the moving theory of GARCH Model under the thick tail distribution used to analyze the fluctuation of the yield in financial market. Meanwhile, the dynamic model of GRACH-EVT(ES) is set up with thick tail distribution (GED distribution) applied to it as well as with its validity. (3)The result of the validity shows that the model of the GARCH-EVT-VaR(ES), based on the Extreme Value Theory(EVT), adapts to the financial time series in terms of thick tail distribution much better, and, therefore, it is one of the most effective tools to portray the risk at the tail of financial market.