The Research On The Management Of The Commercial Banks' Exchange Rate Risk With The Method Of Value At Risk By GARCH

In November 30.2015,JMF announced that the RMB will be included in the special drawing right;this event is a milestone in the process of development about the foreign currency on the country.It is the inevitable product within the increasingly close relationship of the economy in China and the world.marking the RMB international process has taken an important step and indicating that the RMB in the international economic exchanges of the future will be more widely accepted and used by other countries.Commercial banks for their own foreign exchange risk are more focused on long-term or daily level of exchange rate risk under normal market conditions faced by holders of assets,in order to make an effective intervention in advance.Taking into consideration that the VaR is the comprehensive and immediacy measurement,this paper chooses it to measure the exchange rate risk of commercial banks.When commercial banks measure the foreign exchange risk,it often consider the single currency of risk while ignoring the linkage between different currencies.Based on this,from the perspective of a comprehensive measure of risk of departure,the linkage between foreign exchange risk are taking into account,this paper chooses different GARCH to measure the Value at Risk.By exploring the accuracy by different model.it gets the advantage and disadvantage for reference.This paper is divided into five sections.The first part is the introduction,including the background of the topic,the significance of the research and innovation and inadequate of this article.The second part describes the based theory of risk,mainly summarizing the foreign currency risk theory of commercial banks.The third part shows the VaR method and the the different GARCH model.The fourth is the practical part of comparison.The fifth gives the findings of research and some advises to measure risk.In the empirical research,this article selected the data in the about ten years from July 21th,2005 to December 31th,2015 to do research.and the data mainly includes the exchange rate between dollar against RMB,the exchange rate between Euro against RMB,the exchange rate between the yen against RMB and the exchange rate between the HKD against RMB.Based on the statistical characteristics of each currency rate of return,the exchange-rate yield data were build into univariate GARCH model,multivariate GARCH model and the panel GARCH model by selecting t distribution and GED distribution.By backtracking the degree of the VaR based on different GARCH model covering the past actual fluctuation,the conclusion can be draw that the distribution assumptions to the return of rate is not suitable for the distribution of t.On the other hand,when the distribution assumption is the distribution of GED.we can draw the conclusion that the model of BEEK or DCC also overvalue the risk and that the accuracy of measuring the risk through the panel GARCH is superior to univariate GARCH model.By the above.we draw that the panel GARCH is accuracy for measuring risk through considering the linkage effect between different currency rates.Then we give the change of risk in different currency rate or different bank so that we can provide a useful reference for the measurement about the risk of currency rate.The innovation of this paper is:firstly,this paper introduces panel GARCH into the measurement of foreign currency rate in commercial bank so that the VaR fully consider the linkage relationship between different currencies.Secondly,in term of the accuracy in measurement of risk,according to the actual statistical characteristics of series of the return of rate.this paper choices t or GED distribution as the distribution assumption and give up the normal distribution in order to measure risk more accurately.The potential shortcomings of this article:firstly,in the estimation of multivariate GARCH model,this lag is selected as the first order lag without thinking.While this simplifying the model to estimate,but it may miss some major information.secondly,The panel GARCH model,which takes into account the linkage,is easy to estimate,not only its individual effect parameters but also linkage effect parameters,but there is a simplicity in the model setting that the model lacks the expression of the risk transmission relationship among different currencies.