Measurement Of RMB Exchange Rate Risk Based On Extreme Value Theory

November 30 th,2015,the International Monetary Fund announced that RMB will be join the SDR at October 1th,2016,the RMB boarded the international stage with the Dollar,Euro,Yen,Pound together constitute a basket of currencies.In the process of internationalization of RMB,the economic of US is recovering,the Dollar has gradually entered the ascending channel,the devaluation of the RMB is inevitable,a series of factors led to the RMB exchange rate sharply higher or lower,commercial banks and foreign-related enterprises as the main recipient of exchange rate risk,How to control the RMB exchange rate risk become an important issue to relevant institutions,however,how to measure the exchange rate risk is the prerequisite for effective control of the risk,so,finding a more effective way to measure the RMB exchange rate risk has become the most important problem.This paper chooses the daily reference rate for RMB against Dollar,Euro,Hong Kong dollar,Yen and Pound as the research sample.The value at risk and the expected loss of five foreign exchange are selected as the measure indicators,then,Empirical analysis and compared with the RMB exchange rate risk measure effectiveness based on the traditional value at risk method and the extreme value.95%and 99% confidence.Firstly,based on the method of traditional value at risk,compared the result which is based on the historical simulation method,variance-covariance,Monte Carlo simulation and GARCH models.The results show that the historical simulation method is more suitable for high confidence,under low confidence(such as 95%),it will underestimate the risks,The variance-covariance method measures the RMB exchange rate risk,the failure rate is highest,only the yen through the Kupiec back test,besides,it shows that the variance-covariance method is not suitable for measuring the exchange rate risk.Based on the Monte Carlo simulation method to measure the exchange rate risk,it is found that the Monte Carlo simulation can measure the exchange rate risk in extreme cases by simulating the possible paths for the exchange rate return series,Based on GARCH models,It shows that the normal GARCH and GARCH-t cannot measure the exchange rate risk of RMB,Since it has obvious thick tail.The GARCH-GED and the asymmetric EGARCH can be better messure the value at risk.which indicates that the GARCH-GED model can measure the RMB exchange rate risk better.Secondly,based on the extreme value theory,the BMM model and POT modelare selected to measure the RMB exchange rate risk.For the BMM model,the maximum value of block filters out much useful information,It makes the model is less effective under low confidence level,but the ES have better effects,which indicates that the BMM model is only suitable for the ES measure;for POT model,selecting the threshold is the most important thing.firstly,using the overrun function graph to select the threshold and measure the RMB exchange rate risk,through the back test,only the Euro risk value VaR failed to pass the test under 95% of the confidence level It indicating that the method can be a good way to measure the exchange rate risk of RMB.when using the overrun function graph to select the threshold,It has certain subjectivity so we use the kurtosis method to supplement.The results show that two methods to measure the exchange rate risk of RMB is similarly,All results pass the back test besides the VaR of Euro at the 95% confidence level,Comparison of these two methods,Although the kurtosis method is more complicated,but the result is more conservative and credible.Finally,according to the empirical results to make recommendations for commercial banks and foreign companies.