Correlation And Risk Analysis Of Foreign Exchange Portfolio Based On EGARCH-EVT-Copula Model

With the rapid development of China's economy and the intensification of economic globalization,the exchange rate has become one of the important variables in China's economic operation,and financial investors are increasingly favoring foreign exchange investment.However,due to the re-allocation of world resources and the acceleration of the global economy,foreign exchange risk have been expanding.This also has aggravated the frequency and fluctuations of foreign exchange.Therefore,it is particularly important to study the dependent structure,investment ratio and portfolio risk among different types of exchange rates.This paper selects the US dollar,the euro,the Japanese yen and the Hong Kong dollar against the RMB exchange rates as the portfolio sample data.First,we must model the four foreign exchange rate series.but most of the financial series,especially the volatility of the foreign exchange rate series,not only have heteroscedasticity,but also have asymmetry.Therefore,in order to describe the heteroscedasticity and asymmetry of the volatility of foreign exchange rate series,this paper introduces the derivative model of GARCH model-EGARCH model.The four foreign exchange yield series are modeled by the EGARCH model,and four standardized foreign exchange residual sequences are obtained.It is judged by empirical analysis that the standardized residual sequence has thick tail characteristics,so the extreme value theory which can describe the thick tail characteristics of the foreign exchange standardized residual sequence is introduced to reduce the tail risk.The edge distribution of the four foreign exchange standardized residual sequences is fitted by the EGARCH-EVT combined model.In order to fit the dependent structure between the four foreign exchange standardized residual sequences,this paper introduces the Copula function which studies the correlation between variables.Find correlation coefficient matrices between four foreign exchange and perform correlation analysis.The results show that the correlation between the US dollar and the Hong Kong dollar is the strongest,with the correlation coefficient being as high as 0.9313.while the US dollar is negatively correlated with the yen and the euro.Correlation analysis between exchange rates allows investors to capture foreign exchange market information and accurately adjust the investment ratio.Then Monte Carlo simulation method is used to simulate the total return rate of the four foreign exchanges at the moment,and then the VaR method is introduced to analyze the risk of the portfolio.Firstly,the combined risk values of four foreign exchanges under equal weight investment are obtained by using EGARCH-EVT-t-Copula and EGARCH-EVT-Gaussian-Copula models.Comparing the combined risk value VaR and CVaR.The results show that the normal Copula underestimates the foreign exchange risk,which in turn indicates that EGARCH-EVT-t-Copula model analyzes the risk of foreign exchange portfolios more cautiously,accurately and effectively.Finally,the Min-CVaR method is used to find the minimum combined risk value and the corresponding investment weight.For example,when the confidence level is 99%,the EGARCH-EVT-t-Copula model finds the VaR value: 0.0016.The CVaR value is 0.0020.The corresponding investment weights are: 0.84,0.10,0.04,0.02.This paper uses EGARCH-EVT-Copula model to investigate the correlation and risk analysis of the foreign exchange investment portfolio,and provides investors with theoretical reference value.It also has certain reference significance in adjusting the proportion of foreign exchange assets investment and avoiding risks.