Estimating Foreign Exchange Portfolios Risk Based On FIGARCH-EVT-Copula Model

With the reform of RMB exchange rate system and the continuous improvement of China's foreign exchange market mechanism,foreign exchange investment has become another important investment field after the stock investment.In recent years,all nations have began to reconfigure assets,which increases the fluctuation frequency and amplitude as well as foreign exchange risk in foreign exchange markets.Hence,it is extremely important for China's foreign exchange market good development to investigate and control the foreign exchange market risk.Therefore,in this paper,we adopt FIGARCHEVT-Copula model to evaluate the risk of foreign exchange rate portfolio based on the sample data of USD,EUR,JPY and HKD.By employing the classical R/S analysis and modified R/S analysis to test four kinds of foreign exchange volatility series of long memory,the research indicates that the fluctuation of China's foreign exchange market rate series possess long memory characteristic.In details,USD/RMB have the strongest long memory strength volatility series,followed by EUR/RMB,JPY/RMB,and HKD/RMB.Based on the four kinds of foreign exchange volatility sequence characteristic,first of all,the FIGARCH model which can well depict the volatility of long memory and the extreme value theory which better describe the distribution of non normal tail characteristics was introduced to build FIGARCH model.At the same time,the traditional multivariate normal Copula function and multiple tCopula functions of four kinds of foreign exchange portfolio risk were also adopted to obtain FIGARCH-EVT model.Through evaluating investment portfolio value risks for four kinds of foreign exchange,it was indicated that the risk values obtained by GARCH-EVT-Copula model were less than the FIGARCH-EVT-Copula model at the same confidence level,which indicated that GARCH-EVTCopula model possesses the possibility of underestimation of foreign exchange risk,FIGARCH-EVTCopula model can described the fluctuation of exchange the rate of long memory to measure foreign exchange portfolio risk more accurately.In order to better depict high dimensional correlation structure,this paper adopted the Pair Copula and constructed the FIGARCH-EVT-Pair Copula model and employed FIGARCH-EVT-Pair Copula model to measure the risk of four kinds of foreign exchange portfolio.Under the principle of minimal risk,it can be obtained that the optimal investment proportion of the four kinds of foreign exchange and the corresponding portfolio risk value.In details,compared with the traditional FIGARCH-EVT-tCopula model,the FIGARCH-EVT-Pair Copula model portfolio risk values were higher than the FIGARCH-EVT-t-Copula model.Therefore,we used Kupiec and Christoffersen back testing to test the efficiency of FIGARCH-EVT-t-Copula model and FIGARCH-EVT-Pair Copula model,which shown that the FIGARCH-EVT-t-Copula model could not passed the robustness test whereas FIGARCH-EVTPair Copula can pass he robustness test.All these indicated that FIGARCH-EVT-Pair Copula model possess more advantages than the traditional FIGARCH-EVT-Copula model in multiple foreign exchange portfolio risk evaluation.Besides,it was found that the minimum risk portfolio weight difference is very small no matter what kind of Copula function therefore the foreign exchange investment is still concentrated in USD asset.In this paper,we used FIGARCH-EVT-Copula model to measure the foreign exchange portfolio risk,which can provides a reference for further study of foreign exchange risk.