Correlation Research Of Foreign Exchange Rate Based On Mixed Copula-GARCH-EVT Model

Since July 21,2005,when government announced that the RMB exchange rate began to implement the floating exchange rate mechanism which is mainly based on the regulation of the free market,and is driven by a basket of currencies,China's foreign exchange market mechanism has been gradually improved and China's foreign exchange market has become more and more mature.In October 1,2016,the RMB has become the fifth currency,following the dollar,euro,pound and yen,to formally join the SDR currency basket.It marks that China's foreign exchange market has stepped a key step into the international exchange market.With the reform of the foreign exchange market,the volatility of RMB exchange rate is becoming increasingly market-oriented.Meanwhile,the deepening of globalization also makes China' s trade with foreign countries more and more frequently.Therefore,the risk of foreign exchange rate(also known as exchange rate risk)that some foreign-invested enterprises are faced with,such as commercial banks,securities companies,import and export companies,and fund companies,has become increasingly prominent.So when our domestic foreign trade enterprises trade with the foreign enterprises,it is extremely vital to accurately measure the risk of foreign exchange in order to prevent and circumvent the currency risk.The interdependence study among markets is an important part of the risk control and measurement.Because Copula function can not only depict the correlation between time series,but also describe their dependency structure,it is often used as the main tool to explore the interdependence of multiple assets in risk control.However,the correlation structure of different assets in the market is not fixed and will not be a fixed pattern,so simply using a Copula function is difficult to effectively depict the correlation between structures of different assets.But the mixed Copula function,which is consist of copula functions with different properties,can describe the relationship between assets more accurately than only use one Copula function.In this paper,regarding the date of foreign exchange rate between euro and RMB,date between yen and RMB from May 21,2007 to August 10,2015 as the sample,We conducted a marginal distribution analysis in these two kinds of foreign exchange rate data,using the GARCH-EVT model.The empirical results show that the GARCH-EVT model figures out the Conditional heteroskedasticity of the residual sequence and its tail-tail characteristic and is more suitable for the subsequent Copula fitting.We select the Gumbel Copula,Clayton Copula and Frank Copula in the Archimedean Copula function to construct a mixed Copula model.The results show that the mixed Copula model,which is composed of Gumbel Copula,Clayton and Frank Copula,can express the correlation between multiple markets more accurately.