Based On Extreme Value Theory And Copula Function Of The Rmb Exchange Rate Risk Measure

The liberalization of foreign exchange rate market increases its uncertainty and the exchange rate risk that foreign invested economic subjects such as commercial banks and enterprises face aggravates since the reformation of CNY exchange rate at July 2005. How to reinforce the exchange rate risk management becomes a burning question of each economic subject and the most critical step is to measure the exchange rate risk accurately. So far, Value-at-Risk (VaR) has been accepted as the core risk measurement of financial market risk. Many foreign financial institutions have adopted VaR to measure their market risk. However, VaR is rarely applied to the domestic commercial banks and enterprises. But with the development of our financial market, it is necessary to adopt VaR to measure risks.Among existing studies, though few academicians use several kinds of VaR models to measure the exchange rate risk, they rarely focus on the CNY exchange rate and the methods which calculate VaR are simple and rough. Based on this current situation, this paper tries to find an adapted method to measure the exchange rate risk of CNY. Considering that CNY exchange rate series all have sharp peaks and heavy tails, we reject the nomal distribution assumption and utilize the generalized Pareto distribution (GPD) in extreme value theory (EVT) to fit the series' tails, and then we can estimate the VaR. Result of back testing indicates that comparing with VaR values calculated by historical simulation (HS) and variance-covariance method, VaR values calculated by EVT can measure the risk more accurately while dealing with JPY/CNY and EUR/CNY. Namely, Foreign invested economic subjects can adopt EVT to estimate VaR values of these two exchange rates to measure their risks. But EVT-based VaR values underestimate the risk of USD/CNY and HKD/CNY, this may be caused by the continuous CNY appreciation toward USD and HKD. Therefore, to measure the risk of this two exchange rates pricisely, we must combine EVT with other qualitative and quantitive methods, or we can consider the risk measure called expected shortfall (ES). Considering that commercial banks and enterprises hold exchange rate portfolio in most cases, this paper combines EVT with copula theory to build the copula-EVT model. And then calculates the VaR for the exchange rate portfolio of EUR/CNY and JPY/CNY by Monte Carlo simulation to measure its exchange rate risk. The back testing result indicates that, comparing with VaR values based on HS, the VaR values based on copula-EVT model can reflect the risk of this exchange rate portfolio more accurately. That is to say, commercial banks and enterprises can adopt this method to measure the exchange rate risk of this portfolio. They can also reserve funds to deal with the risk or take steps to avoid risk according to these VaR values. In addition, we also estimate the expected shortfall (ES) as a supplement of VaR.