Measurement Study Of The Foreign Exchange Portfolio Risk

Chinese foreign exchange reserve has growth rapidly in recent years, it is over$3trillion until2012. However, dollar-denominated asset has accounted for about more than70%of the share of huge foreign exchange reserve. Apparently, it is the liabilities of the United States to China; in fact, if the RMB continues to appreciate against the dollar, it means that huge dollar-denominated asset held by China will continue to shrink. Therefore, it is particularly necessary to increase the value of the foreign exchange asset at the present stage. While appreciation of the foreign exchange asset, the core link is to measure the risk of foreign exchange asset portfolio, In2007, the U.S. sub-prime bonds housing mortgage crisis broke out, and it developed subsequently into the development of the financial crisis on Wall Street, and eventually evolved to the global financial crisis. In the context of the integration of global economic, after the impact of the subprime mortgage crisis and the impact of a variety of channels, the financial risk had transmitted to Chinese foreign exchange market. Finally, the volatility of foreign exchange market had been increased, and the volatility of foreign exchange portfolio yield had. In this international situation, the risk measurement of foreign exchange portfolio has become more urgent and necessary.If you want to study the portfolio risk, you need to study the correlation between assets first. The process of study on the structure of property has changed from static to dynamic, from linear correlation to nonlinear. The current mainstream research method is a dynamic study, assuming assets have the relationship with time-varying, and can change the corresponding with time changes. This matches the real time financial market conditions. So we consider using dynamic method to study the correlation between foreign exchange assets, as a further measure of market risk. Due to the unclear relationship between foreign exchange assets is linear, we determined separately from the relationship between linear and nonlinear to establish model and describe the correlation of assets structure. On this basis, we choose the risk measurement method which called VaR to risk measurement. Finally, in order to compare the two models of the actual value to degree the fitting, we use return detection method to make model comparison. Like the other time financial assets, Foreign exchange assets volatility have heteroscedastic phenomenon. In the econometric model of time series, GARCH model can research field of economic volatility. In many of the multivariate GARCH model, multivariate DCC-GARCH model has a strong advantage. First, the model allows portfolio fluctuations to consist with other GARCH model, and it makes the model more flexible. Second, because the model requires the estimation of the parameters and combination of assets into a linear relationship, the model is relatively simple. Third, the model uses two stage methods to estimate parameters, so at the same time it requires less parameter, and multiple model estimation is easy to achieve. In a number of asset of linear related to dynamic study, DCC-GARCH model is a good model.For nonlinear correlation portfolio, time-varying Copula model is now the powerful tools to study the correlation. Copula theory has many advantages in practical application. First of all, it is not subject to the marginal distribution of selection restrictions, so we can construct a plurality of flexible distribution of random variables, then separate the marginal distribution and the correlation to study. Second, the Copula function which is derived for consistency and relevance measure can make the strictly monotonic transformation not change. In addition, like the other financial time series data, the foreign exchange market data generally has the kurtosis characteristics. If the exchange rate of return does not follow a multivariate normal distribution, Copula model which is good at depicting the information can better fit the data.The article concludes that under the ground of United States sub-loan crisis, the correlation of U.S. dollar and the euro exchange rate, US dollar and Japanese yen, the euro and t Japanese yen exchange rate changed apparently. But after the sub-loan crisis correlation return to the normal state. We can see that the subprime mortgage crisis has not changed the correlation structure of the three foreign currencies and just has the external shock.DCC-GARCH model and time-varying Copula model can describe the time-varying correlation of the dollar, euro and Japanese yen. The index of the two models all passed the test. The time-varying point can match the recent events, and can be well reflected in the sub-loan crisis of the exchange fluctuation trend. The model has a strong practical. Time varying Copula is better than DCC models in accuracy, and can calculate VaR more accurately which will measure the foreign exchange portfolio risk better.