外汇风险度量研究——基于GARCH类模型及VaR方法  

GARCH Models and Its VaR Empirical Analysis on Foreign Exchange Risk

作　　者：王德全[1] 

机构地区：[1]北京大学软件与微电子学院,北京102600

出　　处：《南方金融》2009年第8期11-15,10,共6页South China Finance

摘　　要：本文应用GARCH模型和VaR方法对美元、欧元、日元和港币对人民币的四种外汇汇率进行了实证分析,得出如下主要结论:四种汇率波动率序列均为非正态平稳序列,存在显著的ARCH效应,GARCH类模型可以有效地刻画其非线性动态波动特性。四种汇率均具有自我稳定功能,且美元和港币汇率波动的持续性显著地高于欧元和日元。在日元汇率中存在显著的非对称效应,而在美元和港币汇率中存在显著的风险补偿效应,欧元和日元的汇率风险大约为美元和港币的6-7倍;欧元汇率最理想的估计模型为GARCH(1,1)和IGARCH(1,1),而美元和港币汇率的首选模型为GARCH-M(1,1)-t和GARCH-M(1,1)-g,日元汇率的首选模型为PARCH(1,1)-t和EGARCH(1,1)-g。This paper empirically studies the daily change rates of four foreign exchange： USD/RMB, EUR/RMB, JPY/RMB and HKD/RMB using the GARCH models. Based on the models, we estimate the VaR of the foreign exchange rates. Finally we draw some main conclusions： there do exist the obvious ARCH effects on all foreign exchange rates; we can select the proper GARCH models which are better fitted with the nonlinear dynamic feature of foreign exchange rates. The leverage effect does exist only in JPY/RMB, namely, the volatility is bigger when the foreign exchange rate gets soaring, and is smaller when the rate gets down. The risk compensation effect does exist in USD/RMB and HKD/RMB, and the risk of JPY/RMB and EUR/RMB is about six or seven times large than that of USD/RMB and HKD/RMB. The GARCH （ 1, 1 ） and IGARCH （ 1, 1 ） are the best models in forecasting the VaR of EUR/RMB, the GARCH-M （ 1, 1 ） -t and GARCH-M （ 1, 1 ） -g are the best models in forecasting the VaR of USD/RMB and HKD/RMB, while the PARCH （ 1, 1 ） -t and EGARCH （ 1, 1 ） -g are the best models in forecasting the VaR of JPY/RMB.

关 键 词：外汇风险 杠杆效应 GARCH类模型 VAR 

分 类 号：F831.52[经济管理—金融学] F224