本文以對稱的GARCH模型和不對稱波動的GJR-GARCH、NGARCH、QGARCH、VGARCH模型來來估算包含日本、新加坡、與台灣等匯率投資組合的風險值。主要目的在於比較這幾種模型對於估算外匯投資組合的風險值，何者有較準確的預測和財務風險的管理效能。實證結果如下：包含台幣、日幣與新加坡幣的外匯投資組合當中，台幣之不對稱特性並不明顯，所以使用對稱與不對稱的模型來估測外匯的風險值，其效果差異不大。由於匯率資料本身的訊息不對稱性較弱，所以使用GARCH模型來估計資產的波動性即可得到不錯的風險管理效果，並不需要使用具有可以反應訊息不對稱的不對稱GARCH模型。
In this paper, we use symmetric GARCH model and asymmetric GARCH model which includes GJR-GARCH, NGARCH, QGARCH and VGARCH model to evaluate VaR (Value-at-Risk) of the foreign exchange portfolio which includes three currencies: Japan Yen, New Taiwan Dollars and Singapore Dollars. The purpose of this paper is to compare the capability of the evaluation of VaR with these five models and finds out which model has better risk prediction and financial management capability. The empirical results are as follows: when our portfolio contains Yen、Taiwan New Dollars and Singapore Dollars, there is no significant difference in these five models. Because asymmetric property of exchange rate of return is weak, we just need to apply symmetric model to evaluate the VaR of the exchange portfolio.
