本文以新台幣對美元、歐元與日元三種貨幣之利差交易，實證檢驗 以新台幣進行利差交易獲利之可能。相較於傳統利差交易僅以兩 國利率之相對大小決定操作部位，我們將匯率基本面變數所隱含的 訊息納入利差交易策略，探究其改善利差交易操作績效之可能。同 時考量台灣管理式的浮動匯率制度之獨特性，我們也將央行干預匯 率市場之替代變數納入模型，檢視進一步提高利差交易績效之可行 性。相較於過去以樣本外預測均方誤之大小評估匯率預測模型之優 劣，本文透過利差交易獲利績效之提昇與否，賦予匯率基本面變數 以及央行干預因素的納入決策之經濟價值。我們發現經濟基本面變 數及央行干預之考量，其績效之貢獻在傳統利差交易所實現之績效 大幅下降的金融海嘯後時期，益加凸顯。同時該利差交易之報酬，與 市場大盤超額報酬的連動性不高，而無法爲投資人承擔系統性風險 之風險溢酬所解釋。

The profitability performance of some NTD carry trade strategies against major international currencies subject to intervention are investigated. Of particular concern in this paper is whether the profitability of the strategies taking into account information on fundamentals and/or intervention is better than that of the naive strategy considering interest differentials alone. The paper establishes significant evidence in support of this idea. The evidence presented is remarkable for the time series sample's post global financial crisis of 2008. Moreover, the returns to the strategies can not be attributed to the existence of risk premia, as the excess returns of either the S&P 500 or TAIEX, proxies for systematic risks, and found to lack power to explain the profit differentials.