對於未拋補利率平價說(Uncovered interest parity)過去有許多文獻，並不支持UIP之成立，而且大多數用一般線性迴歸模型，而近幾年來，非線性迴歸模型之應用大幅增加，許多匯率決定理論之實證發現，匯率的決定模型之參數調整，並不是跳躍的，而是平滑轉換的，故本文利用非線性模型探討荷蘭之未拋補利率平價說，以非線性平滑狀態結構轉換模型進行檢定，結果發現，荷蘭利差調整之非線性調整是以logisitc方式轉換，代表時間序列資料是以非線性的方式在調整，轉換速度是非常迅速的，並且存在著雙門檻，本文再對門檻內外之不同區間，再加以檢定，發現門檻外所觀察之樣本，UIP為成立的，期間為1985年1月至6月及1995年3月至1996年12月，分別成立的原因是，G5達成廣場協議聯合干預外匯及歐洲聯合放寬匯率波動幅度，而門檻內之樣本，UIP為不成立的。 The majority of researches have adopted general linear regression model to examine uncoverd interest parity (UIP) and the outcomes did not support the UIP theory in most studies. Recently, the application of non-linear regression model increases substantially. Most evidences of exchange rate analysis show that the adjustment of parameter on exchange rate model is smooth transition not jump pattern. Hence, this study use non-linear smooth transitional threshold model to examine uncoverd interest parity (UIP) in the Netherlands (Holland). Our results show that the adjustment of interest spread in the Netherlands is logistic transition, which means that time series data is modulated by non-linearly. Moreover, the result also indicate that the transition is highly frequency and with double thresholds. Besides, this study examines the UIP in different intervals of in- and out-threshold samples. We find that the UIP theory is hold on out-threshold samples in January, 1985 to June and March, 1995 to December, 1996 ,because G5 intervention foreign exchange in Plaza Accord and Europe unite the relaxation foreign exchange fluctuation scope but is not on in-threshold samples.
