本論文主要探討美國貨幣政策對美國及亞洲主要國家(中、日、台3個國家)之經濟成長(gross domestic product, GDP)影響是否存在非線性關係。本文選取美國聯邦基金利率(federal fund rate, FFR)為美國貨幣政策之代理變數，以FFR及四個國家之消費者物價指數(consumer price index, CPI)為解釋變數。研究方法則採用Granger and Teräsvirta(1993)及Teräsvirta(1994)所提出的平滑移轉迴歸模型(Smooth Transition Auto Regression, STAR)，檢驗FFR對四個國家之GDP是否有顯著影響力，並說明不同解釋變數對四個國家之GDP有無關聯性。
經實證結果發現，FFR對四個國家之GDP影響存在非線性關係，且皆有顯著門檻值。其中，FFR對台灣GDP之轉換速度為四個國家中最大，受衝擊後調整回均衡值最快，衝擊時間最短；FFR對美國GDP之轉換速度為四個國家最小，受衝擊後調整回均衡值較慢，衝擊時間較長。在不同LSTAR(Logistic Function STAR)模型下，解釋變數對四個國家之GDP有不同關聯性影響。其中又以小型開放經濟體台灣，對於來自大國的衝擊反應最敏感，有較高的解釋能力；其餘中、日、美皆屬大國，解釋能力則較不顯著。 Theme of this study is to verify the non-liner relationship between monetary policy and gross domestic product(GDP) in USA and three major Asian countries(China, Japan and Taiwan). This research uses federal fund rate(FFR) as proxy variable on behalf of monetary policy and four countries’ consumer price indexes(CPI) as explanatory variables. Smooth Transition Auto Regression(STAR) Model, which was developed by Granger, Teräsvirta(1993) and Teräsvirta(1994), is used to examine FFR and GDP of four countries if there is any significant correlation effect or not. It may also explain if there is any correlation between USA and three Asian countries.
Our research suggest that non-liner relationship is obvious between FFR and GDP of four countries; and its threshold is significant. It was noted that Taiwan shows the greatest conversion effect from FFR onto GDP while as USA is minor. By using different LSTAR(Logistic Function STAR) models. It was noted that explanatory variables to four countries’ GDP has different correlation effects, it appears that Taiwan has got a significant impact due to its open a small-scale economy. The impact on bigger countries, USA, China and Japan shows a relatively minor.
