本文以1995年1月到1998年8月東南亞五國（新加坡、印尼、馬來西亞、菲律賓、泰國）的日、月、週三種不同形態的資料，利用單根檢定、向量自我迴歸（VAR）模型，來探討東南亞五國的外匯波動是否具有某種程度的傳遞效果。以單根檢定而言，不論何種型態的匯率資料均無法拒絕單根的存在，顯示東協五國的外匯市場大致來說趨向隨機漫步模式。接著我們以一般化自我迴歸條件異質變異數（ GARCH）模型，進行外匯波動性的萃取，而後以向量自我迴歸模型檢驗各國的匯率波動間是否具有一定的型態（Pattern）。結果發現不論是報酬率或波動性其在日資料所呈現的因果關係均較為顯著，至於月、週資料則較不顯著，換言之，不同資料型態的傳導效果並不一致。而就個別國家而言，泰國與菲律賓與其他三國的關係並不顯著，而新加坡、馬來西亞與印尼間的關係較為顯著。 This paper examines the Granger-causality for returns and volatilities among currency markets of Southeast Asian nations(Singapore、Indonesia、Malaysian、Philippine and Thailand) in the per iod from 1995 Jan, 1995 to Aug. 1998.We use unit root test、vector au toregressive (VAR) and different types of data (daily, weekly and monthly data) to investigate if there are any causality for returns and volatilities among currency markets of Southeast Asian nations. The results are as follows:All types of data don't reject the existence of unit root tests and indicate that the currency markets of Southeast Asian nations follow random walk. Next, we use GARCH models to form volatilities data and use VAR models to investigate if there are any causality for returns data and volatilities data respectively in different types of data. Results showed that the relationship in daily data is more significant than weekly or monthly data. In addition, we also find the relationship among these three nations, Malaysian, Indonesia and Singapore, are more significant than Thailand and Philippine.
