台灣加權股價指數與美國三大股價指數關聯性之研究

過去文獻討論台灣與美國股價指數聯動性，隨著時間國際經濟局勢的變化與樣本區間及研究方法不同，得到不一致的結果。本文利用近期日資料探討台灣股價指數與美國最重要三大股價指數，道瓊股價指數、標準普爾500股價指數、那斯達克股價指數之因果關係，探討是否美國三大股價指數都可以影響台灣股價指數，還是僅有部分美國股價指數可以影響台灣股價指數。另外，也進一步分析台灣股價指數是否具有預測美股股價指數的能力。最後探討台股與美股三大指數，排序四個指數間預測能力強弱，提供最新台灣與美國股價指數聯動性分析供投資者參考。本文資料期間是從2005 年 1 月 1 日到 2019 年 11 月 14 日，以向量自我迴歸模型，分析台灣與美國三大股價指數之相互關連性，並以因果檢定分析台灣與美國三大股價指數報酬的領先與落後因果關係。實證結果顯示：四個指數間不存在共整合關係，表示變數間不存在長期穩定關係；那斯達克股價指數報酬會影響台灣加權股價指數報酬，但是道瓊股價指數報酬與標準普爾500股價指數報酬並不會影響台灣加權股價指數報酬。台灣加權股價指數報酬並不會影響美國三大股價指數報酬，說明台灣與美國股價指數連動性關係，只存在那斯達克股價指數報酬單向影響台灣加權股價指數報酬，而台灣加權股價指數報酬對美國股價指數報酬沒有任何預測能力。

關鍵字

股價；向量自我迴歸模型；因果關係

並列摘要

Results of research on relationships between Taiwan and US stock markets are inconclusive. Because past papers adopted different sample periods and methodologies, and international economics change from time to time. This paper adds to this debate by using latest daily data to examine Granger Causality among Taiwan Stock Index and Three US Major Stock Indices: Dow Jones stock index, S&P 500 stock index and Nasdaq stock price index. This paper investigates whether all US three major stock indices can predict Taiwan stock index, or only some indices can predict. Furthermore, this paper also examines whether Taiwan stock index can predict US three major stock indices. Finally, this paper examines the ranking of predictive power among 4 stock indices for investors with the latest relationships between Taiwan and US stock markets. To explore the Granger Causality among four stock indices, this paper conducts vector autoregression model to analyze and collects daily data from January 1, 2005 to November 14, 2019. The empirical results show that there is no cointegration relationship among the four indices, indicating that there is no long-term stable relationship among the variables; In the Granger Causality Test, only Nasdaq stock index will affect the Taiwan stock index as a one-way causal relationship; however, Taiwan stock index couldn’t predict any US stock index.