本研究主要在探討台灣、美國股價、匯率、利率之動態關聯，採用1991年10月至2013年2月的月資料，首先，進行變數的單根檢定，以ADF單根檢定、PP單根檢定及KPSS檢定法，根據單根檢定結果，均為I(1)數列。再利用Johansen共整合模型來檢定所有變數長期下是否存在穩定的均衡關係，其模型存在共整合向量。以誤差修正模型來探討短期動態調整過程。在誤差修正模型中發現台灣加權股價指數及匯率皆受到美國標準普爾500股價指數落後期的影響；台灣金融業同業拆款利率、台灣加權股價指數及匯率本身皆受到匯率落後期的影響；而本研究之變數除了美國標準普爾500股價指數受到網際網路泡沫及雷曼破產事件的衝擊影響明顯顯著，其他變數皆以亞洲金融風暴時，最為顯著。Granger因果關係實證結果，美國標準普爾500股價指數單向領先影響匯率，而匯率亦單向領先影響台灣加權股價指數，因此，美國標準普爾500股價指數單向領先影響台灣加權股價指數亦符合檢定結果 向量誤差修正模型VECM實證結果方面，VECM實證結果發現台灣、美國在短期內匯率、利率負向關係，最後考慮了變數間長期關係和異質變異性現象VECM-GARCH模型部分，結果發現與VECM的模型估計結果大致相同。 The purpose of this paper is to investigate the dynamic correlation among stock indices, exchange rates, and interest rates in Taiwan and America. The monthly sample data are from October 1991 to February 2013. Based on three different unit root tests-ADF、PP and KPSS, all series became stable after first difference. Johansen’s cointegration method has been applied to make sure the long-run equilibrium relationship among all series, and the methodology of cointegration also provide evidence of unique cointegrating vector. Results from VECM confirm that Taiwan Stock Index (TAIEX) and exchange rates (EX) affectd by lag periods of S&P500 index (S&P 500). The series of Taiwan Interbank Call Loar Rate (TICLR), TAIEX and EX affected by the lag periods of EX. Besides, in this study, all series have significantly effects by Asian Financial Crisis except S&P 500 which has significantly effects by dot-com bubble and the bankruptcy of Lehman Brother. The findings from Grandger Causality Test show S&P 500 leads EX, EX leads TAIEX, and S&P 500 leads TAIEX. The finding that in short-run the relationship between exchange rates and interest rates is negative in VECM and the same results also be found in VECM-GARCH which considered the long-run relationship and heteroskedasticity.
