本文究旨在探討股、匯市之間的長短期非線性因果關係,以亞洲地區之台灣、韓國為例,資料來源以台股走勢、台幣兌美元匯率及韓股走勢、韓圜兌美元匯率為主要四大變數資料,期間始自2009年5月至2011年2月止,並對樣本進行資料處理,只留下兩國共有之交易日資料,以求交易資料具有一致性,共427筆日資料,在運用單根檢定、門檻自我迴歸模型(TAR),及動差門檻自我迴歸模型(M-TAR)進行門檻共整合檢定後,並進一步利用門檻誤差修正模型(TECM)及誤差修正模型(ECM),來檢驗其長短期非線性關係,最後進行Granger因果關係檢定。實證研究結果發現在台灣方面,不論是長短期,皆是股市會影響匯率,但匯率無法影響股市,但在韓國方面,短期時,股市會影響匯率而且匯率也會影響股市,在長期時,以門檻值-0.00872為上下區間分界點,在上區間時,股市會影響匯率,匯率也會影響股市,而在下區間的部份,則是股市會影響匯率,匯率不會影響股市。
The purpose of this article is to discuss the nonlinear causality in long and short term between the foreign exchange and stock market. Take Taiwan and Korea markets for example. The data are from Taiwan stock trend, TWD to USD exchange rate, Korean stock trend, KRW to USD exchange rate. The timeframe is from May 2009 to Feb. 2011. And to keep the data’s consistency, only keep the daily data on which both markets are open. There are 427 daily data in total. After the threshold co-integrate verification by Unit Root, Threshold autoregressive model(TAR) and Momentum-Threshold Autoregressive Model(M-TAR), make further use of Threshold Error-Correction Model(TECM) and Error-Correction Model(ECM), to check the nonlinear relationship in long and short term. Finally do Granger causality verification。 From the research result, we find in Taiwan, the stock market affects exchange rate, however the exchange rate couldn’t affect stock market whether in long term or short term. But in Korea, in short term, stock market affects exchange rate and exchange rate affects stock market as well. But in long term, take threshold value -0.00872 as cut-off point of upper and nether zone. At upper zone, stock market will affect exchange rate and exchange rate will affect stock market as well. But at nether zone, stock market will affect exchange rate, but exchange rate couldn't affect stock market。