過去研究台股指數現貨與台股指數期貨的文獻，較少把不同的時間趨勢考慮進去。本論文旨在研究台股指數現貨與台股指數期貨間在不同時間趨勢下領先落後關係之探討，研究樣本為2003/01/02至2004/12/31的台股指數現貨每日收盤價資料及台股指數期貨接近下午1:30時的交易資料，依Fabozzi and Francis(1977)與Kim and Zumwalt(1979)的判定多空頭方法，從資料期間擷取出台股指數現貨與台股指數期貨之多頭、空頭、多頭轉空頭、空頭轉多頭此4段時間區段，以共整合檢定、誤差修正模型、Granger因果關係、衝擊反應函數及預測誤差變異數拆解，探討台股指數現貨與台股指數期貨報酬在此4段時間區段之領先落後關係，實證結果如下： 1.在多頭期間下，短期是由現貨領先期貨；而長期趨向均衡時則是期貨領先現貨，代表期貨在長期具價格發現功能。2.在空頭期間下，不論長短期皆是由期貨領先現貨，顯示較多的資訊是由期貨市場傳遞向現貨市場。3.在多頭轉空頭期間下，短期是由現貨領先期貨；而長期趨向均衡時現貨與期貨互有領先，但現貨領先強度稍強。4.在空頭轉多頭期間下，以短期而言，現貨與期貨沒有領先落後的關係；而長期趨向均衡時現貨與期貨互有領先，但現貨領先強度稍強。 In the past literature, many scholars investigated the lead-lag relationship between stock index, index futures while they ignore one of the important factor; that is, the phenomenon of ＂cimate mrkets＂. The purpose of this thesis is to investigateate the lead-lag relationships between Taiwan stock index and Taiwan stock index futures in different climate markets. By following the definition of Fabozzi and Francis(1977), Kim and Zumwalt(1979) with respect to different climate markets, we divide the market into four periods, which are the bull market, the bear market, the bull market turning bearish and the bear market turning to bullish. In this thesis, we use daily closing index prices and future prices closing at 1:30 P.M. during 2003/01/02~2004/12/31 and Vector Error Correction model and Granger causality model are used to analyze the data. Our conclusions are summarized as follows: (1)In the bull market, TAIEX leads TAIFEX in the short-run while TAIFEX leads TAIEX in the long-run equilibrium relationship. TAIFEX have a role of price discovery in the long-run. (2)In the bear market, TAIFEX leads TAIEX in the bull and bear markets, leading to a suggestion that TAIFEX transmits information to TAIEX. (3)In the bull market turning bearish, TAIEX leads TAIFEX in the short-run. In the long-run equilibrium relationship, TAIEX leads TAIFEX and TAIEX leads TAIFEX. TAIEX is more sensitive to information than TAIFEX in the long-run. (4)In the bear market turning bullish, there is no lead-lag relationship between TAIEX and TAIFEX markets. In the long-run equilibrium relationship, TAIEX leads TAIFEX and TAIEX leads TAIFEX. TAIEX is more sensitive to information than TAIFEX in the long-run.
