本研究以TAIFEX與MSCI兩種臺股指數期貨來規避其股價指數現貨之風險。運用OLS模型、誤差修正模型、單變量GARCH(1,1)、雙變量GARCH(1,1)與卡爾曼濾淨器等避險模型來估計避險比率，並比較兩種避險工具在不同模型下之避險效果，以尋求最適的避險工具供投資大眾參考。 實證結果發現兩種臺股指數期貨與現貨之時間序列資料並非呈常態分配，且其水準項具有單根的性質，而經一階差分之後（差分項）則皆成為定態數列。此外，兩種臺股指數現貨及期貨間皆存在共整合關係，此乃表示其現貨與期貨間存在有長期均衡關係。樣本外的避險效果比較則說明兩種避險工具在單變量GARCH(1,1)模型下可得到最佳的避險效果，且發現不論在各類模型下，MSCI摩根臺股指數期貨之避險效果皆較TAIFEX臺股指數期貨為佳。

　　　　　This paper considers hedge and basis simultaneously to investigate MSCI Taiwan Index futures and TAIFEX Stock Index futures, and we compare which is appropriate to hedge the Taiwan stock Index. It compares the hedging effectiveness in ECM model, univariate GARCH, bivariate GARCH model and Kalman filter model. The main empirical results re as follows, we find the significance of unit roots and thus the non-stationary of the price series, so price series should be difference to induce stationary. We also find evidence of cointegration between spot and futures prices. In the out-of-sample comparison, The univariate GARCH model outperforms all other hedging models, and TAIFEX Stock Index futures is the best instrument to hedge the Taiwan stock Index.