避險是投資活動中相當重要的項目之一。當市場存在套利交易理論時,則須考慮套利交易對避險的影響,因而須利用具有門檻的誤差修正模型,探討避險組合的避險績效。本研究以2001年1月2日至2016年9月30日的臺灣加權股價指數現貨每日收盤價及臺灣加權股價指數期貨的最靠近13:30交易價格為研究樣本,利用移動視窗法(Rolling-Window)探討樣本外的動態避險績效,針對普通最小平方法(OLS)、向量誤差修正模型(VECM)及門檻向量誤差修正模型(TVECM)的DCC-GARCH模型來做比較。本研究發現OLS的避險績效顯著優於VECM-DCC-GARCH及TVECM-DCC-GARCH模型。此外,VECM-DCC-GARCH與TVECM-DCC-GARCH模型的避險績效則沒有顯著差異,此或許隱含套利交易理論不存在於台灣的指數現貨與期貨市場,因此有門檻的模型反而績效較差。本研究的研究發現可做為投資人的參考。 Hedging is playing an important role in investment. When arbitrage theory exists, it needs to consider the effect of arbitrage on hedging. So the error correction model with thresholds should be used to investigate hedging effectiveness of hedging portfolio. This study examined Taiwan index spot daily close price and Taiwan index futures transaction price that is occurred close to 13:30 from January 2, 2001 to September 30, 2016. The rolling-window method is used to investigate dynamic out-of-sample hedging effectiveness of OLS, VECM-DCC-GARCH and TVECM-DCC-GARCH model. The results show that hedging effectiveness of OLS model is significantly better than that of VECM-DCC-GARCH and TVECM-DCC-GARCH models. Furthermore, there is no significant difference in hedging effectiveness of VECM-DCC-GARCH and TVECM-DCC-GARCH model, which implying arbitrage theory does not exist in Taiwan index spot and index futures market. These findings in this study can be used as a reference for investors.