本文以美國、英國、台灣等國之股價指數與指數期貨為主要研究對象，研究期間取自2001年1月1日至2008年12月31日止，採用CCC- GARCH避險模型，探討加入基差與變幅波動對避險績效的影響，實證結果發現在基差變數的比較上，CCC- GARCH避險模型加入不對稱基差的避險績效為所有修正模型中最高，比加入對稱基差的避險模型與CCC-GARCH避險模型好；在變幅波動變數的比較上，CCC- GARCH避險模型加入以Parkinson (1980) 或 Garman and Klass (1980) 或Rogers and Satchell (1991) 所估算波動率的避險模型並無一致的結果；在基差和變幅波動變數的比較上，CCC- GARCH避險模型加入不對稱基差的避險績效最佳，而不考慮變數的CCC-GARCH模型避險績效最差。最後，以優勢預測能力檢定（Superior Predictive Ability Test；SPA）模型檢定，評估模型預測績效優劣，CCC- GARCH避險模型加入不對稱基差，可提供投資人決定最適避險比率及衡量避險績效之參考。 This thesis takes S&P 500, Dow Jones, FTSE 100 and Taiwan stock indexs as the research object. The sample period covers from 1/1/2001 to 31/12/2008.With the use of the constant conditional correlation GARCH framework, and incorporating the decomposed basis and range volatility into the model to estimate hedging performances. The empirical results indicate that asymmetry effect model provides better hedging performance than the symmetric effect model and CCC-GARCH model. The hedging performance of CCC-GARCH also improves significantly by the inclusion of extreme-value volatility. The volatility estimates, based on the Parkinson estimator, provide better forecasts than those based on the Garman and Klass or Rogers Satchell estimator. Furthermore, use the SPA Test to determine which model has better accuracy in predicting the hedging performance of the actual market. In conclusion, the result indicates that asymmetric basis effect model has the best hedging performances. Asymmetric basis effect model provides investors to decide the hedging ratio of futures and to measure hedging performance.
