使用考慮正負不對稱基差效果之馬可夫狀態轉換時變相關係數GARCH模型進行能源期貨避險

A Markov Regime Switching Time Varying Correlation GARCH Model with Asymmetric Basis Effect for Energy Futures Hedging

本文應用考慮正負不對稱基差效果之馬可夫狀態轉換時變相關係數GARCH 模型(Markov regime-switching time-varying correlation GARCH model with asymmetric basis effect; RS-VC-Basis-GARCH)進行能源期貨避險，這個模型將Lee and Yoder (2007a)所提出的馬可夫狀態轉換時變相關係數GARCH模型(Markov regime- switching timevarying correlation GARCH; RS-VC-GARCH)及Lien and Yang (2006) 所提出之考慮正負不對稱基差效果的時變相關係數GARCH模型(Time-varying correlation GARCH with asymmetric basis effect; VC-Basis-GARCH)相結合，探討同時考慮馬可夫狀態轉換及不對稱基差效果對期貨避險的影響。本文採用原油(Crude oil)、熱燃油(Heating oil)、汽油(Gasoline)及丙皖(Propane)等能源現貨期貨資料進行避險實證，結果發現加入狀態轉換後對所有資料均可改善其樣本外的避險績效，且熱燃油及丙皖在同時加入狀態轉換及正負不對稱基差效果後比只有考慮狀態轉換效果更進一步改善其避險績效。

This article applies a Markov regime- switching time-varying correlation GARCH model with asymmetric basis effect (RS-VC-Basis-GARCH) for energy futures hedging. This model nests within it the Markov regime switching time-varying correlation GARCH (RS-VC-GARCH) proposed by Lee and Yoder and the Time-varying correlation GARCH model with asymmetric basis effect (VC-Basis-GARCH) proposed by Lien and Yang (2006). We investigate simultaneously the effects of Markov regime-switching and the asymmetric basis on futures hedging. Four energy futures, crude oil, heating oil, gasoline and propane data are used for empirical study. We find that taking account of the effect of regime-switching improves the hedging performance out-of-sample for all data and taking account of both the effects of regime-switching and asymmetric basis further improves the hedging performance for heating oil and propane.