Gray(1996)的一般化狀態轉換(generalized regime-switching, GRS)模型允許條件平均報酬與條件變異數同時隨狀態變數變動而變,本文先利用具有高低波動狀態的GRS模型,捕捉股價指數與股價指數期貨市場的個別報酬與波動性行為,並再進一步利用自我迴歸分配遞延(autoregressive distributed lag, ADL)模型,分別探討股價指數與股價指數期貨市場的報酬與波動性的動態關係。本文研究發現:允許條件平均報酬與條件變異數同時隨高低兩種波動狀態變動的一般化狀態轉換模型較能捕捉股價指數與股價指數期貨市場的報酬與波動性行為。就短期而言,股價指數報酬對股價指數期貨報酬的領先關係較為強烈,股價指數期貨波動性對股價指數波動性的領先關係呈現顯著,此外,股價格數與股價指數期貨的報酬與波動性序列分別存在長期均衡關係。
The generalized regime-switching (GRS) model, proposed by Gray (1996), is simultaneously specified for both the conditional expectation and conditional variance with different regimes. In this paper the GRS model with two regimes, high and low, is used to capture the respective conditional expectation and conditional variance in the Taiwan stock index and stock index futures markets, and then the autoregressive distributed lag (ADL) model is used to investigate the dynamics of return and volatility in the Taiwan stock index and stock index futures markets. The resultant shows that GRS model with two regimes is a better-fitted model to capture the conditional expectation and conditional variance in the Taiwan stock index and stock index futures markets. The return of stock index leading to stock index futures is larger than the return of stock index futures leading to stock index in the short-run. The volatility of stock index futures leads significantly to stock index in the short-run. A long-run equilibrium relationship exists between the return of stock index and stock index futures and it does too between the volatility of stock index and futures stock index.