本研究探討現貨市場與期貨市場間的波動現象，由於過去的文獻多針對現貨與期貨的價格行為上研究，較少從事價量關係的實證。故本研究藉由價量關係的基礎，使用不對稱GARCH模型，以檢驗台股期貨與現貨市場間之報酬與波動的關聯性。實證結果如下： 1.在不對稱GARCH-M模型的實證中發現，無論現貨或期貨市場皆存在顯著報酬的不對稱性，其負向的報酬衝擊大於正向的報酬衝擊，且使波動性增加。2.代表風險貼水的係數值呈現顯著，表示市場上的投資人屬於風險規避者，視股票與期貨為風險性資產，表示當風險上升時，投資人將會要求資產報酬的增加。3.價量關係的實證上發現，現貨市場上並沒有充分的證據支持成交量與波動性呈正相關。4.將交易量區分為預期與未預期後，加入在對方的變異數方程式中，以觀察其波動性的改變。整體而言，未預期交易的衝擊效果遠大於預期交易，此種外溢的效果，也影響到對方市場的交易行為，尤其存在期貨市場。亦即當現貨未預期到的交易量會明顯使期貨市場的波動增加，也同時改變期貨市場的交易行為（即期貨市場交易量上升）。5.在不對稱模型GARCH-M中，以最大概似值擇其最佳之模型，發現NGARCH優於EGARCH、GJR模型；在預測方法之比較上，GARCH-VAR優於移動平均法。 The focus of this research is to study the behaviour of volatility in two parallel markets: the spot market and the market for futures on spot index. Previous studies on the relation between spot and futures markets mostly focused on price behaviors but few on price-volume relationships. This study investigates the relationship between trading behaviors and volatility impact for Taiwan stock index futures and stock index using the asymmetric GARCH-M model. Furthermore, we partition each trading volume into expected and unexpected components based on Moving-average approach and VAR model. From empirical results, we obtain the following summary: First, in the asymmetric GARCH model, an asymmetric effect exists on spot and futures markets. The negative impact level is greater than the positive impact level to the conditional volatility. Second, when the risk (volatility) of spot and futures increased, the holders would require risk premium for compensation. Third, by examining price-volume relationships, we find no evidence to support that the increase in volume would raise volatility on spot market. Fourth, after adding the expected and unexpected volumes into volatility equation, we find that the unexpected trading is greater than expected trading, and this spilled effect is over anther markets, especially in the futures market. Finally, among three asymmetric GARCH models, NGARCH is superior to EGARCH and GJR models. As for forecasting approaches, GARCH-VAR application is superior to Moving-average.
