在研究期間中,向量誤差修正模型顯示台指期貨在價格發現中具有較強的主導地位;資訊比例模型顯示台指期貨對於價格發現之貢獻程度較大。藉由迴歸去分析影響價格發現之因素,結果指出,開紅盤日、週末效應及異常訊息的發布為顯著負相關,表示台灣50ETFs的價格發現能力高於台指期,另外交易量增加及波動度變大也會使得台灣50ETFs對台指期的價格發現增強。可能的原因為在假日的過程中,累積的訊息大部分為個別公司或是產業的消息,而非整體市場的訊息,使得開盤後,市場在反應訊息時,現貨的價格發現能力高於期貨。 This paper studies competition in price discovery between the Taiwan 50 ETF (TTT) and TX for the different types of information announcement on Hasbrouck’s (1995) information shares model. The sample period extends from January 2, 2005 to December 31, 2012. Using matched synchronous intraday trading data, Johansen’s maximum likelihood estimator is employed to disclose the cointegration relationship among the TX and TTT.
Results indicate that two price series are a cointegrated system with one long-run stochastic trend. Estimated coefficients of the vector error correction model suggest that price adjustment takes place in the TTT, but not in the TX. The results show that the TX provide more price discovery than do the TTT overall. However the contribution of the TTT to discovery increases in the time surrounding information announcement. Because informed traders with private firm-specific information are more likely to trade in individual stocks, and possibly increase informational contribution of the spot market.