本文主要是探討變數間的變異數因果關係,本文採用義大利學者Massimiliano Caporin 於2007年發表的指數因果模型Exponential Causality multivariate GARCH (EC-GARCH) model。此模型結構由一個指數函數乘上傳統GARCH模型而成,使用指數函數可以將變數間的影響直接的表達出來。有鑑於中國的股票市場在國際間愈來愈重要,本論文主要的四個變數選用上海A指、深圳A指兩指數其報酬率波動及成交量變動。各變數資料期間為1996年1月5日到2012年4月6日的週資料,818個觀察值。藉由檢驗變數間的變異數因果關係,了解變數間影響方向及影響程度,以提供交易者有利的資訊。在檢測四個模型之後,本研究結果發現四個模型皆得到同樣的結果,亦即本期報酬率波動會影響下一期成交量的變動。由此可了解,上海A指及深圳A指,其報酬率的波動扮演較重要的角色,交易者藉由觀察報酬率波動,可了解到成交量的波動情形。
This study investigates on the variance causality by applying the Exponential Causality multivariate GARCH (EC-GARCH)model, which was introduced by Massimiliano Caporin in 2007. The main structure of this model is an exponential factor multiplying the traditional GARCH equation to drive the causality relation. This research mainly focuses on the biggest stock exchange of China, selecting return and trading volume of both Shanghai Stock Exchange A Share Index and Shenzhen Stock Exchange A Share Index as our four variables. In addition, used data is weekly data from January 5, 1996 to April 6, 2012 with total 818 observations for each variable. Our purpose is to analyze the variance causality among four variables to provide some useful information for the traders. As the result, by testing on the variance causality between Shanghai Stock Exchange A Share Index and Shenzhen Stock Exchange A Share Index, we find out an unexpected return volatility shock at time t-1 will induce the investors to invest on the stock at the time. The results emphasize the importance of return volatility on predicting the volume change of Shenzhen Stock Exchange A Share Index and Shanghai Stock Exchange A Share Index. The traders in the stock market may realize the volume change from investigating the return volatility.