以ADCC模型探討衝擊事件對亞洲主要股市間動態相關的影響

A Study of the Influence on Dynamic Correlation of Asia Stock Returns under Impact Events by ADCC Model

李欣儒(Shin-Ju Li)

指導教授 ： 李孟峰

國立臺北大學/商學院/統計學系/碩士(2009年)

摘要

以往對於金融商品報酬間的相關性研究，大多採用固定的相關係數衡量，但事實上金融商品報酬間的相關性經常隨時間而改變。因此本研究以Glosten et al. (1993) 提出的GJR-GARCH模型結合Cappiello et al. (2006) 所提出的ADCC模型探討亞洲主要股市間的動態相關，期望能捕捉隨時間變動的相關係數。同時在波動及相關係數模型中亦加入波動不對稱的考量，使模型更真實反映波動效果的不對稱性，並就十年間影響亞洲主要股市的衝擊事件分別觀察其變化。

本論文以美國股市作為外生變數，探討上海、香港、日本及臺灣股市報酬間的動態相關。實證結果證實美國股市對此四個亞洲股市報酬具有領先效果，並且香港股市報酬與日本及臺灣股市報酬間具有雙向的回饋關係，此結果亦顯示在這三個亞洲股市中，香港股市較具有主導的地位；此外，四個亞洲股市皆有顯著的波動不對稱性。最主要的研究結果即此四個亞洲股市在十年間的相關性顯著的隨時間而變動，因此採用動態相關係數衡量各股市報酬間的相關性，較固定相關係數更能掌握其變化。研究結果亦顯示：研究期間各股市的相關性皆有上升的情形，其中以香港與日本股市報酬間平均相關性最高，上海與臺灣股市報酬間平均相關性最弱，僅上海對臺灣及日本股市曾出現負相關的情形，其餘大部分的情況及各股市間相關係數皆為正，這意味著亞洲股市間存有共移關係，或相互依存效果。此外，根據動態條件相關係數趨勢圖發現，2006年以前上海股市與其他股市報酬間的相關性皆非常低，即使重大衝擊事件也很難對上海股市之股價指數產生影響；而2006年後，由於中國大陸金融政策的開放，上海股市與其他亞洲國家股市間報酬相關開始提高。在次貸風暴與雷曼兄弟倒閉所引起的金融海嘯後，整個亞洲股市報酬相關性皆上升，這亦顯示隨著市場的開放，金融危機比以前更能顯著衝擊亞洲股市間的共移關係，建議股票市場投資人在投資亞洲股市甚至是鄰近區域國家市場時，應將金融外生衝擊作為決策變數，以避免忽略市場訊息所增加的投資風險。

關鍵字

DCC ； GJR-MVGARCH ； ADCC ； 波動不對稱效應

並列摘要

In the past, the interdependence among the returns of financial merchandise was mostly evaluated by constant correlation coefficients. But, in fact, the correlation coefficients among the return of financial merchandises were changed frequently. Hence, this research combines the GJR-GARCH model proposed by Glosten et. al (1993) and Asymmetric Dynamic Conditional Correlation (ADCC) model proposed by Cappiello et al. (2006) to investigate the dynamic correlation of the major equity markets in Asia. And the changes of correlations resulting from the impact events in Asia during recent 10 years are statistically significant. Besides, the asymmetric property of volatilities and conditional correlations is also taken into account in the model. Then the model is not only can estimate the time-varying correlation coefficients among the returns of financial merchandise but also can capture symmetric property of volatility.
This dissertation is focused on investigating the dynamic interdependence among equity markets of Shanghai, Hong Kong, Japan and Taiwan by treating U.S equity market as an exogenous variable. Empirical results show that the return of U.S equity market really leads the returns of the four Asian equity markets. In addition, the return of Hong Kong equity market has feedback relationships with the returns of equity markets in Japan and Taiwan, which indicates that the Hong Kong equity market tends to be the most dominant in the three Asian equity markets. Moreover, significantly asymmetric volatility exists in all the four Asian equity markets. However, the most important finding is that the correlation coefficients of the four major equity markets in Asia changes over time significantly. So using the dynamic correlation coefficient model to evaluate the interdependence among equity markets is much better than using the constant correlation coefficient model. During the study period, we found that the relevance among the four markets was increasing. The average correlation coefficients between the returns of Hong Kong and Japan equity market is the highest, whereas that between the returns of Shanghai and Taiwan is the lowest. Among them, the only negative correlation is the correlation coefficient between Shanghai and two equity markets-Taiwan and Japan. In other situations, the correlation coefficients between any two equity markets are positive, which means that the co-movement effects or interdependence effects appear among the four Asian equity markets. In addition, according to the trend of dynamic conditional correlation, the correlations between the returns of Shanghai equity market and the other equity markets had been very low before 2006. The price index of Shanghai equity market had been difficult to be affected by even significant impact events .But, after 2006, it became more interrelated with other Asian equity markets due to the open monetary policy of China. Moreover, the interdependence of the whole Asian stock markets was increasing since the subprime mortgage crisis and the bankruptcy of Lehman Brothers. It suggests that comovement among Asian equity markets becomes more significantly affected by financial crises along with the opening of markets. Because of this, investors should take financial impact events into account while investing in equity markets of Asian or neighboring countries to reducing the risk of investment due to ignoring the market information.

並列關鍵字

DCC ； GJR-MVGARCH ； ADCC ； asymmetric of volatility

參考文獻

方文碩、王冠閔與董澍琦（2006）。亞洲金融危機期間股票市場的蔓延效果。管理評論，25（2），61-82。

王凱立與陳美玲（2001）。亞洲金融風暴發生前後美國與台灣股市動態關聯之進一步研究。經濟論文叢刊，31（2），191-252。

王冠閔（2007）。不對稱訊息下台灣股、匯市與美國股市蔓延效果之預測檢定。人文暨社會科學期刊，3（1），69-80。

林群復（2004）。中國大陸加入WTO後對其金融市場及兩岸四地股票市場大盤指數連動性之研究。國立臺灣大學國際企業研究所碩士論文，未出版，台北市。

楊聲勇、董澍琦、李昭蓉與黃喬郁（2006）。臺灣股票市場與期貨市場間價格與波動性傳遞關係之探討-EGARCH-DCC模型之應用。中國統計學報，44，417-439。