國際主要股市對亞洲股市相關性研究－DCC-CARR 模型的應用

Abstract

本文主要研究分析各國股市間動態相關係數行程，探討相關係數行程的大小 以及變動情形。以美國、日本和英國股票市場對十個其他亞洲股票市場的日資料 為主要分析對象，研究期間為1997 年7 月1 日至2007 年3 月9 日。應用Engle ( 2002 )提出的動態條件相關係數（dynamic conditional correlation, DCC）模型結 合Chou (2005)提出的條件變幅自我相關模型（Condition Autoregressive Range， CARR）模型估計動態條件相關係數。實證結果發現：第一，地區性股票市場對 亞洲股市的相關性比全球性市場對亞洲股市的相關性還高。第二，美國、日本對 亞洲發展較成熟的市場相關性較高。第三，美國、日本、英國對高經濟成長國家 的相關性較低。第四，在日本對台灣、日本對新加坡、日本對南韓的動態條件相 關係數平均數變化與股價趨勢呈正向關係。

This paper investigates the international stock market dynamic correlations between U.S., Japan, U.K., and ten Asian stock markets. Daily data are from 1990/7/1 to 2007/3/9. We use Condition Autoregressive Range (CARR) model of Chou (2005),as an alternative to the univariate GARCH in the dynamic conditional correlation (DCC) model of Engle(2002) first-step estimation. First, The correlations between Asian stock markets and a local idiosyncratic stock market are higher than the correlations between Asian stock markets and a global stock market. Second, the correlations between the mature Asian stock markets and Japan & U.S are higher than others. Third, the correlations between high economic growth Asian countries and international stock markets are much lower. Fourth, there is a positive relation between the mean of dynamic conditional correlations and stock price index in Taiwan, Singapore, and Korea with Japan.