Extreme Correlation and Asset Allocation in International Stock Markets

• Main Authors: Shan-Shan Shen, 沈姍姍
• Other Authors: Chu-Hsiung Lin
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/7m58cn

博士 === 國立高雄第一科技大學 === 管理研究所 === 96 === This study intends to explore the correlation of extreme returns between stock markets and its implications to the asset allocation. Two extreme value theory based approaches are used to measure the extent of extreme co-movements. One is the conditional correlation approach and the other is extreme dependence method. Empirical results show that unequal extreme dependence is observed from different sub-sample periods and the degree of co-movement gets even stronger during market turmoil. Extreme correlations across national markets tend to be much greater on the downside than on the upside and the co-movement is larger within developed markets than in emerging markets. Left tail dependence structure of the developed market is different from the emerging market. For developed markets, left tail dependence goes up as the threshold level increases, but tail dependence tends to go down in emerging market. With the exception of measuring extreme correlation, we explore the role of extreme correlation to international asset allocation utilizing the mean-lower partial moment (MLPM) portfolios. Empirical evidence demonstrates the diversification benefits of emerging markets and shows that extreme dependence of equities in a MLPM portfolio is significant lower than that outside the portfolio. Moreover, portfolios constructed of low extreme linkage markets have better performance than which is simply formed of low correlation markets. This result suggests that construct portfolios utilizing markets with low extreme linkage, especially low booming linkage, would be an ideal strategy.