Asymmetric dependence structures for regional stock markets: An unconditional quantile regression approach

• Published in: The North American journal of economics and finance Vol. 52; p. 101111
• Main Authors: Dong, Xiyong, Li, Changhong, Yoon, Seong-Min
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.04.2020
• Subjects: Asymmetric dependence; Global economic factor; Regional stock market; Stock market condition; Unconditional quantile regression; C31; F62; Regional stock market; Asymmetric dependence; Stock market condition; Unconditional quantile regression; G11; C58; Global economic factor; F36
• ISSN: 1062-9408; 1879-0860
• DOI: 10.1016/j.najef.2019.101111

•We study the dependence structures for six regional stock markets.•The UQR approach provides more useful results than the traditional CQR approach.•Heterogeneous dependence structures are observed.•Relationship between geographically and economically close stock markets is strong.•US monetary policy and financial pressure impacts are more obvious during bear markets. Owing to the asymmetry of stock markets, this study investigates the dependence structures for six regional stock markets according to different market conditions by applying the unconditional quantile regression (UQR) approach. This approach can address the traditional conditional quantile regression (CQR) approach’s limitation that its distributions are defined conditional on specific covariates. Specifically, we not only examine the detailed linkages among these six regional stock markets, but also explore the effect of global economic factors on them, given the strengthening of both international investment and the globalization of financial markets. The results show these dependence structures are often an asymmetric U-shaped or inverted U-shaped structure, which indicates that the impacts of both other geographically and economically close stock markets and economic factors are more pronounced during bear and bull markets than during normal markets, especially so in bear markets. Moreover, the UQR approach provides stronger extreme-value relationships and more significant asymmetric effects than the traditional CQR approach.