The Day-of-the-Week Effect' Asymmetry in Return and Volatility in China: An Empirical Study

Abstract

This paper aims to investigate the day-of-the-week effects' asymmetry based on the CSI300 Index, the Shanghai Composite Index and the Shenzhen Component Index. The Kruskal-Wallis test established that the day-of-the-week effects exist in Chinese stock markets. In addition, the Brown-Forsythe test was applied to investigating the constancy of variances across the days of the week, which demonstrated that only the Shanghai Composite Index satisfies the null hypothesis of constancy variances. Furthermore, the first-generation threshold models were used to assess the day-of-the-week effects' asymmetry. Comparative analysis of ARMA-GARCH, ARMA-GJR-ARCH and ARMA-APARCH using the Engle and Ng test implied that ARMA-GJR-GARCH and ARMA-APARCH could explain Chinese stock markets better; the Nyblom test showed the parameters' constancy in these models. Finally, and most importantly, with the aid of the sensible proxy variable of the volatility, the Diebold-Mariano test was used to compare the forecasting performance of competing models; the results indicated that for the CSI300 and Shanghai Composite Indices, ARMA-GJR-GARCH displays better performance than ARMA-APARCH. However, for the Shenzhen Component Index, no significant difference was observed between the two models.