Nonparametric analysis of the Shenzhen Stock Market: The day of the week effect

The day of the week effect of the stock market has been studied extensively by researchers. Most of the conclusions previously made were based on parametric models with the assumption of normality. In this article, after examining the null hypothesis of normality of the log-transformed returns from the daily composite index of the Shenzhen Stock Exchange in China from December 25, 1995 to December 31, 2010, we rejected the null hypothesis of normality using three nonparametric tests. We then conducted nonparametric analysis of the transformed data using the Skillings–Mack statistic. The Skillings–Mack statistic is an extension of the well-known Friedman’s test with incomplete observations. We also compared the results based on nonparametric analysis to that from classic parametric models. Our results showed there was a significant negative Thursday effect in the Shenzhen Stock Market during the entire period and several selected sub-periods.