Seasonality tests on the Shanghai and Shenzhen stock exchanges: an empirical analysis

Abstract

This article investigates Day-of-the-Week and January Effects in the Shanghai and Shenzhen stock markets over the period 1990 to 2006 for both the ‘A’ and ‘B’ indices. During this period, these two Chinese stock markets went through the limit period and nonlimit period and then again through a limit period. We examine the seasonality effects both during the different periods and also over the whole period. Our results indicate that the Shanghai A index is prone to higher volatility and also shows some January and Weekend Effects.

Notes

¹ See Mookherjee and Yu (1999a, p. 104).

² The descriptions of the indexes in this section have been obtained from the same database.

³ See Chen et al. (2005, p. 162).

⁴ We checked the data for stationarity. The augmented Dickey–Fuller tests have been conducted for all series. Null hypothesis of nonstationarity has been rejected in all cases. The test has been considered up to a maximum of 15 autoregressive lag lengths. The statistics for 5% left tail critical values of the limit distribution of Tau were SSEAD −40.2, SSEBD −39.4, SZSAD −40.78 and SZSBD −37.05. All four have a p-value of (<0.0001).

⁵ We conducted the Durbin–Watson test to check for autocorrelation and the White's test to check for heteroskedasticity. As the residuals of some of the regressions exhibited autocorrelation and/or heteroskedasticity, we used the Newey and West (1987) correction which corresponds to using a Bartell Kernel with bandwidth parameter L + 1 where L is the lag length. Due to space limitations, we do not report the exact statistics of the diagnostic tests. This also applies to the regressions in Tables 10–13 examining the January Effect. These values are however available on request.