SAIFM Financial Markets Journal
 
The South African Financial Markets Journal
17th Edition May 2013

Volatility and Anomalies in the Johannesburg Securities Exchange Daily Returns

Ziv Chinzara and Stuart Slyper

sing three models that differ according to the degree to which they accommodate and allow risk to vary across days of the week, we examine whether anomalies exist in the Johannesburg Securities Exchange daily returns. Simultaneously, we also examine the relationship between risk and return across the different days of the week. We estimate the measure of time-varying market risk using an asymmetric GARCH model. While the results vary across the three models, we generally find that after allowing for risk to vary across the days of the week, anomalies exist in the JSE daily returns. More specifically, the JSE daily returns seem to exhibit significant positive returns early in the week and significantly negative returns later in the week. Furthermore, the relationship between risk and returns is contrary to the portfolio theory in later days of the week. We outline the investment implications of these findings.

Introduction

Many stock market anomalies have been documented within the past three decades including the January Effect (Keim, 1982), the Momentum Effect (De Bondt and Thaler, 1984), the Daylight Savings effect (Kamstra et al., 2000) and numerous other such anomalies. The day-of-the-week effect is one of these anomalies. These effects are regarded as anomalies since they do not conform to the Efficient Market Hypothesis (EMH). This is due to the fact that their presence implies that an investor could get arbitrage profit by buying stocks on days when their returns are significantly negative and short them on the days when their prices are significantly positive.

The day-of-the-week effect was coined after it was noticed through empirical testing that the return of shares differs according to which day of the week it is. Numerous past empirical studies have observed differences in return between days of the week. Cross (1973: 69) noted that between 1953 and 1970, Fridays had a significantly higher mean return than Mondays. Subsequent studies have also measured the differing volatilities of indexes on different days of the week. US markets between 1973 and 1997 were found to exhibit the highest volatility 2 on Fridays and the least volatility on Wednesdays (Berument and Kiymaz, 2001: 181).

While there have been relevant studies on the South African equity market (cf. Bhana, 1985; Le Roux and Smit, 2001; Basher and Sadorsky, 2006; Mbululu and Chipeta, 2012), the effect of domestic risk on the day-of-the-week effect has not received attention. The current study attempts to fill this gap by investigating whether returns on the SA equity market exhibit the day-of-the-week effect and analysing whether this anomaly (if present) is influenced by time-varying market risk. We analyse both the aggregate market returns and sector-level returns.

The rest of the study is organised as follows: Section 2 reviews relevant theoretical and empirical literature. Section 3 describes data and methodology employed. In Section 4, the results are presented and analysed and finally Section 5 concludes the paper.

Click here to download full article in PDF format

:: Back to top ::

Download PDF
 

Copyright & Disclaimer, SAIFM. All Rights Reserved. Editorial Committee: Ms Karin van Wyk, Mr Adam Reeves and Mr Tim Howse Designed & Developed by [ Live Q ]