Firm size and the day of the week effect on the Johannesburg Stock Exchange.

Abstract

The Efficient Market Hypothesis (EMH) asserts that stock prices always entirely reflect all available information and that stock prices follow a random walk, where future stock prices are not predictable based on historical prices (implying stock market efficiency). If the stock market is not efficient, abnormal returns can be realised by beating the stock market through observing and trading on certain patterns (anomalies) exhibited by past stock prices. Various anomalies have been documented, including the Day of the Week (DOW) effect (the tendency of a stock market to exhibit on average low daily returns in the beginning of the week (mostly on Mondays) and high returns towards the end of the week (mostly on Fridays). Examining the DOW effect is particularly interesting, as it demonstrates daily patterns on which investors can take advantage of this anomaly to realise excess returns on daily basis. One of the reasons that has been put forward as to what initiates the DOW effect, is measurement error as when a variable of interest either explanatory or dependent variable has some measurement error independent of its value. Thereby, leading to the notion that the DOW effect is present in medium and small markets or firms with low merchantability (firm size effect). However, from the South African literature, still has a gap about the existence of the DOW effect across firm sizes on the JSE and its cyclical (appearing or disappearing) changes over time. Firstly, the study examined the existence of the DOW effect on the JSE in firm sizes on a full sample (1995 to 2019) utilising daily log-returns. The best-fit models were selected from a family of GARCH models, EGARCH (2, 1) and EGARCH (3, 1) models better fitted the AltX and the large index respectively and TGARCH (3, 1) and TGARCH (1, 1) better fitted medium and small indices respectively. The results showed that the DOW effect exists on the JSE stock exchange in three out of all the four investigated indices (medium, small and AltX except the large), particularly the DOW effect existed more in returns than in the volatility of those returns. Secondly, a rolling window analysis was utilised to examine the changes of the DOW effect over 1995- 2019 where the best-fit model for each sub-period was utilised. The results showed that the existence of the DOW effect is not constant over time concluding a cyclical behaviour (appearing and disappearing in some sub-periods). However, the highest frequency of appearance of the DOW effect appeared in the medium, small and the AltX indices confirming the notion that the DOW anomaly is mostly found in companies with low capitalisation.