Previous work on the empirical distribution of security returns has found that equity returns are not normally distributed. These findings have brought the applicability of certain asset allocation and pricing frameworks into question. This study examines whether the removal of a priori macro-outliers and micro-outliers leads to improved fits to the Gaussian distribution for single-listed equities on the Johannesburg Stock Exchange (JSE). Single-listed equities refer to stocks (i) listed on the JSE Main Board over the period covered in this study, (ii) that comprise of the exchange’s largest 100 stocks by market capitalisation, and (iii) have been determined, by comparing American Depository Receipt (ADR) trading volume to JSE trading volume, to be mainly exposed to the South African market. Regarding the predetermined outliers, the study categorises macro-outliers as days related to predictable market announcements which are US nonfarm payrolls announcement days. Similarly, micro-outliers are classified as days linked to predictable sector-specific and firm-specific news, which are sectoral announcement, and company earnings announcement days, respectively. The study aims to contribute to the empirical and theoretical literature on the distributional properties of South African equity returns. This study makes use of a filter to narrow the sample of stocks for empirical investigation over the period from 1 January 2016 to 31 December 2017, and analyses daily stock returns on a 65-day rolling basis. Using only those equities, an evaluation of the goodness-of-fit methodology is conducted using graphical methods, and statistical goodness-of-fit tests sorted into (i) empirical distribution function, (ii) regression and correlation, and (iii) moment tests. It is found that the majority of the data exhibits significant departures from normality in empirical distribution function, and regression and correlation tests. The results were statistically significant at three confidence levels. However, in the case of moment tests, the results show a clear divergence between the methods. It is further demonstrated that while the daily stock returns have improved fits to the normal distribution, they remain predominantly positively-skewed and thick-tailed even after the removal of the a priori outliers. On this basis, it is argued that some downside risk measures, and asset allocation frameworks may not be applicable in the South African context.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:uct/oai:localhost:11427/31721 |
| Date | 30 April 2020 |
| Creators | Beau, Thabiso |
| Contributors | Holman, Glen |
| Publisher | Faculty of Commerce, Department of Finance and Tax |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Master Thesis, Masters, MCom |
| Format | application/pdf |
Page generated in 0.0016 seconds