Abstract
Abstract : African stock markets have always received less attention in the literature. Regardless of being deemed small, segmented and illiquid, it will be useful to global investors and interested stakeholders to investigate the dynamics in these stock markets. Given this background, this study investigates the relationship that exists between the South African stock market and selected African stock markets. The study utilizes monthly data between February 2002 and July 2018. Monthly data is deemed appropriate as opposed to daily data which contains too much noise. To establish if long run equilibrium exists between these stock markets the study employs the Johansen test within a Vector Autoregression (VAR) framework. For subsequent analysis, the linear Vector Error Correction Model (VECM) and the Markov-Switching VECM (MS-VECM) are employed to explain the long run relationship between the South African stock market and the selected African stock markets. The Granger causality test establishes if any causal links exist among African stock markets. Results indicate that cointegration indeed exists between South Africa and selected stock markets, although in a weak form. A comparison between the linear VECM and MS-VECM establishes that MS-VECM outperforms the linear VECM as the model’s transition probabilities capture every transition change in the stock market data. A weak form of cointegration suggests that these stock markets jointly offer potential gains for diversified portfolio investments at all time scales. Therefore, international investors should not only target emerging and other international markets, but also these African stock markets if they are to reap long run returns from diversified investment.
M.Com. (Financial Economics)