Market efficiency of African stock markets: A Meta-Analysis

Investments on African stock markets have grown over the years. This is because stock markets in Africa have realized increase in market capitalization, membership, value and volume traded. Notwithstanding these increases, empirical studies on informational efficiency have had mixed conclusions about the markets. Are these mixed conclusions a result of variations in study characteristics? This study aims to find out if study characteristics have a probability of concluding efficiency of African stock markets. We quantitatively reviewed previous studies of informational efficiency of African stock markets by means of meta-analysis. We employed the mixed effect logistic meta-regression model to examine which of the study characteristics is significant in concluding a market to be efficient. The mixed effect logistic model was chosen because it contains both fixed effects and random effects. The model explains an outcome as a linear combination of fixed and conditional random effects. The fixed effects assume equal influence of explanatory variables on an outcome whilst the random effects assume variations amongst observations when analyzing relationships between an outcome and explanatory variables. Our results showed that only the indicator for publication bias is significant at the 5% level and that none of the study characteristics is significant in concluding efficiency of African stock markets. The indicator for publication bias being significant means our analysis suffers from publication bias. This implies that there has been a change in attitude in recent years towards studies on informational market efficiency whose results do not validate the Efficient Market Hypothesis (EMH) unlike the earlier years when the EMH was formulated and acclaimed to be one of the best propositions in economics. Though, none of the study characteristics was statistically significant in efficiency conclusions, we observed that stock markets in Africa are about 4 times more likely to conclude that the markets are efficient if the study tested the weak-form rather than the semi-strong form of the EMH. The results have important implications in that, efficiency conclusions of African stock markets do not depend on any of these study characteristics. Therefore, traders cannot devise strategies to outperform the markets based on any of these study characteristics. We therefore conclude that informational efficiency conclusions of African stock markets are context-specific which occurs randomly and not based on any of the study characteristics.