Effects of Price Clustering on African Stock Markets

Abstract

The phenomenon of price clustering refers to the empirical finding that some prices in financial markets occur significantly more frequently than others. The phenomenon is important theoretically as it challenges the efficient market theory and empirically as it suggests that predictability patterns can be used by investors to devise strategies and investments capable of generating abnormal returns. In this paper, we study the phenomenon for the first time in the context of African markets. Our study includes data from the period spanning 2018-2022 for the stock markets of Egypt, Kenya, Morocco, Nigeria, South Africa, and Tunisia. Our results provide compelling evidence of price clustering within all markets under analysis. Univariate analysis confirms widespread clustering, particularly favoring closing prices ending in zero and five. The results of the multivariate analysis suggest that stocks with higher prices, lower turnover, and lower liquidity tend to exhibit a higher level of clustering. Contrary to the expectations of the Panic Selling Hypothesis, a more intense clustering did not occur during the COVID-19 pandemic. Collectively, our results offer partial support for the Attraction Hypothesis and the Negotiation/Price Resolution Hypothesis.