Are Frontier African Markets Inefficient or Adaptive? Application of Rolling GARCH Models

Abstract

Time-varying calendar anomaly is thinly investigated in frontier stock markets. This study evaluates the day-of-the-week (DOW) calendar effects within the adaptive market hypothesis framework in frontier African stock markets. The study applies rolling analyses of the various GARCH family models to estimate daily stock indices return of Ghana stock exchange, Nairobi securities exchange, Botswana stock exchange and Bourse Regionale des Valeurs Mobilieres (BRVM) for 2000:1-2020:6 periods. The results show changing DOW effects in Kenya and Botswana which is consistent with the AMH. However, DOW effects cannot be validated in BRVM and Ghana. It suggests that each market must be treated with their own peculiarity even though they are ranked as frontier markets. We conclude that the changing DOW effects in the AMH context cannot be generalised in the frontier African markets and the existence of DOW effects must be treated with caution in BRVM and Ghana.

Keywords

• calendar effect
• frontier markets
• garch
• adaptive market hypothesis
• market efficiency
• rolling window.