Calendar Effect Anomalies and Stock Returns Volatility in African Markets: Evidence From Ghana and Nairobi Stock Exchanges

Abstract

his study examines calendar effect anomalies, particularly Day-of-the Week effect and Month-of-the-Year effect and stock returns volatility in Ghana Stock Exchange (GSE) and Nairobi Stock Exchange (NSE). Daily closing prices indices from the two stock markets for the period 2005 to 2014 was used. Using an Ordinary Least Square (OLS) regression with autoregressive term, the findings provide no evidence of day-of-the-week effect for GSE but there exist Friday effect for NSE. However, the study provides no evidence of month-of-the-year effect anomaly in either NSE or GSE. The study also documents that daily returns could be predicted but monthly returns cannot be predicted in NSE. On the contrary, the findings indicate that whiles daily returns are difficult to predict monthly returns can be predicted using past price and returns information in GSE. Furthermore, Generalised Autoregressive Conditional Heteroskedastic GARCH (1, 1) Threshold GARCH (1, 1) and Exponential GARCH (1, 1) were employed to examine stock returns volatility. The results of the GARCH model suggest a high degree of persistent in the conditional volatility of daily and monthly stock returns in NSE. The TGARCH, and EARCH models show significant evidence for asymmetry (leverage effect) in monthly stock returns but no evidence of asymmetry in daily returns was found in NSE. However, there was no evidence of conditional volatility for Ghana Stock Exchange Composite Index (GSE-CI). The study concludes that GSE and NSE are inefficient markets. It is recommended that months are irrelevant in making investment decisions in GSE or NSE but days of the week are relevant in NSE only