Effect of Calendar Anomalies on Stock Price Volatility using GARCH Models: Evidence from Nairobi Securities Exchange

Abstract

Many researchers have shown that the financial market of the Kenyan economy is weakly inefficient in terms of calendar anomalies. The inefficiency in the market may be explained by the volatility of the stock returns. This study was conducted with the main objective being to establish the effect of calendar anomalies on stock price volatility using the GARCH (1,1) model. The main characteristics of the Kenyan financial market are briefly explained. The empirical analysis is done on the transformed daily returns data of the NSE 20 share index. A comparison between the OLS model and the GARCH model is done on the return equation. The day of the week effect is found to be significant in both models where Friday had the highest returns while Monday had the lowest returns but January effect is only explained in the OLS model while the GARCH model does not show any presence of the January effect. The coefficients of the volatility equation are positive, significant and their summation is less than one indicating that the volatility is persistent and is mean reverting, that is, there is a normal volatility that the volatility should return to no matter what. In conclusion the variance in the calendar anomalies studied is time varying.

Keywords: GARCH (1,1), EMH, Financial market anomalies, mean reversion