Does Holidays affect the Returns of Stock Markets Prices? An Application of some GARCH and Prophet Models under Varying Error Distributions

Abstract

Logically speaking, financial time series data such as stock price returns is directly or indirectly affected by some factors such as holidays (Valentine’s Day, Father’s Day, etc.). These effects have not been fully studied by researchers. This paper utilizes GJR-GARCH (with dummy variables as holidays) and Prophet models on Nigeria and Chinese stock price returns so that more realistic results and information on how the holidays affect the returns of stock markets will be obtained. Results of the analysis reveal the existence of holiday’s effects on Thursdays (for NSE) and Mondays and Fridays (for CSE) both under Generalized Error Distribution (GED) assumption of innovations. Additionally, no leverage effect was found in the two returns series. Furthermore, the GJR-GARCH model under all the three considered distributions of innovations on the two data sets shows volatility clustering and persistence. Thereafter, evidence from the minimum RMSE and MAE, the GJR-GARCH model under NSE slightly outperformed the Prophet model, while under CSE, the Prophet model performed better than the GJR-GARCH model.