Revisiting the auto-regressive integrated moving average approach to modelling volatility using Bahrain all share index daily returns

Abstract

Most academic researchers in economics and finance have researched the characteristics of stock prices and how it behaves. The widespread belief that understanding the mentioned behaviour and characteristics will provide critical information in forecasting future stock prices fuels the continued interest in creating approaches to improve existing models' predictive power. This study provides a fresh investigation of stock market index volatility utilising Box-Jenkin's auto-regressive integrated moving average (ARIMA) method. The study discovered that ARIMA (1, 1, 4) best simulates Bahrain's stock market index volatility. According to the research, the fitted ARIMA time series' consecutive residuals (prediction errors) were not statistically connected. On the other hand, the residuals are average, having a mean of zero and a constant variance. Moreover, it can be said that the same model is best if used on a weekly forecast horizon, and its ability to model long-term price behaviour, and thus volatility, is still much to be desired.