Hybridization Methodology of ARMA-FIGARCH Model to Examine Gasoline Data in Iraq

The reason beyond this research is to investigate the long memory with volatility impact embedded in the daily fuel prices (Gasoline) via time-series behavior. Gasoline returns are assumed to follow ARMA-FIGARCH models. Here, a hybrid methodology is introduced in two main steps. Firstly, the results of the estimation of mean value have been achieved by using ARMA models, while the second step is to estimate the conditional variance value by using FIGARCH models. Among these formulating, the final step will be done by combining the previous two steps that yield ARMA-FIGARCH. Particularly, AR(2)-FIGARCH (1,d,2) model will be yielded under the normal distributional assumption of residuals, which indicated a better fit for price volatility of gasoline. Non-Gaussian residuals are also assumed by using student-t distribution. Moreover, AR(2)-FIGARCH (1,d,2) had been selected significantly for daily returns and was preferred due to its success in passing the goodness-of-test fit.