Modelling and forecasting monthly Brent crude oil prices: a long memory and volatility approach

Abstract The Standard Generalised Autoregressive Conditionally Heteroskedastic (sGARCH) model and the Functional (fGARCH) were applied to study volatility of Fractionally Integrated Moving Average (ARFIMA) model, which is primary objective this study. other goal paper expand on researchers’ previous work by examining long memory volatilities simultaneously, using ARFIMA-sGARCH hybrid comparing it against ARFIMA-fGARCH model. Consequently, models configured with monthly Brent crude oil price series for period from January 1979 July 2019. These datasets considered as global economy currently facing significant challenges resulting noticeable volatilities, especially in terms prices, due outbreak COVID-19. To achieve these goals, an R/S analysis was performed aggregated variance Higuchi methods test presence dataset. Furthermore, four breaks have been detected: 1986, 1999, 2005, 2013 Bayes information criterion. In further section paper, Hurst Exponent Geweke-Porter-Hudak (GPH) used estimate values fractional differences. Thus, some ARFIMA identified AIC (Akaike Information Criterion), BIC (Schwartz Bayesian AICc (corrected AIC), RMSE (Root Mean Squared Error). result, following conclusions reached: ARFIMA(2,0.3589648,2)-sGARCH(1,1) ARFIMA(2,0.3589648,2)-fGARCH(1,1) under normal distribution proved be best models, demonstrating smallest criteria. calculations conducted herein show that two are same accuracy level value, equals 0.08808882, result distinguishes our conclusion, can predict prices more accurately than others.