Analysis of Inflation Behavior in Sudan via Long-Memory Models

Abstract

This study aims to analyze and forecast inflation behaviour in Sudan. The research problem is to investigate the presence of long memory in the inflation rate series. The analysis was done via autoregressive fractionally integrated moving average (ARFIMA)and FIGARCH (fractionally integrated generalized autoregressive conditional heteroskedasticity) using a monthly series covering the period Jan. 2002–Dec. 2020. The study revealed the existence of long memory in the inflation series, where the fractionally integrated coefficient is 0.43. We recommend the use of ARFIMA in analyzing inflation in Sudan. ARFIMA was preferred to FIGARCH based on Akaike's information criteria. It is also recommended to adopt a monetary policy that leads to stabilizing inflation and the economy as a whole, such as implementing interest rate adjustments and controlling the money supply to mitigate inflationary pressures.