INVESTIGATING SOME MACRO ECONOMIC VARIABLES IN NIGERIA USING ARIMAX AND ARIMA MODELS

This study investigates the behavior of inflation in Nigeria by modeling and forecasting its dynamics using Autoregressive Integrated Moving Average (ARIMA) and its extension with exogenous variables (ARIMAX) from 1990 to 2023. Inflation is modeled alongside key macroeconomic indicators: exchange rate, interest rate, and unemployment rate. Time series techniques were employed, including unit root testing, transformation using natural logarithms, and fitting of optimal ARIMA and ARIMAX models. The ARIMA (0, 0, 1) and ARIMAX (0,0,1) models were identified as best-fitting models for inflation forecasting. Although both models showed statistical adequacy with normally distributed residuals and no significant autocorrelation, ARIMA outperformed ARIMAX in terms of in-sample forecast accuracy with lower RMSE and MAE values. However, the ARIMAX model provided insights into the role of unemployment as a significant negative predictor of inflation. This study concludes that while ARIMA provides better short-term forecasts, ARIMAX offers a richer understanding of the inflation process by incorporating macroeconomic variables. These findings offer valuable input for monetary policy planning and economic modeling in Nigeria.