An Assessment on the Predictive Performance of Arima-Garch and Arimax-Garch Models on Nigerian Inflation Rate

Abstract

Accurate inflation forecasting is critical for effective economic planning and decision-making in Nigeria, yet it remains challenging due to exchange rate fluctuations, oil price shocks, and shifts in monetary policy. Traditional models such as ARIMA often fail to capture these complex dynamics, particularly during periods of high volatility. This study evaluates the predictive performance of two hybrid time series models, ARIMA-GARCH and ARIMAX-GARCH, using monthly inflation data from 1997 to 2024. While ARIMA-GARCH models the conditional mean and volatility of inflation, ARIMAX-GARCH extends this by incorporating exogenous variables, including the exchange rate, premium motor spirit (PMS) price, and money supply. Stationarity was tested using the Augmented Dickey-Fuller (ADF) test, model selection guided by AIC and BIC, and volatility clustering confirmed with the ARCH-LM test. Forecast accuracy was assessed using RMSE, MAE, and MAPE. Results demonstrate that ARIMAX-GARCH consistently outperforms ARIMA-GARCH, yielding more accurate forecasts. The study recommends ARIMAX-GARCH for inflation forecasting in Nigeria, as it provides robust, reliable predictions that support improved policy and investment strategy.