Modelling and Forecasting Inflation Volatility in Nigeria: A Comparative Evaluation of Symmetric and Asymmetric GARCH-Family Models

Abstract

Volatility persistence remains one of the major macroeconomic challenges particularly in developing markets due to inherent structural adjustment, recurrent economic disturbances, and policy uncertainty which affect the forecasting capability and monetary policy decision making processes. Volatility modeling and forecasts of GARCH family of models (GARCH(1,1), EGARCH(1,1), TGARCH(1,1), GJR-GARCH(1,1) and APARCH(1,1)) are being studied on monthly Nigerian inflation data collected from the National Bureau of Statistics (NBS) between January 2000 to December, 2025. The initial diagnostic tests proved the existence of volatility clustering, non-normal distributional properties and significant conditional heteroskedasticity and therefore nonlinear models of volatility are indispensable. The results have proven significant persistence of volatility with ARCH and GARCH effects across all the different forms of models used. Similarly, the findings using asymmetric models have established different effects of positive and negative shock on volatility of inflation, hence the presence of the need for asymmetry of inflation volatility to be adequately considered. The comparative information criterion and out-of-sample forecast performance confirmed the effectiveness of APARCH(1,1) over the other GARCH models in both in-sample performance and out-of-sample forecasting power as APARCH(1,1) has the capacity to capture both power transformation and asymmetries of inflation volatility. It is concluded that the result emphasizes the significance of modeling volatility asymmetry for a more predictable and stable forecasts of Nigeria inflation that Central Bank of Nigeria should be able to incorporate forecasting tools to use APARCH to get better and more accurate inflation forecast so as to bolster more effective response of monetary policy.