Egyptian Inflation Forecasting: A Performance Comparison Between ARIMA and Chronos-Bolt Models

This paper presents a rigorous empirical comparison between two different paradigms of forecasting, the classical Autoregressive Integrated Moving Average (ARIMA) model and the modern hybrid Chronos-Bolt algorithm, for the purpose of forecasting Egypt's annual inflation rate for the period 1960-2024. An ARIMA specification was established based on the Box-Jenkins approach and the Chronos-Bolt model was implemented based on its standard multi-component decomposition framework. We used a full set of error measures (MSE, MAE, RMSE, MAPE) on a validation sample to measure performance. The results show that the ARIMA model is far better than expected, with much fewer forecast errors across all criteria. The ARIMA model had a RMSE (12.41) lower than half of the Chronos-Bolt model (22.45) and its MAPE (85.98) was considerably lower than the Chronos-Bolt’s (146.11). The ARIMA models indicate that Egypt’s economy will continue to experience instability and significant inflation up to 2029, with an average inflation rate of about 13%. This suggests that inflationary pressures are still robust even when the values are stabilized during the forecast period. This research offers evidence-based recommendations for policymakers, produces more accurate short- to medium-term inflation forecasts for Egypt over a complex hybrid alternative, emphasizing the critical importance of choosing models that are appropriate for specific economic contexts.