Modeling the association between illiteracy and poverty in Egypt: a comparative analysis of linear regression and ARDL approaches

This study examines the dynamic relationship between illiteracy and poverty in Egypt over the period 1990–2023 using annual time-series data. The analysis applies both a simple linear regression model and the Autoregressive Distributed Lag (ARDL) framework to distinguish between short-run dynamics and potential long-run relationships. While the linear regression results indicate a statistically significant association between illiteracy and poverty, the ARDL (1,2) specification provides a more appropriate framework for capturing temporal adjustments and dynamic interactions. The ARDL bounds test yields inconclusive evidence regarding long-run cointegration, suggesting that the existence of a stable long-run equilibrium relationship cannot be confirmed with certainty. Consequently, long-run estimates are interpreted as indicative of potential sustained effects rather than definitive equilibrium outcomes. In contrast, the short-run results reveal statistically significant and cumulative effects of illiteracy on poverty. The error correction term is negative and statistically significant (ECM (− 1) = − 0.199), indicating that approximately 19.9% of short-term deviations from the long-run path are corrected each year. Diagnostic and stability tests confirm the robustness and validity of the estimated model. Overall, the findings highlight the importance of short-run dynamics and emphasize education as a critical policy instrument for poverty reduction and sustainable development in Egypt. By focusing on illiteracy as a key determinant, this study provides new empirical insights into the dynamic education–poverty nexus in a developing country context.