Effects of regional integration, crisis response financing and official development assistance on environmental quality in developing economies with the moderating role of renewable energy consumption

Environmental sustainability remains a major challenge for developing economies amid deepening regional integration and increasing reliance on external financing. This study examines the relationships between regional integration, crisis response financing, official development assistance, and environmental quality, while assessing the moderating role of renewable energy consumption across 65 developing economies from 2000 to 2024. Using a balanced panel of 1625 country-year observations, the analysis combines two-way fixed effects with Driscoll–Kraay standard errors, Cross-Sectionally Augmented Autoregressive Distributed Lag (CS-ARDL) estimation, Hansen's panel threshold regression, Panel Corrected Standard Errors (PCSE), Feasible Generalised Least Squares (FGLS), and XGBoost with SHAP for predictive validation and interpretation. The baseline estimates indicate that regional integration is positively associated with carbon emissions (β = 0.408, p = 0.007), whereas official development assistance (β = − 0.0076, p < 0.001) and renewable energy consumption (β = − 0.0183, p < 0.001) are associated with lower emissions. Crisis response financing exhibits a positive baseline association (β = 0.0013, p = 0.030), but the CS-ARDL estimates are negative in both the long run (β = − 0.0021, p = 0.020) and short run (β = − 0.0017, p = 0.005), demonstrating that its environmental relationship varies across specifications and time horizons. Renewable energy consumption also conditions the integration–emissions relationship: the baseline interaction is negative (β = − 0.0084, p = 0.002), while the CS-ARDL results reveal a negative short-run but positive long-run interaction. Threshold analysis identifies a statistically significant, sample-specific renewable-energy breakpoint of 39.84%, above which the positive association between regional integration and emissions becomes substantially smaller and statistically weaker. XGBoost achieves an R 2 of 0.817, with SHAP identifying renewable energy consumption as the most influential predictor. The study contributes by constructing a composite Regional Integration Index integrating trade openness and regulatory quality and by combining complementary econometric inference with explainable machine-learning prediction. The findings underscore the importance of coupling regional integration and development financing with renewable-energy expansion, environmental safeguards, and environmentally responsible allocation of external financing.