Revisiting the Environmental Kuznets Curve in West Africa: A spatial econometric analysis of growth, globalization, and emissions.

This study re-examines the Environmental Kuznets Curve (EKC) hypothesis for 16 West African countries from 1994 to 2023, mainly employing spatial econometric models that account for cross-border spillovers. Unlike previous EKC studies that consider countries as isolated units, this paper incorporates spatial dependence through the Spatial Durbin Model (SDM) and validates long-run dynamics using Fully Modified OLS (FMOLS) and Canonical Cointegration Regression (CCR) methods, thereby contributing to the EKC discussion. The findings indicate that CO2 emissions are significantly influenced by spatial spillovers, confirming strong regional interdependence, as evidenced by Moran I values of 0.084 (z-value = 2.291), 0.080 (z-value = 2.278), and -0.18 (z-value = -1.81) for the years 2006, 2015, and 2020, respectively. Locally, GDP and its square reduce emissions by 0.46 and 0.55, respectively; both have positive and significant long-run effects of 0.66 and 0.55, thereby rejecting the EKC hypothesis for West Africa. Industrial value added exacerbates emissions, whereas globalization, through foreign direct investment (FDI) and trade, consistently mitigates them, supporting the "pollution halo" hypothesis. Heterogeneity analysis reveals divergent emission trajectories: low-income countries benefit from clean energy spillovers, while lower-middle-income countries experience emission-intensive growth. These findings highlight the originality of applying a spatially grounded and income-sensitive approach to EKC testing in Africa. The policy implications suggest the need for ECOWAS-led regional climate governance, green industrial transformation, and targeted FDI and trade strategies to support a low-carbon transition. By integrating spatial dynamics and structural differences, this paper challenges the universality of the EKC and provides evidence-based pathways for sustainable growth in West Africa.