Financial inclusion and threshold effects in environmental degradation through energy poverty: evidence from Sub-Saharan Africa

Purpose This study aims to investigate the complex and nonlinear relationship between financial inclusion (FI), energy poverty (EP) and environmental degradation (ED) in sub-Saharan Africa (SSA) over the period 2010–2022. While FI is widely recognized as a driver of economic development, its environmental implications remain ambiguous, particularly in regions characterized by limited energy access and structural dependence on carbon-intensive energy systems. To address this issue, this paper adopts an integrated analytical framework combining structural equation modeling (SEM) with advanced machine learning techniques, including double machine learning (DML) and causal forest estimation. This hybrid approach allows for the simultaneous assessment of direct, indirect and heterogeneous effects, while capturing nonlinearities and regime-specific dynamics in the FI–EP–ED nexus. Design/methodology/approach The analysis uses annual data from 41 SSA countries over the period 2010–2022. A hybrid framework is adopted, combining SEM with advanced machine learning techniques, including Random Forest, DML, Causal Forest and SHAP-based interpretability. Threshold regression is also applied to identify regime shifts in the FI–ED relationship according to levels of energy access. Findings SEM results show that FI significantly reduces CO2 emissions both directly and indirectly through reduced EP, confirming EP as a key transmission channel. Machine learning methods reveal heterogeneous and nonlinear effects: FI reduces ED in countries with better energy access but may worsen it in regions with high EP. SHAP analysis identifies GDP and trade openness as the most influential predictors, while FI, EP and renewable energy exhibit strongly context-specific effects. Threshold regression confirms a structural regime shift in the FI–ED link depending on energy access levels. Practical implications These findings underscore the dual and conditional nature of FI in shaping environmental outcomes. They suggest that, in the absence of supportive energy and environmental policies, expanding access to financial services may unintentionally intensify environmental pressures. Conversely, when aligned with clean energy investments and green financial instruments, FI can serve as a powerful lever for sustainable development. Originality/value This study contributes to the literature by explicitly modeling EP as a mediating channel, identifying regime-dependent effects and integrating econometric and machine learning approaches to better capture the complexity of the finance energy environment nexus in SSA.