Remittance and financial inclusion as a determinist of energy poverty reduction in Sub-Saharan Africa: evidence from machine learning with Fourier functions

Abstract This study examines the impact of remittances and financial inclusion on alleviating energy poverty in sub-Saharan Africa. Utilizing a Fourier-augmented machine learning framework from 2003 to 2023, the methodology integrates Fourier Toda–Yamamoto causality tests, panel cointegration, and CS-ARDL models across 36 countries. The results confirm that remittances significantly reduce energy poverty by enabling household investments in clean energy, especially where financial inclusion is strong. Financial services, such as credit and savings, are critical mediators for optimizing remittance utilization for sustainable energy infrastructure. The novelty of this study lies in the application of Fourier functions to detect nonlinear patterns and structural breaks, offering a robust estimation approach that overcomes the limitations of conventional econometric models. This pioneering combination of Fourier methods and machine learning enhances causal inferences in dynamic socioeconomic contexts. Policy recommendations include promoting financial inclusion through tailored banking services, incentivizing off-grid renewables, and fostering partnerships between banks and energy providers to align financial tools with the goals of energy development. Graphical Abstract