Temporal-spatial decomposition and multi-scenario prediction analysis of energy poverty in China

The eradication of energy poverty is crucial for alleviating poverty and transitioning to clean energy. For countries with regional heterogeneity, identifying and predicting the differences in energy poverty and assessing their driving forces at the subnational scale are crucial for achieving the national energy-poverty reduction targets. In this study, comprehensive analytical frameworks were used to explore the spatial and temporal differences and driving forces of macro multidimensional energy poverty (MMEP) in China for 2006–2021. We expanded a temporal-spatial within-between refined Laspeyres index decomposition and used a long short-term memory neural network to conduct multi-scenario predictions of MMEP under various intensities of government intervention. The results indicated an improvement in the MMEP at the sub-national scale in China, with a certain degree of s temporal-spatial heterogeneity that could be attributed to the stability of key indicators within the energy poverty system and the importance of economic regions in reducing energy poverty; we noted varying directions and magnitudes of effects across eight economic regions and 30 provinces. The prediction results show that without additional government intervention, all the subnational scales in China will not surpass their historical best MMEP. Therefore, when implementing policies, governments should consider the within-regional variability of pathways to minimize energy poverty. Our study expands the understanding of the internal driving forces and predictions of MMEP by considering the temporal changes and spatial differences at the macro level. Notably, our study can be used as a reference for policymakers to develop differentiated energy-poverty reduction schemes.