Based on the panel data of 30 provinces in China from 2011 to 2021, this paper uses the Super-EBM-undesirable model to measure the energy carbon emission efficiency (ECEE). On the basis of analyzing the spatiotemporal characteristics of ECEE, the Tobit regression model and the mediation effect method are used to empirically test how digital inclusive finance (DIF) affects the ECEE. The main research results reveal that 1) the ECEE of China shows a fluctuating trend during the study period, with the highest point in 2013 (0.529), and the lowest point in 2021 (0.501). There are significant differences in ECEE among different regions, with the highest in the eastern region, followed by the central and the western regions, and the lowest in the northeast region. 2) According to the empirical research results, the DIF can significantly promote the improvement of ECEE; technological innovation and technological application play a significant intermediary role in the impact of DIF on ECEE. This conclusion is still valid after robustness tests. 3) As for the control variables, the development of infrastructure has a significant inhibitory effect on ECEE, while China’s openness to the outside world and economic urbanization have enormous promoting effects on ECEE. The research results could provide reference for formulating national digital finance related policies and establishing a financial support system for the strategy of “carbon peak and carbon neutrality”. 

基于2011—2021年中国30个省、直辖市和自治区(西藏自治区、台湾省、香港和澳门特别行政区数据暂时缺失)的面板数据, 采用考虑非期望产出的Super-EBM-undesirable模型, 测度能源碳排放效率。在分析能源碳排放效率时空特征的基础上, 构建Tobit回归模型, 运用中介效应方法, 实证检验数字普惠金融如何促进能源碳排放效率的改善, 得到如下结果。1) 中国能源碳排放效率在研究时段呈现小幅度波动变化趋势, 2013年为最高点(0.529), 2021年为最低点(0.501)。各地区能源碳排放效率存在显著差异, 东部地区最高, 中部和西部地区次之, 东北地区最低。2) 数字普惠金融发展水平的提高能够显著地促进能源碳排放效率的改善。其中, 技术创新水平和技术应用水平在数字普惠金融对能源碳排放效率的影响过程中起到显著的中介作用, 并且经过一系列稳健性检验后, 该结果仍然成立。3) 从控制变量看, 基础设施水平对能源碳排放效率起到显著的抑制作用, 而对外开放水平和经济城镇化水平起到显著的促进作用。研究结果可为制定国家数字金融相关政策, 建立“碳达峰和碳中和”发展战略的金融支持体系提供参考。