Smarter and cleaner: How does energy digitalization affect carbon productivity?

Digitalization is a driving force behind the ongoing energy industrial revolutions, catalyzing China's pursuit of carbon neutrality and sustainable development. Leveraging provincial data and annual reports from energy enterprises in China, this study constructs a comprehensive analytical framework that encompasses benchmark regression models, mediating effect models, threshold models, and spatial econometric models. These models are utilized to investigate the multi-faceted impacts of energy digitalization on carbon productivity (CP). The aim is to furnish micro-level evidence and policy guidance for advancing energy transformation and fostering low-carbon development enriched with digital elements. This research employs natural language processing and machine learning techniques to compute an Energy Digitalization Index, examining two critical dimensions: digital industry investment and the inclination toward digital transformation. The following key findings emerge: firstly, energy digitalization (ED) exhibits a statistically significant ability to enhance regional CP, a phenomenon marked by temporal and regional variations. Secondly, the analysis confirms the transmission mechanisms associated with energy technology innovation, energy structure, and energy utilization efficiency, as revealed through the Logarithmic Mean Divisia Index (LMDI) decomposition method. Furthermore, the optimal effect of energy digitalization on low-carbon economies materializes in settings characterized by mature market conditions, modest environmental regulations, advanced digital infrastructure, and reduced resource dependency. Additionally, the spatial Markov chain analysis unveils a conspicuous spatial distribution pattern termed "club convergence" in regional CP, accompanied by a pronounced "Matthew effect." According to the spatial Durbin model, energy digitalization generates favorable spatial spillover effects, primarily in peripheral regions, with a more pronounced short-term influence. Building upon these insights, this paper presents pertinent policy recommendations encompassing the national "digital energy" strategy, regional differentiation policies, and initiatives to stimulate digital technology innovation among enterprises. Our findings furnish robust empirical evidence and constructive policy insights, empowering governments to forge a smarter and cleaner energy ecosystem. Furthermore, these findings offer valuable guidance for other developing nations seeking to implement effective digital strategies.