Science and Technology Finance Policy and Urban Energy Efficiency: Evidence from the Double/Debiased Machine Learning Model

Evaluating the impact of science and technology finance policy (STFP) on urban energy efficiency (UEE) is essential for achieving carbon neutrality and fostering urban sustainability. In this study, UEE incorporating carbon emissions is measured using stable nighttime light data and global 0.1° × 0.1° carbon emission grid data. Taking “Promoting the Integration of Science and Technology with Finance” pilot policy as an exogenous shock, a double/debiased machine learning (DDML) model is applied to investigate STFP’s influence on UEE, using data covering 266 Chinese cities during 2003–2019. Findings show that STFP significantly increases UEE by approximately 9.54% relative to the sample mean. Mechanism analysis indicates the effect operates via green innovation, advanced industrial structure, and fintech development. Heterogeneity analysis reveals stronger policy effects in non-old industrial base cities, cities with high climate policy uncertainty, and non-resource-based cities. Additionally, STFP exhibits spatial spillover effects on UEE that are generally positive within a 40 km radius but turn negative as the spatial scale expands, reaching the maximum negative magnitude within a 2800 km radius. These findings provide valuable insights for optimizing financial policy design and advancing urban low-carbon transformation.