Has China's national big data comprehensive pilot zone reduced corporate carbon emissions? A quasi-natural experiment based on double machine learning

This research investigates how the National Big Data Comprehensive Pilot Zone (NBDCPZ) affects corporate carbon emissions in China, using a sample of A-share listed industrial companies from 2011 to 2022. Within a Double Machine Learning (DML) modeling framework, we find that the NBDCPZ policy significantly reduces corporate carbon emissions, and the finding holds after several robustness checks. Mechanism analysis indicates that the policy achieves this reduction by enhancing corporate innovation capability and promoting digital transformation. Further exploration of the mechanisms demonstrates that it reduces corporate carbon emissions through two key channels: (1) stimulating incremental innovation activities within firms and (2) promoting the adoption of cloud computing, big data, and artificial intelligence technologies. Media attention strengthens the policy's effect on carbon emissions, with negative media coverage playing a particularly critical role. Heterogeneity analysis reveals that the carbon reduction effect varies across firm characteristics, including life cycle stage, production factor dependency, and pollution level. These findings underscore the role of state-led digital institutional innovation in facilitating a synergistic coordinated transition and offer valuable insights for designing targeted, multi-stakeholder environmental policies.