How do data assets mitigate high-pollution corporate carbon emission intensity?
Аннотация
Reducing carbon emission intensity in high-pollution corporations is essential for promoting sustainable development and achieving China’s climate commitments. Based on panel data of Chinese listed corporations, this study constructs a double machine learning model to systematically examine the causal impact and underlying mechanisms through which data assets influence the carbon emission intensity of high-pollution corporations. Empirical results demonstrate that the rational allocation of data assets significantly reduces carbon emission intensity, and this finding remains robust across multiple tests. Mechanism analysis further reveals that data assets achieve this effect by increasing the proportion of highly educated talent, strengthening green innovation, and reducing the bullwhip effect. Stronger ESG performance and broader regional broadband internet access positively moderate the impact of data assets on reducing carbon emission intensity in high-pollution corporations. To further identify the key factors influencing carbon emission intensity and provide insights for optimizing data asset strategies, this study applies SHAP analysis based on a machine learning framework. The results highlight fixed asset ratio, total assets, profitability, number of employees, proportion of highly educated talent, green innovation, asset-liability ratio, and bullwhip effect as the most influential determinants. Further synergistic mechanism analysis reveals that the emission-reducing effect of data assets varies across corporations. High-pollution corporations with higher fixed asset ratios, larger total assets, stronger profitability, lower asset-liability ratios, and larger workforce sizes benefit more from data assets in reducing carbon emissions. These findings clarify the mechanisms through which data assets facilitate low-carbon transformation among high-pollution corporations and provide theoretical guidance for advancing sustainable development.