Digital Transformation and Corporate ESG Performance: The Moderating Roles of Innovation Networks and Analyst Digital Attention

Using panel data on Chinese A-share-listed firms on the Shanghai and Shenzhen stock exchanges from 2010 to 2023, this study employs Word2Vec word embeddings and latent Dirichlet allocation (LDA) topic modeling to identify digital technology content in firms’ invention patent abstracts, measures corporate digital transformation in terms of digital technology innovation output, and examines its nonlinear relationship with corporate ESG performance and the boundary conditions of this relationship. The results show a significant inverted U-shaped relationship between digital transformation and corporate ESG performance. The moderating effects of interfirm technological innovation networks differ across centrality measures: weighted out-degree does not exhibit a significant moderating effect, whereas weighted in-degree provides only marginal evidence of a rightward shift in the turning point. A higher level of analyst digital attention is associated with more pronounced concavity of the curve, but the locations of the turning points do not differ significantly across attention levels. Further analysis indicates that this nonlinear relationship differs across data asset information disclosure statuses. The social dimension exhibits a relatively robust inverted U-shaped relationship, evidence for the inverted U-shaped relationship in the environmental dimension is only marginal, and the governance dimension exhibits a marginally significant pattern of initial decline followed by an increase. This study provides empirical evidence for understanding the stage-dependent relationship between digital transformation and corporate ESG performance and the boundary conditions of this relationship.