Abstract
Environmental, social, and governance (ESG) performance has become a strategic priority for firms seeking sustainable competitiveness. However, the mechanisms through which research & development investment contributes to ESG performance remain insufficiently understood. This study investigates whether green innovation mediates the relationship between research & development investment and ESG performance using an unbalanced panel of 1,444 Chinese A-share listed firms from 2007 to 2022. Fixed-effects models, instrumental variable estimation, and system generalized method of moments are employed to address unobserved heterogeneity and endogeneity. To further strengthen the robustness of the findings, Bayesian regression, Bayesian Additive Regression Trees, random forest analysis, bootstrap mediation tests, and quantile regression are incorporated as complementary analytical approaches. Our results indicate that research & development investment significantly promotes both green innovation and ESG performance, while green innovation exerts a positive effect on ESG performance and partially mediates the impact of research & development investment. The Bootstrap analysis confirms the statistical significance of the indirect effect. Bayesian inference and machine learning analyses provide consistent evidence supporting the robustness of the estimated relationships. The Quantile regression further reveals substantial distributional heterogeneity, with the contribution of green innovation being more pronounced among firms with relatively lower ESG performance. Additional heterogeneity analyses demonstrate that the complementarity between research & development investment and green innovation is stronger in heavily polluting industries than in non-heavily polluting industries. This study extends the literature by integrating conventional econometric methods with Bayesian inference and machine learning techniques to provide more comprehensive evidence on research & development–green innovation–ESG nexus. Our findings offer important managerial and policy implications for enhancing corporate sustainability through innovation-driven strategies.

This publication has 26 references indexed in Scilit: