Assessing the Impact of <scp>ESG</scp> Practices on Financial Performance and Market Valuation: Heterogeneity Analyses Across Regions, Income Levels, and Industries

ABSTRACT This study examines how environmental, social and governance (ESG) practices relate to corporate financial performance and market valuation across regions, income groups and industries. Using an unbalanced panel of 2157 listed firms (17,247 firm‐year observations) from 2016 to 2023, the analysis combines firm fixed‐effects models, fixed‐effects instrumental‐variable (IV) estimator and two‐step System Generalized Method of Moments (System GMM) robustness tests. In the baseline fixed‐effects models, ESG is not significantly associated with return on assets (ROA) or Tobin's Q. In the IV models, however, ESG is positively associated with ROA at the 10% level and with Tobin's Q at the 1% level. However, in the preferred two‐step System GMM specifications, lagged ESG is not significantly associated with either ROA or Tobin's Q once dynamic persistence is modelled explicitly, although the Hansen and Arellano–Bond diagnostics support the reported specifications. Regional heterogeneity is statistically significant for both outcomes, with the strongest valuation effect concentrated in North America. Income‐group heterogeneity appears for Tobin's Q but not for ROA. ESG is positively associated with valuation in high‐income and lower‐middle‐income economies, whereas the upper‐middle‐income estimate is insignificant. Industry evidence shows that ESG is weaker for ROA in environmentally sensitive industries and stronger for ROA in customer proximity industries, but valuation effects are not statistically stronger than in the rest of the sample once formal interaction tests are applied. Overall, the findings indicate that the estimated ESG–performance relationship is sensitive to the identification strategy. Positive contemporaneous IV associations are strongest for market valuation, but the dynamic panel evidence does not confirm a robust lagged ESG effect.