ESG and banking financial performance: A multi-level meta-analysis review

Purpose —This study addresses the longstanding debate concerning the relationship between Environmental, Social, and Governance (ESG) engagement and financial performance within the banking industry by systematically synthesizing the existing empirical evidence. Design/Methodology/Approach —A comprehensive literature review identified 28 peer-reviewed studies, yielding 194 distinct effect sizes. Employing a three-level random-effects meta-analytic framework, we estimate the overall ESG-performance association besides ESG subdimensions (E, S, G), and examine key moderating factors, including data sources, financial performance measures, and the macro-regulatory period (pre- versus post-Paris Agreement). Findings —The results indicate a small yet statistically significant positive association between ESG engagement and banking financial performance. The environmental pillar exhibits a consistent positive relationship, whereas the social and governance dimensions do not demonstrate statistically significant independent effects. Return on equity appears more sensitive to ESG variation relative to alternative performance indicators. Additionally, discrepancies across ESG rating providers reveal meaningful measurement heterogeneity. Temporal analysis further suggests that the ESG–performance association strengthened in the post-Paris Agreement regulatory period. The findings remain robust to publication-bias diagnostics. Originality/Value —While corroborating evidence from broader corporate ESG research, this study represents, to our knowledge, the first meta-analysis focused exclusively on the banking sector. Methodologically, it demonstrates the advantages of a three-level meta-analytic framework for addressing statistical dependencies among multiple effect sizes derived from individual primary studies. By decomposing aggregate ESG into its environmental, social, and governance components, evaluating the moderating role of alternative financial metrics (ROA, ROE, and Tobin’s Q), assessing divergence across data providers, and incorporating a temporal analysis of the regulatory shift following the 2015 Paris Agreement, this study advances the literature and offers policy-relevant implications for scholars, practitioners, and regulators.