ESG Measurement and Ratings Divergence: A Cross‐Jurisdictional Review of Institutional, Stakeholder, and Digital Accountability Perspectives
通过系统综述欧美中三地证据，构建跨国问责框架，区分ESG评级分歧中的合理多元性与可避免噪声，并探讨AI工具在减少噪声与引入新风险中的作用。
ABSTRACT ESG ratings for the same firm‐year often diverge, shaping capital allocation and accountability. Drawing on a structured systematic‐narrative hybrid review, I synthesize evidence across the European Union, the United States, and China and develop a transnational accountability framework that traces divergence through the measurement pipeline. I distinguish context‐relevant pluralism—legitimate divergence rooted in heterogeneous institutional logics, materiality rules, and assurance regimes—from avoidable noise arising from opaque provider choices (coverage, weighting, and controversy treatment) and ungoverned digital analytics. A digital accountability perspective shows that AI‐enabled tools can reduce noise via triangulation and verification yet introduce second‐order opacity through model bias, version drift, and weak governance. I highlight the economic “dark side” of divergence—information‐risk premia, diluted stewardship, and incentives for symbolic compliance—and position ESG controversies and circular‐economy metrics as validity stress tests. I conclude with a research agenda emphasizing multi‐provider replication, mechanism‐based theorizing, and auditable digital assurance.