ESG格付の乖離と企業財務パフォーマンス：測定問題に関するレビュー
ESG Rating Divergence and Corporate Financial Performance: A Review of the Measurement Problem (原題)
(著者不明)
🤖 gxceed AI 要約
日本語
ESGパフォーマンスと企業財務パフォーマンス（CFP）の関係は2,000本以上の研究があるが結果は矛盾している。本論文は、その原因が理論の弱さではなくESG測定の質の低さにあると論じ、主要6格付機関間の相関が平均0.54と低いことを示す。さらに、ESG開示の拡大が格付乖離を広げる可能性を指摘し、IFRSサステナビリティ開示基準やCSRDの限界を評価する。香港市場のH株発行体における国際格付と中国本土格付の乖離も分析する。
English
ESG-CFP research yields conflicting results despite 2,000+ papers. This review argues the issue is poor ESG measurement quality, not weak theory, noting pairwise correlations among six major rating providers average only 0.54. It shows greater ESG disclosure can widen rating divergence, challenging mandated disclosure premises, and evaluates IFRS SDS and CSRD limits. A Hong Kong H-share case illustrates institutionally meaningful divergence.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本ではSSBJ基準の導入と有報でのサステナビリティ開示が進む中、ESG格付の測定誤差が投資家の意思決定や企業の資金調達コストに与える影響を理解する上で重要。特に、開示拡大が格付乖離を広げるという知見は、日本企業が開示戦略を練る際の注意点となる。
In the global GX context
As global disclosure frameworks (ISSB, CSRD) mandate more ESG data, this paper warns that rating divergence may persist or worsen, undermining the usefulness of ratings for investors and the cost-of-capital channel. It adds a critical perspective to the global push for standardized disclosure and highlights the need for measurement reliability.
👥 読者別の含意
🔬研究者:ESG-CFP研究における測定誤差の重要性を再認識し、今後の研究設計に格付選択のバイアスを考慮する必要がある。
🏢実務担当者:ESG開示を拡大しても格付が一致しないリスクを理解し、投資家対応や格付機関との対話に活用できる。
🏛政策担当者:IFRSやCSRDなどの開示義務化が格付収斂に直結しない可能性を認識し、測定基準の信頼性向上策を検討すべき。
📄 Abstract（原文）
One of the best-tested hypotheses in sustainable finance concerns the relation between environmental, social, and governance (ESG) performance and corporate financial performance (CFP). There are well over 2,000 published papers in the last decade, but the results are wildly conflicting, ranging from positive to null to negative, and these findings can arise within the same data sample and methodology. A weak theory cannot explain the discrepancy because three channels (cost of capital, operating efficiency, and valuation premium) lead clearly in the same direction, namely, a positive association. The more likely reason is the poor quality of the ESG measurement. Prior studies document that the pairwise correlations among the six major rating providers covering the same set of companies range between 0.38 and 0.71 with a mean of 0.54, compared to the correlation of 0.99 for Moody’s and S&P credit ratings on the same issuers. Such measurement errors can bias the estimated effect towards zero and make empirical results dependent on which rating provider is used by the authors. Further evidence shows that greater ESG disclosures can lead to wider, rather than narrower, rating divergence, challenging the premise of the recent mandated disclosures. This paper reviews the theoretical rationale for why ESG performance affects financial performance and provides an econometric analysis of how the divergence of ESG ratings occurs and its consequences. It also evaluates the limits of recent developments in regulation, including the International Financial Reporting Standards (IFRS) Sustainability Disclosure Standards (IFRS SDS) and the EU’s Corporate Sustainability Reporting Directive (CSRD). A section is devoted to the Hong Kong market, where both international and mainland Chinese rating agencies rate H‑share issuers under different methodologies, producing institutionally meaningful divergence with portfolio relevance. The paper’s central claim is that the next advance in ESG‑CFP research will come not from more elaborate regression specifications but from resolving the reliability of the measurement instrument itself.
🔗 Provenance — このレコードを発見したソース
- semanticscholar https://wonford.com/wp-content/uploads/2026/09/Manuscript-v2n3-2-2.pdffirst seen 2026-09-12 05:53:16 · last seen 2026-09-22 05:20:17
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。