計量経済学と機械学習の統合：FTSE先進国ESG指数におけるESG-財務パフォーマンスの連関
Integrating Econometrics and Machine Learning: ESG-Financial Performance Nexus in FTSE Developed ESG Index (原題)
Randy Adhiputra, A. A. S. Gunawan, Ferry Hadiyanto
🤖 gxceed AI 要約
日本語
FTSE先進国ESG指数採用企業を対象に2016-2025年のESGと財務パフォーマンスの関係を、二方向固定効果パネル回帰と機械学習アンサンブルで分析。ESGはTobin's Qと有意な正の関連を示し、効果は高マテリアリティ産業に集中。機械学習は非線形構造を明らかにし、ESGは操業・資本構成変数より重要度が低いことを示した。
English
Using two-way fixed-effects panel regression and tuned tree-based ML ensembles on FTSE Developed ESG Index firms (2016-2025), this study finds ESG positively associated with Tobin's Q, marginally with ROA, and not with ROE, with effects concentrated in high-materiality sectors. ML models reveal nonlinear patterns and show ESG ranks below operational and capital-structure variables in feature importance.
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
ESGと財務パフォーマンスの非線形・マテリアリティ条件付き関係を示す本成果は、SSBJ開示や統合報告におけるマテリアリティ評価、投資家向けESG説明責任の根拠として日本企業・投資家に示唆を与える。
In the global GX context
This work reinforces the global evidence base for ISSB/TCFD materiality-based disclosure by demonstrating that ESG-financial links are nonlinear and sector-materiality-conditional, informing how investors and standard-setters interpret ESG ratings and valuation relevance.
👥 読者別の含意
🔬研究者:ESG-財務連関の分析に計量経済学と機械学習を統合する方法論的枠組みを提供する。
🏢実務担当者:自社のESG取り組みが市場評価（Tobin's Q）に結びつく可能性を、マテリアリティ産業別に示唆する。
🏛政策担当者:ESG開示のマテリアリティ基準設計において、非線形・産業別効果を考慮する必要性を示す。
📄 Abstract（原文）
Environmental, Social, and Governance (ESG) performance is increasingly used in investment analysis, yet its financial relevance remains debated. This study investigates the relationship between Corporate Financial Performance (CFP) and ESG performance of firms listed in the FTSE Developed ESG Index for the period 2016-2025. This study integrates the econometric inference using two-way fixed-effects (FE) panel regression and machine-learning prediction using tuned treebased model ensembles. Model performances are compared using the Root Mean Squared Error (RMSE) metric. The results showed FE estimates a positive and statistically significant association between ESG and Tobin's Q (coefficient 0.0040, p = 0.011), a marginal association with Return on Assets (ROA), and no significant association with Return on Equity (ROE), also the ESG relevance is concentrated in high-materiality sectors. All tuned models outperform OLS and Random Forest showed the lowest RMSE in almost all the CFP with the score 0.0461 for ROA, 0.4089 for ROE, and 1.4150 for Tobin's Q. ESG ranks behind operational and capital-structure variables in feature importance. The Partial Dependence Plot further reveals nonlinear structures that differ across outcomes. These findings demonstrate that integrating econometric and machine learning methods reveals materiality-conditional and non-linear patterns in the ESG and CFP relationship that linear single-method analyses obscure, with effects concentrated in high-materiality sectors and market-based valuation.
🔗 Provenance — このレコードを発見したソース
- semanticscholar https://doi.org/10.1109/icimtech202671078.2026.11699005first seen 2026-10-02 05:46:57
🔔 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。