Graduate Theses and Dissertations (2019 - present)
Date of Award
8-2026
Document Type
Dissertation
Degree Name
Ph.D.
Department
Business Administration
Committee Chair
Amelia Annette Baldwin, Ph.D.
Abstract
This dissertation examines whether environmental, social, and governance (ESG) disclosure sentiment extracted from SEC Form 10-K Item 1A risk disclosures aligns with Bloomberg ESG scores. The study combines SEC textual disclosures, Bloomberg ESG ratings, firm financial data, executive compensation data, and executive gender characteristics. Natural language processing and artificial intelligence methods, including FinBERT, Loughran-McDonald sentiment measures, BART-MNLI zero-shot classification, LLaMA-based ESG scoring, and readability analysis, are used to evaluate ESG-related disclosure signals. Results reveal that SEC 10-K Item 1A risk-factor sentiment and Bloomberg ESG scores represent related but distinct constructs. ESG sentiment in Item 1A is poorly aligned with third-party ESG scores. Corporate ESG narratives do not consistently align with third-party ESG scoring. Readability does not provide a meaningful explanation for the divergence, suggesting that differences may stem more from selective emphasis and narrative framing than from disclosure complexity. Results also show that firms with female CEOs and a critical mass of female executives demonstrate higher ESG scores. This suggests that gender diversity in executive leadership is positively associated with ESG performance. This dissertation contributes to ESG disclosures, sustainable finance, greenwashing, corporate governance, gender diversity, and AI-based financial analysis by showing both the promises and limitations of using machine learning to evaluate ESG narratives. The results emphasize the need for multiple sources of ESG information, greater transparency in rating methodologies, continued standardization in ESG reporting, and further examination of how leadership characteristics influence ESG outcomes.
Recommended Citation
Davis, John Barnett, "A Machine Learning and Artificial Intelligence Sentiment Analysis of ESG Disclosures" (2026). Graduate Theses and Dissertations (2019 - present). 262.
https://jagworks.southalabama.edu/theses_diss/262
Included in
Business Administration, Management, and Operations Commons, Business Analytics Commons, Business Law, Public Responsibility, and Ethics Commons, Economics Commons, Finance and Financial Management Commons, Management Sciences and Quantitative Methods Commons, Other Business Commons, Public Affairs, Public Policy and Public Administration Commons, Strategic Management Policy Commons