Revolutionizing ESG risk assessment through machine learning: Insights from U.S. corporations
Abstract
Environmental, Social, and Governance (ESG) factors are increasingly essential in evaluating corporate performance, driving demand for accurate ESG risk assessments. However, smaller companies often face challenges in obtaining validated ESG scores due to resource constraints. This study explores the use of Machine Learning (ML) to predict ESG risk scores for U.S. companies, leveraging data from Wharton Research Data Services (WRDS), Yahoo Finance, and Sustainalytics. The XGBoost model demonstrated the best performance, significantly improving the accuracy of ESG risk predictions. These findings suggest that ML can enhance ESG risk assessments, offering valuable insights for investors, regulatory bodies, and corporate management.
This paper has been withdrawn.