ESG performance and financial distress prediction of energy enterprises

Abstract

In the current drive to cut global carbon emissions, energy companies are facing intensifying policy pressures. This study investigates the impact of ESG (Environmental, Social, Governance) performance on the risk of corporate financial distress in the energy sector. Using a tripartite methodology of sentiment, topic, and word frequency analysis, we measure the characteristics of texts of ESG reports. These ESG-related textual variables, combined with company carbon performance and other variables, are integrated into the CatBoost algorithm to predict financial distress. The empirical findings indicate that text words, topics and sentiments derived from ESG reports prove to be effective in forecasting financial distress in energy companies. Additionally, the CatBoost used in this study surpasses other models such as logistic regression and decision trees in predictive capability. This study demonstrates how incorporating textual analysis of ESG reports enhances the predictive accuracy for financial distress in energy companies, highlighting the important role of comprehensive ESG evaluation in financial risk assessment.