Exploratory analysis and development of ESG corporate ratings

Abstract

Environmental, social, and governance (ESG) risks have become increasingly important for businesses, investors, and stakeholders. Effective ESG risk management requires the development of robust ESG rating models that accurately capture the different dimensions of ESG risks. However, the existing ESG models have several limitations, including the lack of consensus on materiality criteria and the absence of global reporting standards. This study categorizes the existing ESG models into market-based and academic research-based models, highlighting their main differences. Furthermore, after analysing the latest updates regarding ESG regulations and global reporting standards, the study emphasizes the need for consensus on materiality criteria for each sector and regulatory cooperation among institutions in different countries or sectors to achieve ESG goals. Overall, the global ESG measurement systems are still incomplete, leaving room for further improvement. The study aims to develop an ESG risk rating model, which explores the conditional independence of individual risks using three multivariable methods Logit/Probit and Panel Data Regression. The statistical data collected from global businesses includes cases with both the presence and absence of ESG events and a wide range of predictive variables that refer to financial, administrative, and managerial indicators, both common and idiosyncratic. The results indicate that larger and financially stronger companies, particularly in essential industries, are more prone to experiencing ESG incidents.