A proposal for refining the ESG methodology used by rating agencies

Nowadays, global uncertainty has consequences for financial investors. Some of them consider sustainable investment as a refuge from instability. Environmental, social, and governance (ESG) scores are fundamental tools for assessing a company's sustainability and other ethical and social aspects. Due to this challenge, several rating agencies provide ESG grades and scores for companies, investors, and stakeholders. However, the uncertainty associated with the linguistic descriptions of the rating scales, along with partial information regarding their methodology, can complicate the decision-making process for investors. For this reason, this paper proposes a new approach to constructing ESG scores by integrating two methodologies: ordinal proximity measures and the extended best–worst method. The former considers the different perceptions that investors may have concerning the ratings on the scale. At the same time, the extended best–worst method seeks to define criteria weight coefficients using fuzzy set theory to manage ambiguity and imprecision. These procedures are applied to obtain linguistic assessments for 115 companies worldwide in the energy utilities sector, considering a specific investor's preferences. The dataset was obtained from Refinitiv Datastream and encompasses the period from 2018 to 2021. The findings demonstrate that the proposed procedures complement and enrich the rating methodology used by Refinitiv and can enhance the investment decision-making processes under uncertain situations.