Predictive roles of environment, social, and governance scores on firms’ diversity: a machine learning approach

Purpose - Environmental, social and governance (ESG) scores are compelling for firm strategy and performance. Thus, this study aims to explore ESG scores’ predictive roles on global firms’ diversity scores. Design/methodology/approach - A total of 1,114 global firm-year data from the Thomson Reuters Eikon database was analyzed using machine learning algorithms like rpart, support vector machine, partykit and evtree. Findings - The results reveal a positive association between diversity, resulting in greater comprehensiveness and relevance. Broadly speaking, the two factors with the most significant values for calculating the overall diversity scores of businesses are ESG scores and social scores. ESG scores and environmental scores are the most effective predictors for the diversity pillar and people development scores. In contrast, community and social scores are the most important predictor factors for the inclusion scores. Originality/value - The research is particularly pertinent to managers and investors considering ESG issues while making decisions. The results indicate that leaders and practitioners should prioritize ESG elements and diversity problems to enhance performance.