Chasing ESG performance: How methodologies shape outcomes

ESG metrics play a crucial role in sustainable finance but face growing criticism for their inability to accurately capture actual sustainability improvements. This study investigates how methodological choices can introduce distortions in ESG scores, with a primary focus on Refinitiv ESG data, while offering insights applicable to other providers as well. We show that data aggregation and score normalization through percentile ranking significantly impact the ability to reflect genuine sustainability progress. Specifically, the inclusion of new, smaller in size entrants with more missing data can artificially inflate the scores of top-ranked companies, obscuring actual sustainability improvements in the underlying metrics by relying on peer comparisons. Moreover, our analysis reveals that once companies achieve an A-rating category, they are unlikely to be downgraded, further highlighting the impact of these methodological decisions on the dynamics of ESG scoring. Overall, our analysis of three key sectors from 2012 to 2021 reveals that less than 45% of total absolute score variation is attributable to company disclosures, emphasizing the influence of score construction methodologies. We show that replacing percentile ranking with a simpler Performance Ratio methodology could mitigate these issues, offering more representative scores. • This study explores how methodological choices, such as percentile ranking, in ESG scoring might distort representation of sustainability performance. • We benchmark Refinitiv’s scores, which are based on percentile ranking, with scores built with an alternative method: the Performance Ratio. • With percentile ranking, top scores are inflated by the incorporation of smaller firms with more missing data. At the same time, the progress of the existing peers deflates such scores.The Performance Ratio approach exhibits robustness towards these two distortions. • Across three sectors from 2012–2021, less than 45% of absolute score variation is driven by company disclosures, highlighting methodology-driven distortions of percentile ranking. The Performance Ratio approach captures instead more than 83% of absolute score variation.