Access Official Technical PDF

Share Summary:LinkedInEmail

ESG rating divergence matters to industrial companies when investors, customers and other stakeholders use sustainability scores to assess corporate performance. Yet different rating agencies can assign notably different scores to the same firm. A study by Florian Berg, Julian F. Koelbel and Roberto Rigobon examined this disagreement and found that it reflects not only different views about what matters, but also differences in how agencies measure the same issues.

For manufacturers, automation companies and other industrial organizations, the practical lesson is to look beyond a headline score. Understanding how a rating is constructed can help explain why providers disagree and make it easier to evaluate sustainability assessments in context.

How ESG ratings are constructed

A rating agency turns company information into a summary score through several choices. The study groups these choices into three elements:

  • Scope: Which environmental, social and governance attributes are included?
  • Measurement: Which indicators are used to assess each attribute, and how are the indicators interpreted?
  • Weights: How much influence does each attribute or indicator have on the overall score?

Consider water management. A provider might assess water use, recycling, water-related risks or management programs. Another might include a different subset, use different measures for a similar issue, or assign water a different weight in its overall calculation. The resulting scores can diverge even when both agencies describe their work as measuring corporate sustainability.

These distinctions are relevant to industrial operations, where sustainability topics can include resource use, emissions, worker health and safety, and governance. A score alone may not show which topics were assessed or how the underlying evidence was interpreted.

What explains ESG rating divergence?

The researchers analyzed ESG ratings and indicator-level data for 924 companies covered by multiple providers. Their baseline analysis used 2014 data and included ratings from KLD, Sustainalytics, Vigeo Eiris, RobecoSAM, Asset4 and MSCI. To compare providers, the researchers developed a taxonomy that grouped related indicators into common categories. This was necessary because it was rarely possible to find two agencies measuring precisely the same attribute for the same company.

The study’s headline decomposition attributed 53% of rating divergence to measurement differences, 44% to scope differences and 3% to differences in weights. In other words, the largest source of disagreement was often how agencies assessed an issue—not simply which issues they chose to include or how much importance they assigned them.

The researchers also identified a rater effect: systematic differences associated with the rating provider itself. In their fixed-effects analysis, adding firm-and-rater effects increased the explained variation by 14 to 16 percentage points, depending on the specification. This indicates that provider-specific practices can contribute to differences beyond the company and sustainability category being assessed.

Why an overall correlation can be misleading

Agreement among rating agencies is not uniform across companies. When the researchers examined firms with the strongest agreement, the correlation between ratings was 90%. For the low-agreement group, it was 32%. An aggregate correlation can therefore conceal substantial company-level variation: agencies may broadly align on many firms while disagreeing sharply on others.

Scores and rankings also answer different questions. A score represents an agency’s assessment under its own framework; a ranking places a company relative to others in a particular set. The study tested its analysis using rankings as well as normalized scores, reflecting the fact that users may care about relative position rather than the numerical score itself. Neither format removes the need to understand the underlying methodology.

Practical implications for industrial companies and users

For industrial companies

When two providers assign different scores, comparing their underlying categories and indicators can help identify the source of the gap. The disagreement may reflect different coverage, different evidence or different measurement rules. A structured review of those elements can make the comparison more informative than focusing on the overall score alone.

For investors and decision-makers

The appropriate approach depends on the decision being made. A user may select one provider whose methodology fits a particular purpose, while recognizing that its score reflects that provider’s framework. Alternatively, users can compare multiple ratings or work with indicator-level data to form a broader assessment. The study’s authors caution that using a single rating should be carefully justified, especially in empirical analysis.

For analysts and researchers

Specific categories can be more useful than a broad composite score when the question concerns a particular sustainability issue. The researchers also recommend using average ratings when the goal is to represent a consensus view, rather than assuming one agency captures the market’s overall assessment.

Conclusion

ESG rating divergence is not simply a matter of competing opinions about corporate responsibility. It can arise from differences in what providers assess, how they measure it, how they combine indicators and provider-specific practices. For industrial organizations and the people evaluating them, the practical response is to examine the methodology behind a rating and connect the score to the decision at hand. A more careful reading of the evidence can distinguish genuine differences in company performance from differences in the way that performance is assessed.

Access Official Technical PDF

About The Author