Credit Risk Ratings vs ESG Ratings

If you have worked with ESG ratings, you most likely heard something like: “How can we trust ESG ratings if the correlations between ESG raters are so low compared to credit risk ratings?”. Let me explain in the following post why comparing the two makes little sense.

As we explain in our “Aggregate Confusion: The Divergence of ESG Ratings” paper, there are mainly two reasons for the disagreement between ESG raters. First, a little bit less than half of the disagreement can be explained by different definitions of sustainability. This source of divergence will persist as different actors have different values or ideas about what is material. And who are we to tell anyone what is important to them? Put differently, ESG ratings are multi-dimensional constructs that change their dimensions between users and over time. Credit risk ratings are not subject to these value judgments. They only reflect credit risk.

The other, slightly more important, part of divergence is measurement. The problem with most ESG issues is that we often do not observe the actual outcome over time. For example, if you measure discrimination, you might resort to a proxy variable such as the gender wage gap. A year after the first assessment, you will not be smarter about how women are treated unless you use different data points, i.e., the measurement error is the same. Credit ratings, however, only measure a single event that is default or loss associated with default. Credit risk agencies can improve their methodologies through backtesting because this is a binary event observable with very little noise.

Hence the higher correlations for credit risk ratings.