Author: 2RSquared – www.2rsq.com
Our post shows that ESG scores can capture different information to Quality scores and hence investing in ESG is not simply investing in Quality. This point has been a topic of recent debate with some studies instead showing high overlap.
Our view is the outcome of any study may depend on the data provider used. Consequently, when making ESG integration decisions, the effect of using different data providers needs to be explored so those decisions can be made as robust as possible to the variability inherent in ESG data.
Setting the scene
We proxy Quality by using a profitability score expressed as gross income over total assets, normalised to a percentile rank score from 0-100. The equity data is powered by FactSet.
The ESG score is the headline score measured from 0-100. The ESG ratings are supplied by OWL Analytics.
We focus on US stocks from 31st Dec 2009 to 30th Sep 2021. Our original universe consists of around 1,250-1,350 medium and large cap stocks depending on the time point. Every quarter we use a sub-set of the universe with a minimum 20-day average trading volume over $1 million and with both an ESG score and a profitability score. The resultant number of stocks ranges from around 750-1,000.
For the US universe we have looked at, there does not appear to be a strong relationship between OWL Analytics’ ESG scores and profitability as a proxy for Quality. Using information coefficient and stylised Top & Bottom portfolios can illustrate this in an intuitive way and provide a more visual complement to regression techniques.
We note that other research has shown strong relationships between ESG and Quality. Our findings are not contradictory to this other analysis. It is simply that different data providers develop their scores in different ways, hence different providers scores will have different relationships to markets.
When considering an investment thesis and integrating ESG through a particular type of data (in our current example ratings), one must recognise the impact can be different depending on the ESG data provider used. The consequence is that when making ESG integration decisions, the effect of the integration needs to be examined over several data providers so that fundamental integration decisions can be made as robust as possible to the variability inherent in ESG data.
As a practical example in the context of our study, if there was consensus amongst data providers (based on numerical analysis) that ESG scores and Quality scores were highly correlated, an ESG focused investor should just use ESG scores, and a quality investor should just use Quality scores. Instead, if there was no such consensus and on aggregate the correlation was low, combining ESG and Quality into a single factor score could make sense and add value to investment returns while at the same time obtaining a better ESG portfolio