Comparing ESG ratings for better decisions
Abraham Lioui, professor at EDHEC and director of the "Data Science, Economics and Finance" department, presents - in an article published in French on The Conversation - his work on funds integrating environmental, social and governance criteria and the associated ratings.
In 2020, more than $15 trillion of assets in the US were managed by funds incorporating environmental criteria. An increase of 50% compared to 2018! This trend is indicative of the ever-growing popularity of so-called "responsible" investments.
But behind this increase in the attractiveness of funds incorporating environmental, social and governance (ESG) criteria lies a paradox. On the one hand, there is great heterogeneity in rating methods, which results in extreme differences in the ranking of companies. It is therefore impossible to discriminate between companies on the basis of these ratings.
On the other hand, funds do not shy away from proclaiming that they invest in responsible companies and individual and institutional investors seem to believe them. However, investors need to consider the financial performance implications of these differences in ratings. This is the subject of a study we recently published in the Journal of Banking & Finance.
ESG everywhere, consistency nowhere?
In the last ten years or so, social pressure, awareness-raising and regulatory intensification have made ESG criteria a significant barometer of company performance. One sign is the sheer volume of scientific literature on the subject. Nearly 5,000 papers published since 2010 attempt to analyse and deepen our knowledge of this trend, and to draw lessons on the market's ability to integrate these developments.
But opinions differ as to the meaning to be given to market reactions: does being "ESG-friendly" lower the cost of capital? Does the market give a premium to responsible companies like a bonus? The differences in ratings make it difficult to reach an academic consensus on these questions. But not only that! One source of confusion remains linked to the different methodologies used to calculate the "greenium", the name given to the potential bonus for companies that work to preserve resources. This is why we wanted to first deconstruct the dominant methodologies before proposing another, more relevant one, which is not subject to the limitations we have outlined.
Give me your notes, I'll give you your bonus!
To calculate the greenium, two methodologies are used. First, companies are divided into two blocks according to whether they are "green" (good students) or "brown" (can do better). The greenium is measured by the performance gap between the two blocks. But how do you measure the performance of each block? The first method, which is ad hoc, consists of using a weighting by the market capitalisation of companies. The second, economically based method weights companies by their ESG rating. It is therefore intuitive that the two methods will rarely give the same greenium for the same ranking!
The first method uses the ESG rating to separate companies into two blocks but the weighting is based on market capitalisation. Two companies with the same market capitalisation and belonging to the good block will have the same weighting regardless of their rating. In particular, this will be true if they have the same ESG rating or if the ESG rating of one is five times higher than the other.
The second method focuses on weighting stocks within a portfolio by ESG rating. This allows for changes in the rating over time to be taken into account in order to vary the weightings. But it ignores other company characteristics such as market capitalisation.
These two methods, based on a different premise, deliver heterogeneous greeniums for the same ranking, sensitive to additional factors and depending on the universe chosen by the company producing the ranking. We wanted to propose a synthesis.
Levelling to compare
Based on the second methodology - so-called cross-sectional (CS) - my co-author Andrea Tarelli and I wanted to propose an economically based approach, which takes into account the changes in greenium over time implied by rating changes and which neutralises other company characteristics such as size and profitability.
For each rating, we are able to construct a pure ESG greenium, i.e. one that neutralises the unique characteristics of the universe used to produce the rating. We filter the greenium premium for each month of our dataset (which covers about 20 years), and not on average as is usually done. As a result, the good performer premium constructed in this way is not blurred by other characteristics and takes into account the huge changes in investor behaviour, making it readable and allowing for a fair comparison between the different rankings.
Using three different rankings, we show that the differences are not insignificant, as they can be as much as double. Hence the importance for investors to choose their ranking carefully.
The richness of heterogeneity
Our method does not aim to bring ESG rating agencies into agreement, but rather to provide investors with a decision-making tool. We are convinced that the heterogeneity of the rankings is an asset: it would be irrelevant to assign the same value to a limited number of extra-financial criteria; this would considerably reduce the scope of ESG and the ability of companies to make their own transition and give free rein to their creativity.
On the contrary, proposing a spectrum of ESG criteria assessment means embracing the diversity of solutions and efforts implemented by companies to respond to environmental and societal challenges. It is therefore not the homogeneity of the ratings that needs to be worked on, but their readability, so as to be able to compare them and make informed choices, and this research makes a modest contribution in this direction.