ESG Indices help to mitigate and assess the risk of companies against ESG factors and helps socially responsible investors to navigate around ESG risks. But how accurate is ESG data? This rather ambitious question was the subject of some – perhaps necessarily – open-ended debate at a roundtable held at the TBLI Nordic conference two weeks ago. The roundtable featured Sylvain Chateau (COO & Co-Founder, Beyond Ratings, Emily Steinbarth (Quantitative Research Analyst, Russell Investments), Lauren Smart (Executive Director, S&P Trucost ltd), Michael Viehs (Associate Director, Hermes Investment Management), moderated by Glenn Frommer, Managing Partner ESG Matters IVS.
Data and Ratings
A prominent criticism of ESG data accuracy is the confusing of data with ratings. A full ESG assessment is made up of many different ESG data points, where there is an inherent logic to ESG ratings that doesn’t necessarily compare across data points. For example, as Smart suggested, sales-side opinions can make a difference to which ESG data points are considered. In addition, there can be hundreds of data points, and this is likely to increase as ESG data evolves over time. As nobody knows what the future looks like, reliance increases on analysis and assumptions about certain trajectories, but in terms of forward-looking ESG data, there can only be views.
One underlying data point, though, is carbon data, where there are still biases in terms of expected disclosure towards large or medium cap companies – which is only 3% of the global market. 70% of these report carbon data but about a third still require work to standardise their reporting (the Greenhouse Gas Protocol helps with this, Smart said, but not all companies are quite there yet). And what happens to the other 97% of the global market? It’s also not just a question of looking for straight-out reporting inaccuracies, but also at data gaps where there’s no disclosure at all and different ways of reporting requiring standardisation.
But how much does “perfect” data matter, and should we be waiting for perfect data, or does being, say, 90% there suffice for action? Waiting for perfect data can mean missed opportunity. There also needs to be more consensus on all the ways data is reported, in terms of the standards and metrics employed across the industry. And, consequently, what the standard of “perfect” disclosure is – is it perfect because the company says so? Or are there other data points that would provide greater accuracy?
Another big issue for ESG ratings, Chateau suggested, is that for one single company there can be very different ratings. For example, Elon Musk’s Tesla was very well rated in one way but was awarded the poorest rating ever by another ratings agency just because they didn’t have the same perception of what needs to be assessed. From this perspective the underlying necessity is to make people converge on what a good perception of an ESG rating is, so the foremost issue becomes transparency.
Managing the Old with the New
Added to this is the dialectic of accuracy and reactivity. If ESG indicators are usually submitted annually or two or three times a year, it can be difficult to do anything reactively if the data is already a year or more old. So how can one help investors using older data? On the other hand, it’s better to have something rather than nothing, even if estimates could be more precise, meaning that it’s a continual process of the connection between ESG ratings and ESG data and how these are used effectively in the investment processes. There is also a structural problem in terms of the pressures on too few analysts covering too many companies, according to Chateau; and companies cannot be expected to deliver perfect transparency with the amount of data available. It could make sense to make companies more auto-corrective, e.g. make the company pay for its ESG rating, which benefits everyone.
On the standards of ESG rating companies themselves, there are talks about regulating rating activity, Chateau explained: if a company cannot defend itself against a standard framework, it becomes very hard to make the market converge towards an objective perception of ESG demands. In terms of ESG frameworks and the connection between the financial world and the real world, ESG frameworks were originally built with a lot of conviction and an ethical approach, whereas they now have the opportunity to lean on more materiality, where it can be demonstrated which indicators have an impact on creditworthiness and statistics show correlations or trends, in turn suggesting that not all ESG data is always relevant and that much also depends on the level of economic development in a given country.
Abundance Means Finessing
For Steinbarth, the onus is on investors to push for higher disclosure, where the two dimensions of data quality are quality and quantity. In terms of quantity, whereas disclosure has increased for carbon emissions over the past 6 years to 70% from 50%, other areas of ESG aren’t going to yield similar levels for quite some time; conversely, it could also be argued there is already too much ESG data, where it doesn’t necessarily mean it’s all high quality. In Steinbarth’s view, this requires coalescing around data that is already available from a qualitative perspective and have discussions at the industry level as opposed to on a company-by-company level.
Viehs provided the insight that ESG ratings quantify something that is not necessarily quantifiable by providing numbers to clients that are relatively easy to understand in terms of their own needs, which means, again, that ESG data is not always accurate. ESG data is important to integrate as it provides a certain signal of where companies rank in terms of ESG performance, but it’s important to not make more of it than it is. The most material ESG ratings matter the most, Viehs said, concurring with Chateau. A 1-100 or A-D scale is not necessarily meaningful.
As ESG is not a one size fits all approach, it doesn’t therefore make sense to impose definitive ESG beliefs onto fund managers but rather to tailor them according to a fund’s investment philosophy and strategy. The same goes for asset owners who have different beliefs in terms of emphasising environmental or social and governance issues, and also different geographical jurisdictions, where regions, countries and laws diverge widely. A single strategy, therefore, cannot satisfy an entire asset base.
Image: (c) TBLI