ESG data standards are a mess. Perhaps this is not such a huge problem
Or why diverse data can be a net benefit
While conducting our research for Wangari and speaking to various industry experts over several months, one subject kept coming up. ESG data standards are a mess, experts told us. One finance professor even went as far as claiming, only partly sarcastically, that the entire ESG data market was a scam.
Luckily, Wangari never intended to become an ESG data provider! Jokes aside, the ESG data market looks messy even for industry veterans. Many companies, such as MSCI, Sustainalytics, and others, have thus resorted to condensing their results in ESG scores.
The buyers of ESG data — this includes asset managers, banks, and insurance companies, and us at Wangari — are not liking these scores, however. Each data provider makes its own set of scores, based on what they think is important to them and their clients. The result is that every asset can be assigned with dozens of different scores. The values of these scores can vary wildly for a single asset.
Many ESG data buyers therefore are resorting to using the raw data. At Wangari, with its strong focus on detailed financial modeling, doing this was a no-brainer. At other financial institutions, however, this is a less obvious choice because transforming raw data into financial insights is a lot of work.
A director at a leading ESG data provider put it to us like this: “Clients don’t want a frozen pizza. They want the raw ingredients from us, so that they can make the pizza themselves.”
Unfortunately, using the raw data is a messy process too. For tomatoes and basil, there is strict and well-established regulation regarding their size, color, consistency, and freshness. For carbon footprints and labor rights, the regulation is work-in-progress.
Diverging standards for raw data
Businesses that issue data about their ESG performance need reporting standards. They do not necessarily know what to measure and how conduct these measurements, so standard frameworks can guide them. People using this data also want to compare it with that of other businesses. Hence, frameworks help ensure that every datapoint is measured, and that these measurements all follow the same methodology.
The problem is that developing such a framework requires deep expertise. In addition to this, potential limitations of a framework often only get noticed after it has already been rolled out.
Here is an example from outside the ESG world: During my PhD in particle physics, I created a framework to establish whether a dark matter theory was coherent with experimental observations or not. Months after publishing our work, researchers from other universities asked us why a certain class of dark matter particles was not included in the framework. We had thought that this class was not sought after by other researchers and had hence left it out for the sake of simplicity. This assumption turned out to be false, and we had to add that feature later on.
In the world of ESG, many notable expert groups have developed ESG reporting standards. There are so many different ones that we at Wangari refer to this multitude as the Ocean of Acronyms. Here is a quick overview over the most important ones:
In the European Union, medium and large companies are subject to the Corporate Sustainability Reporting Directive (CSRD). Companies subject to the CSRD have to report according to European Sustainability Reporting Standards (ESRS). Prior to 2024, large companies had to follow the the Non-Financial Reporting Directive (NFRD).
Another popular set of standards has been developed by the International Sustainability Standards Board (ISSB), which was set up in 2021 by the International Financial Reporting Standards (IFRS) Foundation. This includes two sets of requirements, one for sustainability-related financial information and the other for disclosing specific information about climate-related risks and opportunities.
Released in 2018, the Sustainability Accounting Standards Board (SASB) Standards contain specifications on disclosing financially material sustainability information across 77 industries. They currently remain available for use, although they should eventually be replaced by the standards developed by the ISSB.
The standards set by the Task Force on Climate-related Financial Disclosures (TCFD) were first released in 2017 and have since been used by over 4,000 companies. They have, however, disbanded and have handed over their work to the IFRS Foundation.
The Taskforce on Nature-related Financial Disclosures (TNFD) has released new standards in 2023. As of early 2024, hundreds of companies are planning to adopt these standards.
This list is just a taster; there are many more frameworks. Most of them are voluntary, notable exceptions being the CSRD and the NSRD.
Voluntary disclosure is better than nothing, but it skews the results. It is in the interest of companies that already do well in ESG to disclose their progress, while companies that perform worse might opt against disclosing theirs. In addition, voluntary disclosure favors larger businesses that have budget to assign such large tasks to their employees. Small businesses often lack the capabilities to conduct so many measurements and hence miss out on a potential opportunity to shine in the public eye.
When a number of companies discloses their ESG performance while many others do not, and when those that do disclose it choose varying frameworks or even develop their own — that’s what one calls a mess.
Signs of convergence
There are signs that the days of the wild west of ESG data might be counted. The ISSB has snapped up many different frameworks in order to integrate them into their own. This has notably been the case with TCFD.
The main contender to the ISSB framework is CSRD, the latter of which is known to be more granular and detailed. Companies in many non-EU countries seem to favor the ISSB framework, however, because it is simpler in its structure and easier to adopt.
In a similar fashion, the market for ESG data providers has markedly consolidated in the last few years. This might be good news for financial institutions seeking to buy data. Figuratively speaking, they do not have to bring quite such a large machete to get through this jungle.
The upsides of data diversity
We have already discussed some of the downsides of this mess. Data points from one company to another can be difficult to obtain and to compare. Without a minimum of skill in data engineering, data science, and ESG frameworks, this situation can lead to skewed results for investors.
On the other hand, such a huge diversity can lead to many lucky findings (we say this word frequently at Wangari) that one would not have found in more homogeneous datasets. Imagine a company that initially disclosed their ESG data using their homegrown framework. A few years later, they adopted CSRD.
This company belongs to a sector where some data points, such as land use, do not vary much from one year to the next. If they disclosed their land use in a very detailed way with their homegrown framework but later got less granular by using CSRD, then one can reasonably assume that the data points from the homegrown era would be the roughly the same as in the first years with CSRD.
This might be useful information because various data points are interconnected. If, for example, their land use has not changed but their energy use has gone up, perhaps they are producing more units of product per square meter of factory floor. Or perhaps it was a rough winter and they needed to heat their premises more.
Whatever the case for this company, having various different frameworks might lead to insights one would not have had otherwise. Such insights, of course, needs to be addressed with critical judgement and corroborated by other findings. They do, however, offer a unique window into the operations of a company that might ultimately be the deciding factor for or against an investment.
Conducting such a diligent and sophisticated process for every company is out of scope for many financial institutions. At Wangari, we aim to grant them access to such insights anyway.
We do this by ways of automating this process using advanced data science and various fancy AI-powered algorithms. In the coming weeks and months, some pieces in this blog will cover our trials, errors, and most interesting findings from this endeavor.
Unity in diversity?
Lacking the resources to process so much messy raw data, some financial institutions just buy all available ESG scores from all data providers. They then use a fairly sophisticated algorithm to calculate a weighted average. The result that they obtain is their final score which informs their investment decisions.
At Wangari, we are skeptical of this approach: ESG scores are even less standardized than raw ESG data, and averaging the data eclipses the opportunity for detailed scrutiny that would lead to what we call lucky findings. In addition to this, there is some danger from the garbage-in-garbage-out phenomenon: If a model is fed with faulty data, the chances are high that the results are sub-par.
There is some merit to the idea that averaging many diverging data points, faulty as they might be, would indicate where the market consensus is. In our view, however, this does not help investors beat the market. We believe that it is better to uncover the hidden secrets that one will only find by turning over every stone and digging deep into each detail.
Our work would certainly be easier if ESG data were less messy. However, it would also prevent us from obtaining lucky findings, i.e., insights that everyone overlooked but that are important parts of the future trajectory of the company in question.
Despite the fact that some clearing up has begun, the landscape of ESG data remains a huge mess. At Wangari, we take pleasure in uncovering the raw diamonds that are buried in this mess.