Rethinking ESG: creating a shared, open platform for ESG data and rankings
We are walking a thin line between substance and obfuscation. A call for radical transparency in the ESG reporting ecosystem.
We are entering the age of “ESG” investing. What does ESG stand for? To most people it refers to the environmental, social and governance aspects of a business. It involves looking at how good a company is at managing and disclosing non-financial factors that are relevant to the long-term health of its business and its license to operate.
ESG includes a wide array of mechanisms and metrics, some of which are reflected in laws and regulations, some in evolving norms and shared expectations, and others in institutional standards like stock-exchange listing requirements. A few examples include:
- As of March 2019, companies in France are legally required to calculate and publish a gender equality index; companies that do not achieve a threshold by a certain date are fined;
- In June 2017, The Taskforce on Climate-related Financial Disclosures (TCFD) released final recommendations instructing companies how to report climate risks uniformly;
- As of October 2018, companies in the State of Delaware can obtain a certification from the Delaware Secretary of State as to their transparency in sustainability reporting;
- Stock exchanges in India, Norway, Singapore and South Africa have also introduced listing requirements that require disclosures on corporate governance and on environmental and social issues.
ESG definitions are typically the result of multi-year deliberations among investor groups, academics, unions, non-governmental organizations, and public bodies, each with their own objectives. It is a hotly-contested, multidisciplinary field that is defined differently across countries, industries and stakeholders.
More and more investors and investment products integrate ESG factors into their investing process. This is a reflection of a growing sentiment among institutional investors and retail investors that investments should align with their values. In nearly every report about ESG investing (see here for a 2018 Goldman Sachs report, or here for a 2019 report by the World Economic Forum), you will find an impressive statistic that captures the size of this trend. At the start of 2018, USD 11.6 trillion of professionally managed assets — or one in every four dollars invested in the United States — were under ESG investment strategies, a sharp increase from 2010, when the amount was about USD 3.1 trillion.
The challenge of integrating ESG factors
Yet look a little deeper, and it becomes clear that there is no consensus on how to report, measure or incorporate ESG metrics. ESG evaluations are largely based on voluntary, non-standardized, and unaudited disclosures. Some companies present ESG data as part of integrated reports that combine traditional financial disclosures with non-financial data, while others present stand-alone versions of their sustainability reports. The length of sustainability reports ranges anywhere between 20 pages to 200 pages. Deciding what information should be disclosed can be puzzling for companies given the wide range of ESG issues, the range of stakeholders and multiple definitions of “issue materiality”.
The growing supply of ESG information has increased complexity for investors. A lot of ESG disclosures by companies consist of recycled “boilerplate” language, which is of little use. At the same time, companies make an effort to reflect their specific context in their disclosures, making it even hard for investors to extract useful information and compare it across companies. Therefore, while investors are increasingly interested in ESG data linked to financial performance, they find it challenging to access data that is reliable and comparable.
Some efforts are underway to push for the harmonization of non-financial reporting and bring ESG reporting within the regulatory perimeter of financial market regulators. In Europe, there is a process underway on the harmonization of non-financial information, but guidance remains “non-binding”. In the United States, an influential group of academics and investors petitioned the U.S. Securities and Exchange Commission to mandate ESG disclosures, but this initiative appears to be on hold. A range of standards developers also continue to publish and update a range of ESG-related principles and guidelines. Two well-known examples are the standards promoted by Sustainability Accounting Standards Board and the Global Reporting Initiative which both include a detailed series of recommended disclosure metrics, broken down by sector and industry. The harmonization of ESG regulations is good in principle, but faces major obstacles given the complexity and variety of ESG issues and lack of consensus among regulators.
The rise of the ESG data provider; caveat emptor
The continued disparity among ESG regulations and growing interest in ESG integration has created a new type of power player in finance: the ESG data provider. ESG data providers use the confusing state of the ESG reporting ecosystem to their advantage by transforming a chaotic web of information into clean, digestible rows of information. Data providers make the information available in various forms, including indices, ranking tables and standardized ESG metrics. The components of several popular financial investment products are derived from ESG ratings or negative screens which involves leaving out “sin” companies like tobacco companies or arms manufacturers.
ESG data providers perform a valuable role in the markets by providing investors with ready-to-use information. Without their products and services, ESG investing could not have taken off the way it has, since only a handful of institutional investors have the scale to do the required analytical work on their own.
But for ESG investing to become a sustainable phenomenon, the transparency of ESG data needs to be drastically improved. Every ESG model is a representation of reality and requires making many assumptions. Consider a flagship “ESG Risk Report” for Qualcomm, the U.S. telecom equipment manufacturer. The company earns a raw score of 50 on “Data Privacy Policy” and a weighted score of 11.5. How are the weights of factors set? And how are certain component factors selected over others? Figuring out how ESG scores are defined, calculated and weighted seems to require a PhD degree, as well as countless hours to dig through hundreds of pages of documents. Other concerns exist around how quickly scores are updated. In financial markets, time is measured in milliseconds or less, yet many ESG scores are updated only once a year. ESG is an inexact science, but it is time we all acknowledged this and had a frank discussion about how to make it better. John Maynard Keynes referred to the problem with complex models when he wrote in General Theory in 1936:
“Too large a proportion of recent ‘mathematical’ economics are mere concoctions, as imprecise as the initial assumptions they rest on, which allow the author to lose sight of the complexities and interdependencies of the real world in a maze of pretentious and unhelpful symbols.”
An over-reliance on complex mathematical models to price mortgage-backed securities was one of the contributing factors of the 2007–2008 financial crisis. Is it alarmist to suggest something similar might be happening?
Not surprisingly, the lack of transparency of the ESG data value chain (see the exhibit below) has many regulators privately worried a crisis could arise from the growth in financial products linked to opaque ESG ratings. If enough investment products rely on a set of ESG scores to determine their investment allocations (i.e. like some ESG index funds do), volatility in ESG scores — perhaps triggered by an update in scoring methodologies — could result in big fluctuations in asset prices. Another potential red flag is that a small group of organizations with massive data assets are attaining high levels of market control. While there are efficiency benefits to concentrating ESG research tasks among a handful of specialized firms, a fault in the models at one of the leading ESG data providers could distort asset allocations at funds relying on that data provider.
More transparency through a shared, open ESG platform
To stave off an erosion in trust in the ESG ecosystem, we need to prod industry groups, institutional investors, and/or the data platforms themselves, to achieve for ESG data what SoundCloud is attempting to do for music: democratize and open up the entire ecosystem. (SoundCloud is a Berlin-based startup that provides a platform for artists and others to upload, promote, and share audio.)
In a shared, open ESG platform, all critical steps in the aggregation, processing and analysis of ESG numbers would be made fully transparent. Underlying data sources would be made visible for all to see. Marked-up versions of the corporate disclosures that scores are based on could be inspected for accuracy. If anyone has a comment on an individual data point or if someone wants to suggest an improvement or correction to a calculation, that could all be done on the platform. For the first time, it would expose the entire ESG data value chain to a high level of scrutiny, which would provide a more nuanced and robust view of the ESG performance of financial assets.
We should demand nothing short of end-to-end transparency into the entire ESG reporting ecosystem. It might sound audacious, but it is possible. SoundCloud isn’t the only example of a platform that has successfully democratized access to a valuable, digital good. For other examples, one can look at how Kaggle is attempting to democratize data science, how Crunchbase publishes ample data about startups, how Quantopian is providing trading algorithms to the masses, how cities provide the public with open access to their data, or how Reddit is making it easier to share content and have discussions. In fact, Quandl is a Canadian data platform (recently bought out by Nasdaq) that hosts paid and free alternative datasets for finance, while CSRHub is a platform that aggregates ESG scores from more than 600 sources. Either data platform might be a candidate to spearhead an open data initiative for ESG data.
A key benefit of a shared, open data platform for ESG data is that it shifts the burden of policing the ecosystem from a small number of regulators to an army of citizen-regulators. We could leverage the collective brain power of the community, by giving everyone tools to track, scrutinize ESG data, and publicize their opinions on corporate disclosures and the calculated rankings. The reliability of measurements will be revealed over time, giving asset managers the means to hold ESG data providers accountable. Over time, as we collect more data, we could train machine learning models to monitor data flows in real-time, lowering monitoring costs. The result could be a system of checks and balances that can keep pace with the adoption of ESG as an investment metric without imposing costly regulations — exposing the data to scrutiny more efficiently and effectively than any regulation or compliance program ever could.
In some ways, the ESG landscape seems so intractable to make it hard to build a simple, usable, open data platform. While it is true ESG analysis requires expert topical knowledge, rating one company along a single dimension in a particular industry (e.g. “customer data privacy” in the “healthcare sector”) is similar to rating the next company in that industry. The ESG domain could therefore be divided into a matrix of sub-domains, which could be managed by domain experts. A platform moderator would moderate discussions for spam. For the data to be useful to investors, it would have to be clean, reliable and easy to use. There could be rewards for top contributors (donors could award “kudos” that can be converted to fiat currency). Perhaps users of premium datasets would pay a small fee. Perhaps the initiative would be partially financed using funds from institutional investors, who all share an incentive in the success of this platform. Efforts would also have to be made to prevent user interactions on the platform from going to waste, since this can have a stifling effect on community participation. In summary, a lot of details would have to be worked out, but if enough stakeholders support an open data initiative, there is a high chance it could succeed and improve the reliability of ESG data signals.
Conclusion
Corporations are being pushed by society to pay more attention to social objectives: to look after customers’ well-being, reduce their impact on the environment, and nurture the communities in which they operate, among other things. Measuring how well a company performs along those dimensions is extremely difficult, but necessary if we want to make it possible for investors to meaningfully incorporate ESG issues in their process. Today’s opaque ratings system needs to change. Having transparency into the origins of ESG data that drives investment decisions is a critical, but missing element in today’s ESG reporting ecosystem. Making relevant data publicly accessible on an open platform could be the first step towards increasing the reliability, transparency and flow of ESG information.
