Avoiding Pitfalls In ESG And DEI Measurement Software
I sat down with George Lee, Founder and CEO of Hydrus.ai in San Francisco to talk about the Environmental, Social and Corporate Governance (ESG) reporting, how it is exploited, and how companies can leverage data, analytics, and fraud to improve their visibility to investors. What started out as technocratic accountability has turned into political gaslighting. Considered a leading environmental company for its electric vehicles, Tesla was removed the S&P 500 ESG Exchange Traded Fund (ETF) because of “lack of a carbon strategy.” Lee explains his experience working with corporations on their ESG, the risk of inaccurately reporting non-financial data, and why a software-driven, transparent, and auditable approach is key to adding balance to a field that is increasingly polarized.
Q: ESG seems like a great idea on the surface, but what went wrong? It used to be that we welcomed a company’s reporting about what it was doing for the environment. Now many don’t trust it.
A: “Greenwashing” in ESG, deceptive marketing that a product or service is less environmentally friendly than asserted, has always been a problem and goes back decades. The earlier terminology was CSR or Corporate Social Responsibility. Over time, “Environmental, Social, and Governance” emerged as the three key sustainability issues facing firms.
One proposed value proposition of ESG/CSR is to improve a firm’s sales, marketing, and brand. However ESG can backfire on a firm and lower these assets. Various stakeholders (e.g. lenders, ratings agencies) demand more ESG information, creating headaches for reporting and incomprehensible standards applied broadly to companies with little visibility – especially around Environmental and Social factors. In turn, activists and the media often exploit companies’ reported non-financial metrics that to make negative accusations and statements.
Indeed the process of ESG reporting has always been messy; organizations hire ESG exports to cobble together supposed ESG data. These Excel sheets and reports generally that lack auditability. Software has historically underserved this sector and is made up of many tools not intended for both quantitative and qualitative data management.
Fortunately Hydrus.ai resolves this challenge and ties data to the relevant audit, accounting, and other parameters. The software controls and audits the data collection and reporting process.
Q: It’s almost like the data isn’t “real” anymore. How do we know which data be trusted?
Wall Street firms have power to demand that management complies with ESG standards to continue receiving investment. Blackrock and other large financial institutions learned they can make ESG investing a profitable business by charging fees for “green” investment products. These products are built with vague and frequently negative sector based screening such that investments in categories like firearms, oil/gas, tobacco, and more are automated excluded and deemed not ESG.
If the company does not meet these subjective ESG standards, they are divested from and screened away from investments not only in exchange traded funds (ETF), but also pension funds and sovereign institutional investments (particularly in Europe with Sustainable Finance Disclosures Regulation or SFDR).
This transition of financial investment has single handedly increased US national security risk over infrastructure because industrials, energy, manufacturing, transportation, and many traditional sectors are underinvested. As such they are arbitrarily deemed non-ESG when other measures could determine that they legitimately meet these factors.
Only audited, transparent data with a disclosed methodology can be trusted, for example how were the Greenhouse Gas calculations performed, which methodology was used (AR4 vs. AR5 etc.).
Q: What about Diversity, Equity, and Inclusion (DEI)? Are their numbers accruable and trustworthy? Can you give me some examples of the pros and cons?
DEI data has been available for some time but it is not perfect. There can be some significant and potentially harmful gaps between a firm’s intention and the conclusions drawn from DEI data. It is critical to know exactly how the DEI data was sourced and reported. Some human resource professionals have made some major mistakes in judgement.
For example, I am ethnically Asian, but Accenture, a Fortune 500 global consulting firm with over 500,000 employees, categorized me as “Caucasian/White” in their Workday HR system as. I never specified my ethnicity when I joined the company, so how did they “assume” my ethnicity?
This begs the question, what does self-identification mean? Which categories should exist for gender, ethnicity, and other areas of DEI? There is no commonly agreed upon standard for DEI reporting.
Companies like BlackRock say they will “boost the number of black employees by 30% by 2024”, but this means little without context. If there are 100 employees and 10 of them are black, a 30 percent increase amounts to three people. Many DEI reports, statistics, and communication are inherently deceptive.
Similarly BlackRock said it would “donate $5 million to organizations focused on improving racial equity.” However this assertion of “equity” is not tied to a tangible, recognized measure. It amounts to self-serving virtue signaling. In any event $5 million is a drop in the bucket for BlackRock.
Q. Are there biases in the measurement tools themselves? How can that be addressed?
Many software entrepreneurs may be ideologically aligned with ESG and DEI goals and hence design software to overemphasize or underemphasize relevant data. There is some market attempt to correct for this as investment tell companies to focus on business risks, costs, and opportunities. This sounds a lot like “Business 101.” The bottom line is that ESG measurement is nuanced.
Most ESG and DEI software platforms were built for purposes unrelated to these goal, e.g , social impact data tracking, health and safety, greenhouse gas measurement and so on. Copy-pasting these measures for corporate communication can be risky.
It is important that the ESG software accurately reflects the purported measure rather than the mere translation of the thinking of the software developer or entrepreneur. To get around this, firms should audit first, then define the relevant data, and beware of ideological bias which could mistake the firm’s information.
Q. Companies’ ESG data is used in regulatory proceedings. How can policymakers be more critical of the data?
ESG ratings, standards, and reporting are highly subjective. Ratings agencies like MSCI, Refinitiv, S&P, Morningstar, and others hire legions of offshore workers in emerging countries to score organizations on ESG and DEI measures. There are no common standards for these measures. And yet institutional investors and regulators increasingly require corporations to report this information. Policymakers should consider how to reduce the abuse of ESG information.
The same PhD academics who publish academic studies also work for the ratings agencies, creating conflict of interest and “principal–agent” problems like the ‘08 financial crisis. MIT’s Roberto Rigobon’s “The Aggregate Confusion Project” observed bond ratings are uniformly correlated at 0.9, yet ESG scores are only correlated at 0.6.
ESG standards and reporting can’t be globally standardized due to local differences in culture, values, and industry making it impossible to assess the same ESG principles from one company to another. Indeed the blind demand for data could diminish innovation and growth may diminish and indeed, create incentive for overstatement.
Hydrus solves the problem with an integrated, highly differentiated approach. We cut through the noise by automating the collection of raw data from a variety of sources from finance to energy to HR. As all the data is aggregated into a single system and cryptographically linked to the relevant actor or transaction, the ability to audit and analyze the data is greatly improved. Customers have greater confidence that the resulting information is accurate, meaningful, and actionable.