ESG Data Diffusion, and What Advisors Can Do About It
As wealth clients start to invest with their values, financial advisors find themselves in the confusing position of having to sort and select ESG data - an industry that was not built for them and is increasingly diverse. In fact, research reveals that there are concerning levels of variance and contradiction between providers and their company scores. Who should you trust and how to decide?
In this article, I dive into the reasons for this variance, and why I believe the fragmentation will likely expand. More importantly, I’ll explain why I think the industry has addressed the problem backwardly and how to bypass the noise to get ESG data right in your practice.
Why ESG Data Varies
The conventional wisdom is that ESG data diverges because there is no common standard. This is true. But the problem runs much deeper. ESG data scoring is fundamentally subjective, for reasons I will outline. This leaves room for each data provider to promote their own ranking system, pursue differentiation, and then claim their approach is better, for which they have a financial incentive to do so. It is thus no surprise that the rating for a single company often ends up varying greatly across different providers.
Take General Electric as an example. Let’s look at data from two of the most widely used ESG data providers: Sustainalytics and MSCI. According to Sustainalytics, General Electric has an ESG risk rating 46 out of 100, or “severe.” Whereas MSCI ranks GE as “A,” or “average.”
Confused yet? This pattern repeats across much of the listed universe.
After comparing ESG ratings among five agencies — KLD, Sustainalytics, Vigeo-Eiris, Asset4, and RobecoSAM— a team of researchers at MIT Sloan found that the correlation among those data providers’ ESG ratings was on average 0.61. (1) (2) Worse yet, according to research done by State Street Global Advisors, there was a correlation of only 0.53 among the ESG ratings of MSCI and Sustainalytics. In other words, almost half of the ESG ratings of companies are inconsistent across these data providers.
See Figure 1 and Figure 2 below for a further breakdown.
What’s driving this divergence is how each data provider takes into account scope, weightings, and interpolation of missing data, (1) (2) as well as subjective decisions made by asset managers during portfolio construction.
Scope divergence is a result of data source variance among different data collection providers. This occurs often in ESG reporting because there is no one way to measure the impact a company has on a certain issue. For example, one data provider could include gender pay gaps, the number of women in leadership positions, paid family leave, job security during and after pregnancy, and contraceptive coverage in its ratings scope for gender inequality in the workplace, while another might not consider paid family leave. (1) (2) When different data points determine a company’s ranking, it is no surprise that the data will start to vary across data providers.
Weightings divergence is when data providers evaluate or measure an attribute using different degrees of importance. (1) (2) Returning to the example of gender equality in the workplace, one provider might value the number of women in leadership positions higher than job security, whereas another provider might value those equally. Not to mention that during the act of measuring and weighing these different attributes, companies might use different indicators to evaluate a company’s practices. (1) (2) For example, one company might measure women in leadership positions by the number of women currently in seats, while another will look at women in leadership over the entire company’s history.
Another big reason for the divergence in ESG reporting is through the interpolation of missing data. ESG data collection providers source much of their data from public disclosures. (3) However, not all companies disclose data points needed to accurately measure a company’s ESG ranking, or they simply don’t collect all the data internally needed to then accurately report. This is because there are no formal reporting standards from regulatory agencies on what ESG data a company should publicly disclose. (4) This leaves companies solely responsible to determine for themselves what ESG factors are important and what they will disclose to investors. (4)
As a result, ESG data providers need to use statistical models to create estimates for data that the companies do not report themselves. (3) These statistical models are based on research, trends, and averages from companies within a similar industry. (3) While aiming to close the gap on undisclosed data, these models can never provide 100% accuracy and will lead to variance across different ESG data providers.
There’s another big reason for divergence at a portfolio construction-level, which few admit or want to talk about: there are subjective, value-based trade-offs that need to happen when combining ESG themes during portfolio construction. For example, when replicating the S&P 500 index with ESG overlays, how many tons of carbon do you trade-off for additional women on the board? There are no right answers here, but someone has to make that decision. The result of that decision has significant implications for the ethical profile of the overall portfolio.
Trend Towards More Diffusion, Not Convergence
Many people think that the levels of divergence in ESG data are due to the industry being in its infancy. Perhaps they think this will soon be “solved” by the US government, which just needs to step in to regulate and create a consistent framework. I deem this to be unlikely any time soon. Under a Biden administration, the government might have an appetite for requiring, say, carbon disclosure. But our government has no appetite for making subjective, ethical decisions on behalf of asset managers, in terms of what investors should value and how much.
Despite all this, I have unfortunately met many investment managers who are delaying their market entry as they wait for consolidation or greater perfection in ESG data.
The view that ESG data should or will converge is wrong-headed: ESG data dispersion is a feature, not a bug. In fact, the trend-line is toward a marked increase and proliferation of ESG data signals.
This phenomenon is a natural result of the facts that:
A) Many of the decisions outlined above which drive diffusion (scope, weightings, interpolation, and portfolio construction) are fundamentally subjective; and
B) There is a competitive market for ESG data provision, in which differentiation is rewarded, not punished.
For example, for those of us who have been in the industry for some time, we can all recall one association or multi-stakeholder initiative after another, likely backed by some well-meaning foundations, attempting to create a unified framework for corporate ESG reporting. As far as I have seen, each one has failed to gain traction, ostensibly because each player is seeking to force fealty to their home-baked framework, rather than to someone else’s. Such programs defy the fundamental market incentives towards diffusion, and thus fail to reach their stated goals.
In the coming several years, and further accelerating this phenomenon, a number of venture-backed startups, many of whom are using/claiming machine learning approaches, will complement the old guard of ESG data providers. These black boxes will only further mask any potential basis for convergence.
Why the Industry Got It Wrong, and What to Do About It
The competing ESG data providers, their variance, who to trust, etc., is confusing in part because the industry has been looking at it backwards. Only by completely re-imagining the problem can we end confusion for advisors and clients.
How has the ESG industry gotten it wrong? First, it’s important to understand that SRI (Socially Responsible Investing, the predecessor acronym to ESG) was built up in the institutional space. When you hear about tens of trillions moving into SRI/ESG, that’s virtually all institutional assets.
As such, the ESG data industry catered to these institutional asset managers. The name of the game was to generate a jaw-dropping 100-page MSCI report, preferably with maximum snooze-value, grey borders, and fine print, and drop it on the desk of a hapless Mercer associate, with a heavy thud. The hope is no one reads the thing. Instead, we can all “leave it to the experts” and trust they’ve done their homework.
As Sustainalytics, TrueCost, ISS, Asset4, Bloomberg, etc., entered the market, it became an arms race in this direction.
Frankly, this is part of a much larger industry dynamic, in which financial services is defined by a “product-push” supply chain, rather than “client-pull.” Asset managers traditionally invest significant overhead in product development, including paying for ESG coverage, run the fund through a regulatory process, and then ram it out for maximum scale.
This is the opposite of any dynamic retail industry, such as food, fashion, or music, where engaged customers demand value and easily switch providers, driving the supply chain and driving high rates of innovation.
In our industry, the incentives are stacked towards thick reports and bad websites, in the hopes that clients never engage, ask questions, or examine fees.
Fortunately, this is starting to change. Asset managers have competed themselves to zero on product margins. In the future, margins and value will be defined by client engagement, experiences, services, and personalization. For example, you can now make money on financial planning, or steal customers from competitors with digital tools. Meanwhile, product procurement - actual trading or buying ETFs - is free!
In short, asset management is transforming from a product to a service industry.
As wealth clients get more engaged, ESG is moving downmarket and into this new “service industry” zone. If you are an RIA (rather than a fund manufacturer), and you’re sitting here reading this article on ESG data, you are proving the point.
In the wealth channel, we thus need to transcend the old paradigm of ESG data scoring, squabbles, and hefty tombs. That was all built for a product-based institutional asset management industry.
As a financial advisor, you have a choice to get lost in the details of divergent ESG datasets built for portfolio managers, or bypass these debates entirely and focus on what your clients want...which is not ESG data. While the underlying ESG data must be credible, the real focus should be on the storytelling the data produces for the client - i.e. client account-level “impact reporting.”
No one cares that their “E score went from CCC to BB.” What wealth clients want to understand is their personal impact, in relatable and tangible terms. They want credible story-telling. How many trees did my portfolio save this quarter? How much money towards vaccine development did I invest? How many women leaders did I support with my investments?
None of this is in the current ESG datasets. Rather it results from re-bucketing all that ESG data, multiplying it out by the client’s portfolio holdings, and then providing an engaging interface and experience designed for lay people.
I call this the “Client-First ESG Revolution.”
We’ve been seeing and building around this phenomenon for years at OpenInvest. Tangible, concrete, personalized impact reporting has always been our most popular feature on our asset management platform. Due to major demand from other players, we have recently spun our reporting service off as Portfolio Diagnosis (PD).
PD is an automated tool for advisors and platforms to report on the impact profile of client accounts, in relatable terms, regardless of asset managers or vehicle types.
Fortunately, this puts us in a very different position than other ESG data services. When you create value through technology and client experiences, you don’t have a big incentive to hide your ESG data sources and methodology. While most data providers treat their methodologies as proprietary, keeping advisors in the dark, we aim for full transparency. (3) (This is uncoincidentally aligned with our ethos of bringing honesty and transparency to financial services - hence the name, “OpenInvest.”)
Furthermore, such an approach can bypass the ESG dispersion problem, or at least the stress it creates. We know that no one source is authoritative; and in fact, some are distinctly better on some issues than others. Thus, we’ve created a layer for normalizing and ingesting multi-source ESG data. We currently have over 15 different data sources feeding our system, with over 200 indicators running through them, with new sources added constantly. The world is moving towards more and more transparency. This is a good thing for ESG portfolio construction and reporting, and we want to provide the tooling that can take advantage of this.
There are many other benefits to this approach. But for the purposes of this article, the main value is that we can abstract away the diffusion problem. What matters is creating the right kind of “funnel” to ingest all that data, process it appropriately, and then generate something significant and meaningful for normal people, in an engaging format.
Today, investment management is being defined by engaging the client, holistically, credibly, and in real-time, with the information and stories they desire. Let’s help them understand where their portfolio stands on the issues that matter to them most. In turn, they will start to realize that investing is not some opaque, immoral activity, but rather an important way they can shape the world around them, for the better.
And that’s something they’re going to tell their friends about.
- 1) Mayor, T. (2019, August 26). Why ESG ratings vary so widely (and what you can do about it). Retrieved November 19, 2020, from https://mitsloan.mit.edu/ideas-made-to-matter/why-esg-ratings-vary-so-widely-and-what-you-can-do-about-it
- 2) Berg, F., Kölbel, J., & Rigobon, R. (2019, August 20). Aggregate Confusion: The Divergence of ESG Ratings. Retrieved November 19, 2020, from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3438533
- 3) Kumar, R., & Weiner, A. (2019, March). The ESG Data Challenge. Retrieved November 19, 2020, from https://www.ssga.com/investment-topics/environmental-social-governance/2019/03/esg-data-challenge.pdf
- 4) Bender, J., Bridges, T., He, C., Lester, A., & Sun, X. (2019, June 01). A Blueprint for Integrating ESG into Equity Portfolios. Retrieved November 19, 2020, from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3080381