Beyond ratings: Vendors look to fill ESG data gaps

Sam Duncan and Dinah Koehler founded an environmental, social and governance (ESG) data start-up in 2018, but before that, both worked in firms that focused on impact investing—Leapfrog Investments and UBS’s sustainable equities team, respectively. In these roles, they struggled to obtain quantitative data that gave a holistic picture on ESG performance or risk of assets in impact portfolios.

“When we went on the search for quantitative facts on absolute CO2 emissions, for example, or absolute water usage, absolute waste generation, absolute people with access to financial services, absolute gender diversity—even at board level and on management teams—it’s really hard to find those facts,” Duncan says. “They are behind a wall of A’s and B’s—of ratings, basically.”

These were the frustrations that led the two women to found Net Purpose, which closed a funding round late last year led by specialist venture capital (VC) firm Illuminate Financial. Net Purpose is among many early-stage and established data vendors looking to exploit niches in investor demand for ESG data—demand driven largely by an unwillingness on the buy side to rely solely on the ESG ratings available from large data vendors like MSCI, Bloomberg, and Morningstar’s Sustainalytics.

“What the buy side wants is the underlying data [behind ratings] so they can have their own view and build their own internal scores. They don’t want to use the same references as everyone else, because then they have no IP or thought process of their own,” says Paul Sinthunont, a senior analyst with Aite Group.

Asset and investment managers say that while the ESG rankings from the big data vendors are useful as reference points, and they buy them, these ratings are limited as they are opaque and uncorrelated. Mason Gregory, associate director of investment solutions at MFS Investment Management told WatersTechnology earlier this month, “We are more interested in raw data than we are in ESG ratings. We would rather draw our own conclusions based on raw company data that is either disclosed by the companies themselves or collected by third parties like NGOs.” 

Sinthunont this month finished a survey of 52 buy-side firms, mostly in the US and Europe, and of various sizes. He says the firms in his sample consume data from at least two or three vendors. Some of the less traditional vendors that appeared in his survey results included Yves Blue, which has a platform for portfolio analysis; climate risk specialist Urgentum; and Equilar, which counts some of the biggest banks and asset managers, including BlackRock and BNY Mellon, among its customers. Equilar is a provider of corporate governance data, and has partnered with FactSet to make its data feeds available on the FactSet client marketplace.

Vendor offerings

Corporate ESG reporting has improved over the past five years, Sinthunont says, so vendors are not looking to provide data that delves deep into particular companies. Rather, “it’s about getting wider coverage on companies that don’t report. There are regions and sectors that have regulations around reporting and those that do not, and there are no guidelines on what to report—so, for example, emerging markets and in private markets,” Sinthunont says. 

A significant amount of a portfolio can be intangible with ratings data, he says. 

“A lot of the providers say, ‘OK you have a fixed-income portfolio, an equity portfolio: We can measure your ESG risk score. But here is this other 30 percent we are not touching.’ That is not very useful, because you aren’t getting a holistic view of what your impact is or what your ESG risks are,” Sinthunont says. 

It is these gaps that vendors are now looking to fill, with more and more specialized offerings. Data vendor Preqin, for example, a specialist in the alternatives industry, has started to offer ESG solutions, leveraging its private market experience to offer data for a sector where there are few incentives for participants to make disclosures. 

“To differentiate themselves, vendors no longer say, ‘We have the best coverage across the globe.’ Now it’s about specialization: We have ventured into the private asset space. Or you can upload your portfolio and we have a lens we can put over that to help you understand your exposure to, for example, the UN Sustainable Development Goals (SDGs), or understand how many of the companies you invest in get their money from fossil fuels,” Sinthunont says.

Net Purpose is trying to differentiate itself by providing data on social and environmental performance in quantitative form. It helps that corporates are better at disclosing this information now than they were 20 or even 10 years ago, with 90% of the S&P 500 doing so, Duncan says. “We are now grabbing those quantitative facts and putting them on the platform,” she adds.

Net Purpose extracts data from thousands of source filings from companies’ sustainability reports, their 10K filings, their websites, and from other providers such as the Carbon Disclosure Project. The vendor’s subscribers can access this data on its platform. The extraction process is done with artificial intelligence (AI), but the company has found that it needs to retain human analysts for reliable data.

‘Throwing AI at it’

Net Purpose’s product manager, Steve Cheng, says providing ratings has been one way that vendors have approached ESG. Another way is “throwing a lot of AI at it”—using technology to analyze massive unstructured datasets, like filings.

In the experience of Net Purpose, at least, getting quantitative data this way is a tougher technical challenge than these vendors will admit. The challenge is that the vendor’s AI scrapes data from unstructured sources like PDF documents, but not only is it looking for, say, the word “emissions” (an easy enough task), it also has to extract the quantitative value associated with those emissions: How many metric tons of carbon was that? And was it in fact metric tons, or was it cubic tons? Was it thousands or millions of cubic tons?  

“The idea is that if you give the AI a big enough training set, you will get meaningful results. But it’s actually hard to get accurate data through that kind of method. We learned quickly that it’s important to have a human in the loop to deal with all the exceptions to this data, because the data we are dealing with is complex,” Cheng says.

This complexity is partly because corporate-reported ESG data is rather like financial data in its own early days of reporting, Cheng says: “It’s got one foot in the marketing camp.” In other words, companies can still choose to disclose what they like, and of course they are going to report what makes them look good to investors.

The Net Purpose AI can flag anomalies or exceptions in reports—for example, if a corporate filing reports a dramatically smaller carbon footprint than it did the year before. But it takes a human analyst to understand if this has happened for a legitimate reason (perhaps the reporting company implemented a lot of cleaner technology in that year, or divested itself of a carbon-intensive business) or not.

“It needs a human to say, ‘Well, this has jumped by 200%. But I noticed in the company filings that they have introduced some new tech or done some M&A.’ So we have humans in the loop to do that last sense check,” Cheng says.

Net Purpose is running more experiments this quarter to see if it can automate more of the quality assurance workflow, he adds. 

Portfolio analysis

Another start-up on the Illuminate roster is YvesBlue, which closed a $2 million seed funding round late last year. Founder Anna-Maria Wascher also founded an investment advisory firm called Flat World Partners, which specializes in impact investing. YvesBlue was born of her frustration in realizing that she knew of no solutions that could actively track the portfolios her team constructed for clients.

Users of the YvesBlue platform can upload their portfolios to it, and assess them against standards like the UN SDGs, carbon emissions standards and other indexes.

Wascher says the company is positioning itself as a cost-effective alternative to the teams of analysts that both banks and investment managers pay to understand the ESG risk in their portfolios, and the consultants they pay to help them with reporting under regulation like the EU’s Sustainable Finance Disclosure Regulation.

“Large institutional teams are spending millions to build solutions in-house themselves, or they’re piecing them together with consultants, in Excel spreadsheets or PowerPoint, and working with large data providers to implement reporting,” she says. “Wealth management clients want to understand carbon budget, animal cruelty, alignment with the UN Sustainable Development Goals and so on. That requires firms to pay for multiple data providers, and this is what YvesBlue is trying to simplify,” she says.

Advantage: Big boys

The large data providers have the advantage over the upstarts, Sinthunont says. As ESG investing matures, investment professionals increasingly want to view ESG data in the context of more “traditional” financial data, rather than in isolation. “ESG is no longer just lying with the responsible investment team, and someone occasionally asks them for data or reports. Investment professionals now expect to look at financial and ESG data at the same time, on the same platform. That’s a lot more convenient than having to log in or alt+Tab between different screens,” Sinthunont says.

The incumbents are already looking to build holistic solutions like this, he says, partly through a spate of acquisitions.  

“The next level is to look at ESG on a multi-asset level. We still see a lot of trouble with things like investment risk, where you have a multi-asset portfolio investing in equities and fixed incomes; maybe there is some private debt in there, real estate investment trusts. Having the ability to aggregate that, and have an ESG model that is overarching, is difficult. The holy grail is to get that: traditional investing concerns with ESG on top, in one view, and on one platform,” Sinthunont says.