Social distancing: Putting a $ value on the ‘S’ in ESG

While the roots of environmental, social, and governance investing date back to the 1960s, it’s only in the 21^st century that the practice has seen more widespread adoption. So it’s not hyperbole to say that ESG investing is still in its infancy. And of the three pillars, “social” is proving to be the problem child.

Bérénice Lasfargues is an expert in the field of ESG, as she holds degrees, certifications and real-world experience related to sustainability and the environment. For more than four years, she’s worked at BNP Paribas Asset Management, rising up to sustainability integration lead, where she’s played an important role in building out the firm’s sustainability and impact scoring methodology. Frustration with data is a consistent theme among asset managers. But this is especially true when it comes to quantifying that pesky “social” silo.

“Social is definitely the most difficult to analyze,” Lasfargues says. While the environmental and governance factors are well established and easy to understand, the social element of ESG remains the least well understood of the three. That said, there are signs of growth in the sector as a result of the Covid-19 pandemic, when factors affecting human capital could be clearly linked to the health of companies and the broader economy.

“The challenge of social has always been data,” she says, noting that social data lacks much of the specificity of environmental and governance-related datasets, which have targets, requirements, and quotas. “But with social, there’s a higher challenge around data gathering and standardization. And when companies report, sometimes they only report part of the story. For example, if a company reports how many women it employs, you need to be able to see that at the board level and at all levels, and you would also want to see turnover rates.”

ESG data overall suffers from a lack of standardization, with data providers using individual and proprietary methodologies and definitions, making it hard to compare like-for-like ESG investments—one of the reasons that firms like BNP Paribas Asset Management and more specialized asset managers like RadiantESG Global Investors collect data themselves and build their models and scores in-house from scratch.

Mapping, say, a factory in Thailand back to its parent company and products is not easy. We need to be able to map data back to a company to be able to include it in our investment process, and mapping supply chains can be very difficult

Kathryn McDonald, RadiantESG Global Investors

Broadly speaking, while environmental includes climate-related issues, and governance covers corporate and structural inputs, social data refers to the human aspect of the ESG equation, covering topics such as diversity, gender balance, equal pay, workplace health and safety, and modern slavery or human trafficking among companies in a supply chain.

Perhaps it’s easier to understand that a company that has factories in areas that are subject to increased flooding or wildfires as a result of climate change will experience supply chain disruptions, which would hurt the business. There are also numerous studies that show that having diversity at the management level leads to improved company returns.

But while people say they care about things like workplace safety, and are against modern slavery, do these convictions actually change their spending and investing habits? There’s evidence that this is indeed increasingly the case. Take, for example, UK fashion outlet Boohoo.

In 2020, Boohoo shares fell from £4.13 to £2.29 (from $5.15 to $2.87 at 2020 exchange rates) on the London Stock Exchange over one month, following revelations that workers at the UK-based factories manufacturing clothes for the retailer were paid £3.50 ($4.47) per hour—less than half the 2020 minimum wage for adults over 24 years old. Additionally, those workers were being forced to work 12-hour shifts, or risk being fired, despite Covid-19 lockdown rules.

In the immediate aftermath of these revelations by The Sunday Times newspaper, Scottish investment manager Abrdn (then known as Standard Life Aberdeen) divested its shares in the company, and other retailers stopped carrying the brand’s clothes. One year later, human rights and workers’ advocacy groups said that despite assurances from Boohoo, they could find little evidence of improved worker protections in the company’s supply chain. Boohoo’s share price has steadily declined since, and currently trades around £0.35 ($0.45).

While ESG investing is becoming increasingly politicized in the US, new rules and standards are helping to advance sustainability efforts, so investment firms can’t ignore the space. And while social data lacks standards and struggles with consistent definitions, some say the “S” may be the most impactful—and potentially valuable—of the three.

“Governance is the pillar that has been around the longest and is best understood. It comes primarily from company filings and regulatory reports. Information is easily and cheaply accessible, most of it is audited and is widely used. It’s well-trodden ground,” says Kathryn McDonald, co-founder and head of investments and sustainability at RadiantESG Global Investors. “Meanwhile, environmental has gotten the most attention over the last 10 years, given the existential threat of climate change and the realization that if we don’t change course immediately, we’re in deep, deep trouble—not just economically.”

The other thing about environmental data, she adds, is that there’s a lot of it: a lot of raw data, numbers, scores and ratings, and other ways to quantify environmental impacts. And it has been collected for decades by scientists.

In contrast, the social element of ESG has not been systematically or quantifiably captured compared to its counterparts, may be reported using different formats and definitions, and is unaudited, McDonald says.

DIY for ESG

For RadiantESG, the decision to pursue the “do-it-yourself” route stemmed from needing to ensure the data was reliable and unbiased.

“When we started Radiant, our intention was to bring as much ESG and impact material upstream into our investment process, and to bring that information to bear when picking stocks,” McDonald says.

And while vendors’ models are all “thorough and well-constructed,” she says, each uses subtly different materiality mapping and different underlying data, and can have different targets—for example, some may be more oriented toward managing risk while others are more geared toward generating alpha. The result is scores and sub-scores that don’t agree.

“So, you can have really good models that don’t agree—and that’s fine: After all, we don’t expect every analyst to give the exact same estimates for GDP growth or company earnings,” McDonald says.

But Radiant wanted to ensure it had a definitive result it could trust. “This is why we build our own view of a company from scratch, and why we go out and source raw information. We search far and wide for data. We have relationships with all vendors, but they largely get their information from the same places, such as company reports. And government data is very hard to work with because it’s generally exceptionally poorly organized and often doesn’t come with identifiers,” she says.

In short, finding and collecting the data is no small task. But that’s only half the story. Then, there’s the process of mapping information to companies.

“Identity mapping is not straightforward. Mapping, say, a factory in Thailand back to its parent company and products is not easy. We need to be able to map data back to a company to be able to include it in our investment process, and mapping supply chains can be very difficult, especially because of the opacity of private companies in the mix,” she says. “With public companies, we can usually find things out. But once you hit a private company, the trail goes dark.”

And if Radiant can’t find sufficient data on a company, it doesn’t take a position.

Richard Rothenberg, CEO of big data company GlobalAI, which provides insights into financial and ESG factors, says his company is working on alleviating the challenge. First, its platform maps a company and its organizational structure, and its supply chain—no small task in itself. Then it maps the names of senior executives, topics and sub-topics related to the company’s business, which it uses to train its AI to recognize semantic similarities.

The AI also uses multiple languages. “For many multinationals, sometimes 50% of revenues come from overseas, and in those regions, maybe 65% of the data is not in English, and there can be a time lag before that becomes available in mainstream US sources … and gets priced into a security’s valuation,” Rothenberg says.

Forest for the trees

Not only is social hard to measure, but it’s also hard to see what the target should be.

For example, another factor in predicting performance over time is to look for evidence of action vs. statements of intent. That’s another reason why RadiantESG builds its own datasets: because this is a data element that vendors may not provide. For instance, an example of intent would be whether or not a company has a policy on workplace safety or modern slavery, whereas an example of action would be practical steps taken to reduce employee injuries at work, or to identify and weed out any human rights violations.

“Ultimately, what we’re trying to understand is not just where a company says it is today and how it ranks compared to its peers, but where it’s going—its path of travel, and its impact,” McDonald says.

BNPP also developed its own ESG ratings system, after finding no correlation between the data and ratings offered by vendors.

“Taking an off-the-shelf rating schema really limits you. So, we have proprietary ESG sourcing. We get raw data from providers and we are always trying to find better-fit data. Currently, we use three providers and do our own normalization,” Lasfargues says.

The firm has its own internal ESG data program, with specific scoring for different types of asset classes, and which looks at each element of ESG separately.

“We are heavily invested in data and creating data,” she says. “The value is in drilling down. As a whole, ESG is indicative. But for a portfolio manager, the value is in drilling down into the E, the S, and the G individually, looking at themes, and then at the indicator level. A lump sum ESG score with embedded biases isn’t always the best thing.”

SocialMarks analyzes all elements of a company’s “social” activities around three core pillars—people, brand, and community—to score and rank companies against their competitors. Its co-founder, Grady Lee, says the “E” of ESG is the easiest to quantify: “The target is to get to zero.” While easier said than done, there’s a clear-cut goal. Social and governance, he says, is a bit more ambiguous.

“In ‘S,’ there’s no absolute, and nobody knows what they’re doing: Say your company contributes 1 million hours of volunteering—well, is that good or not?” he says. “And when we looked at everything that was dumped into social, it was just this whole bunch of data on things like diversity and inclusion. It wasn’t actionable.”

According to a recent survey by the Index Industry Association, 81% of asset managers say ESG has become more of a priority in their investment strategies over the past year, and 62% of respondents say they incorporate social factors. However, more than half of the asset managers surveyed find it hard to evaluate the “wide-ranging and intangible nature of S factors,” according to the report.

“We found ourselves in meetings with asset managers who were asking what social impact means,” says Mark Webster, managing director at SocialMarks. “For example, they want to know the social risk that a client might expect.”

The vendor collects its data from a variety of sources, including websites, company reports, proxy statements, and board meetings, and over the past year has built its rankings into a product and enlisted London-based data strategy and sales consultancy Eose to help it market the data to financial firms, and integrate its API with desktop data products already used by clients in the financial markets.

“They have a desktop that’s typically used by investor relations professionals to benchmark their organizations against others. We want to bring their data to a wider audience,” says Eose CEO Suzanne Lock. Beyond its uses for investor relations and monitoring sustainability and diversity internally, it also has potential to help investors and advisors make more informed decisions, for research and analysis, and for assessing risks and opportunities. “The ‘S’ in ESG is still underserved. Most data is very broad but very shallow.”

The Index Industry Association’s survey notes that the “data vacuum” identified in its 2021 report is “diminishing,” but that “significant challenges still remain.” Unsurprisingly, a lack of data standards and a lack of agreed ratings and methods among providers “continue to frustrate the search for more accurate, relevant and timely measurement.”

GlobalAI’s Rothenberg says many companies use their own definitions of ESG, so the meaning of “S” can differ among firms. Think about it this way: order a pint of beer in the US, and one in the UK. The English “Imperial” pint is 20 fluid ounces, 4 ounces larger than its American equivalent. Same name, same measure, different amount.

“Some of these definitions are qualitative, some are quantitative, but even where quantitative indicators exist, the units may be different—and that creates a comparability problem,” Rothenberg says.

With limited data available for many social factors, and with different definitions rampant, a full analysis requires deep research. For example, Rothenberg gives the example of training: What type of training is it? Is it training for an employee’s current role, or training that builds their career? How much money does an employer spend on training per employee? Or how much does an employer spend on employee health and safety versus their rates of injuries or litigation?

ESG, in general, has always been difficult waters to traverse for the buy side. In 2021, Mary-Catherine Lader, then the head of Aladdin Sustainability at BlackRock, explained how the data that fuels ESG investing needs to improve. It’s nothing new. But as new regulations require more granular detail into ESG disclosures, the data environment will improve. The last bastion of that evolution may well just be social, but for those who can corral this information earlier than others, the benefits can yield a clean understanding of a company’s future potential or demise … just ask Boohoo investors.