The Problem with ESG

The blind leading the blind.

Photo by Kier in Sight Archives on Unsplash

Background

A friend recently related an incident at his workplace. His organisation was looking to renew their partnership with an airline for miles to be used as customer benefits. The paperwork on the partnership renewal now included a section on ESG — how the initiative could be “ESG-positive” for the organisation. Based on further consultation, a person from the organisation’s recently setup ESG department mentioned that the miles would affect the organisation’s ESG metrics; they wanted to see if anything could be done to procure “green miles” instead.

If this story sounds familar, it’s time we step up to call bullshit on the ESG movement. As gaslighting goes, ESG stands among the top topics this century. Of late, I have seen numerous articles talking about how ESG is critical to business success, and how data analytics is critical to enabling it. I read an article claiming that we can use big data to track the carbon emissions, employee diversity, etc. of a company. Really? It’s all fluff.

And so I dedicate my 47th weekly article to unpacking the misinformation around this practice.

(I write a weekly series of articles where I call out bad thinking and bad practices in data analytics / data science which you can find here.)

ESG Means …?

For those still unfamiliar with what ESG means, the acronym stands for “Environmental, Social, Governance”. It’s often shorthanded as “sustainability”.

Let’s first unpack each of the dimensions of ESG:

• Environmental = greenhouse gas emissions, energy efficiency, waste management.
• Social = employee diversity, labour practices, community engagement.
• Governance = board composition, executive compensation, shareholder rights.

ESG suffers from the same problem of latent variable definitions (which leads to unstandardised definitions because of the equivocality of interpretation) as what I highlighted in my earlier article on The Problem with DEI. In the Social dimension, sub-dimensions such as employee diversity and community engagement are challenging to define, and harder still to objectively measure or get trustworthy data on. And even if the industry could come to some agreed common definitions (which I doubt), there is still the question of reference points to help us quantify what is “good”, “bad”, and “acceptable”.

Benefits of ESG?

The construct of ESG seems to be motivated by theory and idealism. Take “board composition” as an example. There has been literature written that boards don’t really matter as much as they think they do, given the asymmetric and highly managed flow of information between the organisation and its board. To this day, there is insufficient evidence that board composition (however it’s defined) affects an organisation’s business performance. For example, Perry, T., & Shivdasani, A. (2005) show that firms with a majority of outside directors on the board are more likely to initiate asset restructuring and employee layoffs. And boards don’t represent shareholder interests.

Many of the studies around the impact of ESG are simply stating correlation and not causation. Obviously, successful companies are more likely to have the time and resource to “invest” in ESG — putting in place a reporting process, having dedicated headcounts, etc., thus further reinforcing the said correlation. Large companies are also more equipped to “greenwash” their ESG numbers (i.e. fiddling with the numbers; cooking the books). If ESG data were seriously audited, everyone would fail.

Let’s look at some ESG numbers being bandied about:

1. 54% of banks around the world feature climate-related data in their financial statements (source: KPMG). So what?! Banks have never been the major contributors of greenhouse gas emissions. Sure, they can go paperless, but they would incur the computational carrying cost of digital automation and AI. And if banks actively reduce employee headcounts as a result of digitalisation, should that count towards their reduction in greenhouse gas emissions?
2. 50% of consumers are willing to pay a premium for sustainable brands (source: IBM). Adjusted for the 23% of the world that lives in poverty, you are essentially claiming that 65% of consumers would pay more in a world where prices are already inflating? Seriously? While a recent joint study by McKinsey and NielsenIQ shows modest growth improvements in brands with ESG claims vs none, many retail executives continue to report frustrations that their ESG initiatives are not generating sufficient consumer demand for their products. “There are many stories of companies launching new products incorporating ESG-related claims, only to find that sales fell short of expectations.”
3. 77% of retail investors globally say they are interested in investing in companies or funds that aim to achieve market-rate financial returns while also considering positive social and/or environmental impact (source: Morgan Stanley). This is classic virtual signaling; actual take-up is way below what retail investors claim. 39% of investors in Asia indicate that they have ESG-related investments, compared to 23% in Europe (2023 research from AXA IM); and within Asia, money-savvy Hong Kong stands at 29%. In another vein, based on research from the Association of Investment Companies in the UK, interest in ESG-oriented investing fell from 65% in 2021 to 60% in 2022 to 53% in 2023. All this points to the reality that real investment returns trumps idealism. In fact, a 2023 article published by the Wall Street Journal indicate that “… corporations that remain neutral on social and political issues outperform companies that lean left.”

Conclusion

Beware the buzz! The ESG construct suffers from both data definition and data instrumentation issues. The definition of ESG is so broad that it encourages all kinds of data interpretations and data manipulations. The fact that there are 140 ESG data providers should be a warning sign that things don’t add up. Much of ESG data is voluntarily disclosed, and hence, asymmetric and non-objective. And this is the primary source of inputs for many of these ESG data providers; they are additionally trawling the internet for news articles, and filling in the many data gaps with their own proprietary estimates. As a data analyst, I wouldn’t touch it with a 10-foot pole.