Quantifying ESG – is it possible, useful and long-lasting?

How best to define environmental, social and governance (ESG) metrics in the context of quantitative trading, the role of regulators and data providers in standardising ESG ratings, and indeed their usefulness or otherwise and their life expectancy was debated on the opening day of the QuantMinds International conference on 13 November 2023 in London.

Panellists from the buy- and sell-side at LBBW Asset Management, Lloyds Banking Group, SMBC, and a Professor of Financial Engineering at the University of Leuven joined the Climate Finance & ESG Summit stream in London to assess if ESG ratings would last 2, 5 or 10 years and what form they might evolve into as the green agenda advances? 

“I hope the ESG moniker lasts for as short an amount of time as is possible,” thundered one of the panellists, as he encouraged the industry to move to numbers that make more sense and to be more modest in applying them to reduce the risk of greenwashing.  

“It’s getting very political and that is not our job. Managing risk is our job,” a panellist reminded the quants in the room.  

In addition, there is “a systemic concentration risk” if you push everyone into the same limited pool of ESG approved instruments. “It’s giving me flashbacks to the 2008 financial crisis,” continued the panellist. A diversity of options would be better.  

Another panellist agreed, as he outlined how his financial institution (FI) splits investment opportunities into those covering ‘brown’ companies and ‘green’ companies versus a climate change metric, in order to provide this diversity. This allows investors to pursue short, long or targeted climate change strategies, as they see fit. Sustainability can more easily be dialled up as a metric in this approach as well. 

E to S & G and the data issue 

The panellists’ unhappiness with the present situation mainly stemmed from a belief that comparing the E with the S and G elements wasn’t necessarily always useful, especially if scores are aggregated into an overall ESG rating, which one panellist said he “had an issue with”.  

Aggregation doesn’t help those that want to focus purely on environmental and climate change risk factors, or indeed sustainability as a standalone element.  

“How can you compare women in the boardroom v emissions scores,” queried another panellist. 

Additionally, concerns were raised about the data quality in the ESG arena, where some data points are easier to score than others such as a temperature rise score, which can be notoriously difficult to nail down accurately. 

Artificial intelligence 

A lack of common global standards was also discussed, alongside the role of artificial intelligence (AI) in this area, although one panellist didn’t think AI had a role as, “you could be putting a black box on top of another black box”. By which he meant the ESG quantification process was in his opinion too opaque already without factoring in any AI-driven bias or lack of accountability. 

However, another panellist countered that you could use large language models (LLMs) as a cheap way to validate an ESG rating from a provider and get some extra cheap data. 

The third panellist thought that what he called advanced statistics in his PhD student days was just being renamed AI in this context. “Whatever you call it, and machine learning (ML) techniques can be applied, you could deploy the technology to make an impact in scoring biodiversity for example vis-à-vis supply chain resiliency and so on.” 

ESG life expectancy 

Another panellist commented he thinks ESG will stay. “But we may not call it ESG in the future, as this field evolves still further.”  

“Regulators also need to focus more on establishing a common baseline to provide investors, traders, and others with clearer compatibility metrics.”  

Non-financial factors are presently overlooked, in the opinion of another panellist that joined the debate at QuantMinds International.  

Attendees were ultimately reminded to mix their own firms’ assessments with credit rating agency (CRA) information, solvency and debt data, and ESG ratings from data providers like MSCI or ISS. 

“You shouldn’t rely solely on a single rating or element. ESG ratings can, for instance, give you a good longer-term view in regard to the transition towards a greener economy.” But they aren’t the only metric, and as can be seen by the debate, they aren’t perfect yet either.