ESG Benchmarking: Part #2 - AI, the "New Hope of ESG"

Following my first post on Traditionalists (the disclaimer in that post still stands here!), we are now venturing into AI-powered ratings known as Alternative Ratings or Contemporaries. Contemporaries are heavily reliant on the technological power of data scraping and AI, using sentiment analysis (SA) to come up with ESG benchmarking for companies. SA uses words, adjectives, tone, and aspects of a subject/thing to determine the connotations, opinion, or stance derived from the sentence, document or source of information provided. There are various sources of data to use for Sentiment Analysis, including but not limited to:

  1. Product reviews

  2. Stock market information

  3. News articles

  4. Debates

  5. Posts

  6. Social media and micro blogging (a rich source of information)

There are quite a number of Contemporaries in the market now. However, there is not much information about how things work algorithmically, and different Contemporaries have different approaches. One of the more established contemporaries, TruvalueLabs (TVL). TVL uses various data sources including news, journals, and trade blogs, but avoids social media due to random noise and unreliable data from the source (Hughes, Urban, and Wójcik, 2021). 

How good are they? Can they really replace the Traditionalists?

The big data approach taken by Contemporaries is one of their biggest strengths. They take an “outside-in” approach to ESG rating, and rely on external data scraped off the internet and public sentiment paired with noise (aka controversy) to help determine a company’s rating. This is thought to be “democratic”, since they are not relying on company disclosures (Hughes, Urban, and Wójcik, 2021). 

The use of technology also allows Contemporaries to work with far more data: TVL can scrape data from 100,000 sources daily in comparison to the under 4,000 sources the Traditionalist MSCI would use (Hughes, Urban, and Wójcik, 2021). This wider access allows Contemporaries to have a broader perspective on a company to help determine its ESG rating. 

GreenWatch is an AI tool used specifically to tackle the problem of greenwashing. It verifies the authenticity of “green” claims made by a company by analysing its environment-related statements made in corporate communications, including how forthcoming the company is regarding its stance on climate change, against its carbon scores. Companies are then placed into one of four categories: (1) green leaders; (2) hidden green champions; (3) green incrementalists; or (4) potential greenwashers (Bloomberg, 2021).

GreenWatch is an AI tool used specifically to tackle the problem of greenwashing. It verifies the authenticity of “green” claims made by a company by analysing its environment-related statements made in corporate communications, including how forthcoming the company is regarding its stance on climate change, against its carbon scores. Companies are then placed into one of four categories: (1) green leaders; (2) hidden green champions; (3) green incrementalists; or (4) potential greenwashers (Bloomberg, 2021).

Having access to a treasure trove of external data with a democratized perspective comes with various advantages. For one, it may help to combat the issue of greenwashing faced in the Traditionalist approach, a potential solution to the problem! Additionally, the Contemporary approach may mitigate the problems caused by the subjectivity of ratings and lack of standardization present in Traditionalist approaches by avoiding as much subjectivity as they can in their rating method.

Contemporaries also have advantageous flexibility to help determine ESG ratings. This is due to an important feature of TVL’s algorithm called the dynamic “Impact %” which they use to calculate weightings of key issues for companies: if there are more noise generated for certain controversies, the weighting for a related key issue is adjusted for the company. 

Finally, one very important feature brought by technology is the fact that Contemporaries can keep track of companies in real time. Contemporaries constantly scrape data and can update investors in real time about company ESG ratings and their respective issues. One such case study showed that TVL was able to decrease ratings for a commodity trading company called Glencore in light of various controversies they were involved in, whereas MSCI ratings were only adjusted after their annual review of Glencore (Hughes, Urban, and Wójcik, 2021). This means that investors have better access to more up-to-date information and data to make an investment decision. This means that investors have better access to more up-to-date information and data to make an investment decision. The images below show the controversies picked up by TVL in a relatively short period of time as well as the data points and ratings that were given to the company.

Hughes, Urban, and Wójcik (2021): “Spotlight events for Glencore. Source: Truvalue Labs.”

Hughes, Urban, and Wójcik (2021): “Spotlight events for Glencore. Source: Truvalue Labs.”

Hughes, Urban, and Wójcik (2021): “Insight Materiality trendline for Glencore from the TVL online platform. The dotted line refers to the daily pulse score, from which the Insight score (solid line) is derived as an average. The purple circles refer to ‘spotlight events’, ESG events which create a lot of data points. Larger circles equate to more datapoints. Source: Truvalue Labs.”

Hughes, Urban, and Wójcik (2021): “Insight Materiality trendline for Glencore from the TVL online platform. The dotted line refers to the daily pulse score, from which the Insight score (solid line) is derived as an average. The purple circles refer to ‘spotlight events’, ESG events which create a lot of data points. Larger circles equate to more datapoints. Source: Truvalue Labs.”

TVL provides a good insight of the capabilities and solutions Contemporaries provide to issues faced by Traditionalists. This includes a more democratized and transparent approach, access to far more data sources in comparison to Traditionalists, more objectivity, standardization, and up-to-date information to help investors make better decisions with relation to ESG and investment.

When broadly comparing the Traditionalist and Contemporary approaches to ESG ratings, it seems that Contemporaries win on various fronts: they can update investors quickly and try to be as objective or transparent as possible. However, not all Contemporaries may follow the exact same methodologies and get the exact same ratings. It is all dependent on the data sources they use, the key issue frameworks they choose, and how their algorithms work. 

Unfortunately, Contemporary data is not widely open to the public, and the data mentioned here is all only from TVL. With a lack of access to their data, we cannot undergo certain stress tests to see if there are unique features of Contemporaries, such as picking up on issues that Traditionalists do not see. For the moment, they seem to be able to do what the Traditionalists do with less subjectivity.

Check out my first post on AIs and ESG benchmarking: ESG Benchmarking: Part #1 - ESG, AI, and the Investment Industry

By MUHAMMED Hazim, Segi Enam intern, 23 Aug 2021 | LinkedIn

Edited by Nadirah SHARIF