Why AI can overcome the ESG data challenge
ESG analysis remains a challenge for investors. As the recent staunch criticism of the EU’s sustainable finance taxonomy demonstrates, there is still no consensus on what exactly makes something ‘ESG friendly’. And with no standardised disclosure requirements, ESG data varies from company to company. Further, large companies often self-report on their supposed ethical or ‘green’ practices, leaving investors at risk of being exposed to greenwashing.
Some solutions to this problem have surfaced, but they are ultimately still subject to fault. Many rating providers have started providing generalised ESG ‘scores’ for companies. Unfortunately, these do not provide absolute clarity for investors, as there can be large discrepancies between how different ratings agencies perceive the ESG credentials of different companies. Tesla, for example, is a heavily debated ESG stock – it has received praise for accelerating the use of electric vehicles, but also criticism for its working environment and labour practices.
Intelligent ESG analysis
Instead of waiting for ESG data to become standardised and more readily accessible, investors should, in the meantime, be focusing on how best to assess a potential investee’s green credentials. The use of artificial intelligence (AI) has emerged as a solution for this analysis challenge and is quickly transforming how asset managers find ESG opportunities.
For analysing transitional issuers – corporates that are not yet leaders in ESG, but that have a sustainability plan in place and evidence of meeting ‘green’ goals – fundamental analysis is necessary. To paint a full picture of whether a company is a good or bad corporate citizen, an investment manager must dig deeper than surface-level commentary from a company, which may contain promising ESG-relevant buzzwords, and instead trawl through data from various sources. ESG data is currently scattered across multiple sources with no unified standards, which makes it particularly difficult for investors to dig through these documents.
As the name suggests, AI brings a level of intelligent analysis that can be invaluable when it comes to looking beyond the numbers and assessing a company qualitatively. AI can take those scattered documents and search for concepts rather than specific words to save time, in the process omitting piles of irrelevant information. AI can also make use of sentiment analysis algorithms, which allow for analysis of the tone of a conversation. This feature can be used to ascertain how serious a company’s management team is about ESG by analysing transcripts of presentations from an earnings call, for example.
Marsham already uses AI to enhance its ESG analysis. Sevva.ai, a disruptive research assistant, employs artificial intelligence and machine learning to minimise the time we spend sourcing and filtering through information, allowing us to make better decisions faster.
Sevva.ai can read and understand thousands of documents in an instant, and then collect, verify and summarise this information. It forms a key part of our fundamental assessments for our portfolio companies, as it can drill down into the underlying sources that power the ratings to offer a real-time analysis of a corporate’s ESG status. The use of AI saves us around 70% of the time we would usually spend on investment research, which in turn reduces our costs – and these savings can then be passed onto clients.
ESG in its current form is still a relatively young phenomenon. Investment managers must continue turning to AI in order to accelerate our understanding of it, collecting and analysing more ESG-related data than is possible manually, in just a fraction of the time.