With the prominent focus on ESG, asset managers find themselves in a challenging position: On the one hand, they rely increasingly on ESG ratings, not only for meeting regulations but also for establishing trust and satisfying the growing customer demand for sustainable products. On the other hand, ratings based on the evaluation of past data can always only be retrospective. Real-time signals and their impact on a company’s sustainability cannot be considered. While the traditional approach continues being relevant this leaves space for improvement.
CSSF has repeatedly highlighted the importance of sustainable finance for building confidence in the financial sector and positioning it for the future. Their recent survey on the appetite for sustainable investment products in Luxembourg, commissioned together with Fondation ABBL pour l’Education Financière and the Luxembourg Sustainable Finance Initiative, is just one example of various activities currently underway. Also on a global level, CSSF is working together with its peers to create a financial system that serves the world.
Mastering the data flood
The advance of artificial intelligence, or AI, is about to change things. The technology is already there, and now the industry will have to imbed the new tools into their processes. So far, ESG investors are confronted with a flood of often unstructured data that requires manual screening for potential risks and opportunities. Ratings are conducted by the established rating data providers who predominantly focus on large listed companies. A sustainability assessment in the context of a more moderately sized enterprise, e.g. a microfinance project, is therefore much more difficult as structured data on micro or individual entrepreneurs is usually not available via the established data providers. Other options, such as the direct collection of information on site, are often not possible or not economically feasible.
A further problem is that many structured indicators are updated monthly, quarterly or even annually. While investors have been able to follow stock market prices in real time for a long time now, sustainability aspects and risks are usually assessed with a considerable time lag. Therefore, investment decisions are often based on outdated information. Also, ESG ratings are largely based on information published by the rated companies themselves and sometimes have a corresponding bias. Even though more and more rating agencies conduct their own research, the sheer amount of information cannot be managed manually.
From the investor’s point of view, the goal is «T+0» or «Today + zero», in other words: up-to-date information without time delay. AI promises to speed up the information gathering process even in fragmented information markets.
NLP for customised ESG analysis
Algorithms can learn from large data sets which combination of letters or words is typically relevant in a certain context. This type of text analysis is called Natural Language Processing, or NLP for short. A system that Universal Investment is currently developing with partners and that is currently in testing with pilot clients independently analyses more than 500,000 articles from more than 100,000 online sources such as newspapers and online magazines, news agencies, legislative initiatives, announcements and others. Out of all these sources, the system filters ESG-relevant information and classifies it into different topic areas.
All individual events surrounding a specific company are recorded and weighted. For each event, an impact duration defines for how long the event will be included in the overall score. Based on all weighted events, the system then calculates a news-based overall ESG score and calculates it separately for the three categories Environment, Social and Governance. The system further produces a specific score for each sub-segment. The user can seamlessly move between an overall overview and a detailed analysis according to their information needs.
The resulting ESG score provides a real-time indicator of a company’s exposure to relevant events that impact its ESG rating. A breakdown of individual news articles helps to make the score comprehensible and the information behind it accessible.
There is a strong need for real-time ESG information, and the race in product development is in full swing. At Universal Investment, we have laid the foundation in 2019. We are convinced that substantial change in ESG assessment and reporting provides value to both asset managers and asset managers, and that this change will occur in the next years.
Martin Groos is a member of the management board at Universal Investment Luxembourg, a knowledge partner of Investment Officer Luxembourg.
Related articles on Investment Officer Luxembourg:
- EU places ESG reporting on par with financial data
- In-house valuation teams add value to AIFM clients
- Sofia Harrschar: Investors find stability in alternatives