Overcoming the Limitations of ESG Data and Ratings
Leveraging technology to overcome ESG data challenges
Being a newcomer to a market can be a significant challenge, as strong barriers to entry might exist. On June 27, 2022, ESMA published a letter that it sent to the European Commission to present the findings of its Call for Evidence on Market Characteristics of ESG Rating and Data Providers. One of them is that: “For contractual purposes, users mentioned two main key drivers: price-quality ratio and existing commercial relationships with the company that provides ESG ratings”. Considering in addition that contracts run on average over several years, breaking into the market is no easy task.
But taking a fresh look at market practices is also a powerful way to create value. Disruption can deliver renewed offerings with a stronger focus on client needs. It is no surprise that the ESMA letter highlights several limitations of current ESG data and rating providers. These are well known and have already been evidenced in studies from the European Commission, IOSCO and MIT to name a few.
These limitations are the very reason why Clarity AI was founded back in 2017. Overcoming them is part of our DNA. Bringing social impact to markets is one of our goals and to do this required a fresh approach based on technology and transparency. It took us three years to develop the platform that could provide our clients with the capabilities to allow them to form their view on sustainability performance of their investments. We have been constantly striving to improve ever since.
Focusing on the main limitations highlighted by ESMA, here are a few illustrations of how we have tried to deliver relevant solutions:
1.) Data quality, coverage and granularity
We use three main sources of data: external data providers (over 70), our own data collection processes and our own estimation models. This allows us to develop reliable data sets at scale. We compare multiple sources and identify the most reliable one for each data point. Through technology (Machine Learning and natural language processing) we expand datasets and develop new metrics where no reported data exists. To answer new regulatory requirements (such as the reporting of principal adverse impacts, PAIs under the European Union’s Sustainable Finance Disclosure Regulation, SFDR), we develop a specific and tailored approach to best match the regulators expectations.
As a result, our data coverage is two to three times larger than our closest competitors and we provide unique systematic transparency on the source and reliability level of each data point.
2.) Complexity of methodologies, lack of transparency and aggregated confusion
Our mission is to empower our clients so that they can form their own view. We acknowledge the fact there is no one size fits all approach to sustainability assessment. Beyond high quality data, we therefore provide transparent and science-based methodologies that are accessible for each of our modules. Because our clients might have specific sustainability preferences, we offer customizable scoring metrics to ensure those needs are met. For ESG risk for instance, we rely on SASB financial materiality matrix but we offer the ability to adjust each and every weight at pillar, sub-pillar or KPI level, to select between a best in class or a best in universe approach, and to choose specific data relevance thresholds. This functionality is available in one, easy click for entire portfolios.
3.) Inadequate number of skilled resources to ensure timely follow-up on portfolios and error corrections
As a tech company (two-thirds or our 300 staff is tech), we do not rely on human analysts to deploy adjustments at scale. There is no other way to efficiently cover up to 40k issuers and close to 300k funds. This agility also enables us to keep up with the pace of regulatory evolution. For instance, it took us only two months to implement the two new taxonomy environmental objectives on biodiversity and pollution. Every two weeks we automatically update new reported taxonomy alignment figures publicly disclosed by companies. Lastly, relying on robust AI models (Maintained by a team of over 25 data scientists) allows us to implement data consistency checks, to identify outliers and proceed with error corrections when needed.
4.) US-centric biases linked to concentration of main providers
Clarity AI’s holding company is based in the US. Most of our operations staff are however located in Europe (Spain and Portugal). More importantly perhaps, all our data are hosted on servers located in Europe, an important guarantee when it comes to regulation at the European level. Finally, we can balance different perspectives of sustainability: reflecting the more North American view of impact on enterprise value through our ESG risk module as well as the impact side of double materiality as embedded in European regulation through our impact modules. Because we know that biases exist, we offer capabilities in our tools to overcome them (as mentioned above for instance by adapting the scoring universe depending on geography, size or data availability).
5.) Conflict of interest
Our scores are outputs of algorithms and do not rely on qualitative assessments by analysts. The underlying methodologies are transparent and applied systematically. Moreover as illustrated above the scoring processes are customizable by our clients. Finally, not only do we stay away from any advisory services for corporations, we also pride ourselves for being totally independent from any investment solution providers, thus having full ability to focus on delivering best of breed sustainability assessment capabilities.
As ESMA puts it, “The market for ESG rating and data providers is indicative of an immature but growing market”. At Clarity AI we found that the best way to overcome the market limitations to help the Sustainable investment industry grow at scale on a sound basis, was to define ourselves neither as an ESG data nor as an ESG rating provider but as a sustainability tech platform.