ESG Through a Data Science Lens

Published : 3 years ago, on 15/02/2022

For institutional investors with ESG goals, data is the name of the game. But with inconsistent data sets, no unified evaluation methodologies and a lack of global standards, investors can struggle to gather insights into portfolios. Data science can help make that easier.

By Paul Fahey, Head of Investment Data Science, Northern Trust

Across the industry, we’ve seen an increase in the number of institutional investors who have committed to ESG investing. In fact, according to “The Art of Alpha: It’s All About Investment Data Science”, a 2021 Northern Trust white paper[1] based on a survey of 300 global asset managers, well over half (59%) currently factor ESG data into their investment process, a number that is likely to grow in the coming years

As institutional investors increasingly target sustainability goals, they will need fundamental and diligent analysis at every level: investment processes, compliance practices, organizational design, governance and reporting.

Yet today, an explosion of data has created an environment where both sides of an investment are struggling to keep up:

• Transparency and accountability are the foundation of ESG investing, and investment managers feel the pressure – based on regulations, board accountability or investor demands – to demonstrate that their products are, in fact, green, and not greenwashed.
• Institutional asset owners, meanwhile, have a fiduciary responsibility to their beneficiaries and, in some instances, the public. They need access to data points and metrics to verify a portfolio is achieving stated objectives, to rate ESG factors, and to help manage related risks and compliance.

While today multiple vendors offer access to ESG data, there is a lack of shared standards around which datasets are tracked, how they are tracked, and how to draw out and act on insights from that data.

Without shared industry standards of ESG analysis and reporting, and with data streaming in from multiple sources and in inconsistent formats, often manually tracked in spreadsheets, many questions arise. How can institutional investors make their small sustainability investing teams work smarter? Can they access the ESG data needed to drive portfolio-level decisions? What signals can they derive from the data that will help guide these decisions?

Data science can streamline the broad, non-standardized world of ESG data analytics

Data science allows analysts to sort through huge amounts of information quickly and efficiently. The technology can integrate data from multiple sources and find patterns that help measure, analyze and report ESG investments across key applications.

1. Screening – A fundamental use of ESG data is to screen for appropriate securities. Data science can allow managers to get deeper on their screening, including comparing securities to understand their impacts, especially across the different segments of environmental, social and governance, and across industries. For example, a data science program can compare the score of a credit card company with an auto maker. Clearly the environmental factors of the car manufacturer weigh heavier on its score than that of a credit card company. However, other data points may make one company more appropriate than the other, and the right data program can allow investors to analyze and weigh different factors on a level playing field to understand which one truly is the better fit for a portfolio to meet the investors’ sustainability objectives.
2. Risk and performance analysis – Asset managers need to measure and report on outcomes if they are to meet the expectations of their investors and regulators. The need for better sustainability data will increase as regulators place more demands on investors for reporting on ESG-related risks, much of which differs across jurisdictions. This means being able to provide more granular reporting on ESG holdings. For example, one investor may put special emphasis on gender equality risk while another cares primarily about carbon emission risk. Being able to quickly pull together custom data views on different holdings will be crucial for managers in gaining their clients’ trust and, for that matter, access to their capital.
3. Portfolio optimization and simulation – As investors enter the ESG realm for the first time, being able to understand how changes affect their portfolios on a pre-trade basis is crucial. Data science tools and feeds can allow investors to simulate changes to their portfolios and understand the potential impact on ESG scores.
4. Reporting and benchmarking – Today, the burden of aggregating, analyzing and reporting ESG portfolios falls to investment teams that are already stretched thin. Digitizing this process will allow managers to provide detailed evidence of how they achieve returns and whether the results are based on skill and ESG knowledge. This is crucial for a client base that has access to similar tools, and that is doing comparable analytics. As the datasets become larger and more granular, it will become increasingly difficult for the investment teams, regardless of size, to effectively analyze the data and provide meaningful intelligence without digitization.

To keep up with the many pressures that are emerging around ESG investing, relying on data science to enable decisions and to communicate those decisions to stakeholders will be key.

As ESG increases in focus, regulatory frameworks are likely to grow, and institutional investors will be ahead of the curve if they focus on how to put their data to work now.