Slow and fast ESG scores: what do they tell us?

This insight, part of a series with Probability & Partners, explores the convergence of slow and fast ESG scores, highlighting the benefits of combining both for a comprehensive view of corporate sustainability to inform investment decisions. In this third insight we examine:

• The importance of ESG Performance: Combining slow (from corporate disclosures) and fast (sentiment and media-based) ESG scores can enhance investment decisions and detect greenwashing risks.
• Enhancing ESG Assessment: Fast scores, retroactively backfilled to the early 2000s, raise questions about their correlation with slow scores and potential greenwashing risks.
• NLP and AI Use: These technologies analyse news and social media to generate timely, objective ESG scores, offering an alternative to traditional metrics.

Assessing the ESG (Environmental, Social, and Governance) performance of companies has become a routine task for banks and asset managers, particularly for evaluating firms in their lending or investment portfolios. These annual scores are primarily derived from companies' sustainability and annual reports, as well as other, often self-reported, information. While these scores quickly gained traction in the financial world for their simplicity and clarity, a potential issue becomes evident: companies, particularly those with ample resources, may be tempted to overstate their sustainability efforts—especially in the Environmental or Social domains—presenting them in a favourable light. As a result, their ESG scores may not always align with reality, leading to a phenomenon known as greenwashing – a major source of reputational risk for financial institutions claiming to invest in “good” ESG companies.

Recently, a new generation of ESG data providers^[1] has emerged gathering ESG-related information from public sources by analysing the textual content of news outlets and social media blogs. Leveraging tools like NLP (Natural Language Processing) and AI models such as BERT, they generate timely insights on ESG issues, including controversies, and assess sentiment related to Environmental, Social, and Governance factors across a wide range of public companies. These rapidly updated (often daily) ESG scores offer a potentially more objective alternative to the traditional, slower-moving scores.

Although these fast scores only emerged recently, most providers have retroactively backfilled historical data, with archives dating back to the early 2000s. This raises several key questions:

• How do fast and slow scores compare, both now and over recent years? Are they highly correlated, or do they differ significantly? Have they converged or diverged in the last few years?
• What can the alignment or disparity between these scores reveal about the true ESG performance of firms and the risk of firms overstating their sustainability efforts?

These and other questions are addressed in detail in our paper, "Navigating the ESG Landscape: Slow vs. Fast ESG Scores", which also introduces a methodology for creating a greenwashing risk indicator by combining fast and slow scores. However, we can already share some key findings from the paper.

First, it appears that over the past 20 years, the gap between slow and fast ESG scores has narrowed significantly. By 2024, the average difference between the two scores had decreased to under 10 percentage points, down from a 30-point gap in 2005. This suggests that, on average, companies have improved their self-reporting practices and become more adept at communicating their ESG initiatives. However, there are still significant differences in the average score divergence between companies of different industries and sizes, for example, larger firms have higher score divergence and the smaller companies almost none. This could reflect the fact that larger firms have more means at their disposal to invest in communication and PR and to show themselves in a more favorable ESG light.

The greenwashing risk indicator can be developed by dynamically comparing slow and fast ESG scores, either collectively or for individual E, S, and G pillars. When the scores are aligned, and the number of controversies is low, the risk of greenwashing is minimal, as the self-reported data (slow scores) aligns well with public sentiment, as reflected in news and social media (fast scores). In cases where fast scores exceed slow scores—though rare—this misalignment suggests the company may be benefiting from a positive public perception that isn’t fully supported by its actual sustainability efforts. Alternatively, it could indicate that the company is not effectively communicating its ESG performance, or it may be intentionally downplaying its achievements as a strategic choice to under-promise and overdeliver, or to avoid attention until positive sustainability results are assured. Finally, when fast scores are significantly lower than slow ones—especially if accompanied by numerous controversies—this discrepancy could signal that the company is overstating its sustainability performance which warrants further investigation.

The white paper offers many additional insights and examples; however, we can conclude that the emergence of fast ESG scores, when combined with traditional metrics, provides investors and lenders with a clearer view of companies' actual ESG performance, leading to more informed lending and investment decisions. This dual approach enhances the reliability of ESG assessments, promoting better-informed investment choices and improving overall transparency and accountability in ESG practices. Ultimately, the integration of slow and fast ESG scores creates a robust framework for evaluating corporate sustainability, fostering a more accurate and comprehensive understanding of ESG performance within the financial industry.

[1] Discover how LSEG and MarketPsych partner to provide trusted ESG analytics: MarketPsych | LSEG