In July 2023, the record-breaking heat soared, coinciding with striking images of wildfire smoke enveloping Manhattan’s towering skyscrapers. Amidst these visuals, financial experts criticized ESG metrics as often “subjective, fluffy, and easily manipulated.” Consequently, allegations of greenwashing targeting sectors ranging from oil and gas to finance persist in both media headlines and corporate deliberations.
On one side of the spectrum, there is a push to eliminate investments based on environmental, social, and governance (ESG) principles. Conversely, there are urgent calls for an acceleration in the sustainability movement. However, a consensus emerges on the inadequacies of current methods in gauging the efficacy of ESG initiatives, highlighting inconsistencies and unreliability.
The shift towards a net-zero future looms large on the horizon. Presently, numerous firms, managing assets worth $130 trillion, have joined the Glasgow Financial Alliance for Net Zero (GFANZ), pledging to hasten the decarbonization of economies. As the financial services industry spearheads the drive towards energy transition, there is a pressing need to revamp strategies to instill trust amidst the challenges of ESG measurement. Urgent action is imperative.
Enhancing the Significance of ESG Metrics
Currently, a plethora of third-party ESG data and ratings providers such as ESG Book, Moody’s, S&P, Bloomberg, MSCI, Refinitiv, and Sustainalytics exist. These providers amalgamate hundreds of data points into weighted categories encompassing environmental, social, and governance aspects to generate ratings. The abundance of data dimensions complicates the differentiation between reality and fiction.
For instance, consider the varying overall ESG ratings assigned to a major oil and gas company by three different providers: Provider 1 rates it as “BBB” (average), Provider 2 deems it “good,” while Provider 3 flags it as “severe risk.”
Financial institutions are strategically positioned to bring clarity to the current landscape. A 2022 Dow Jones survey of financial leaders revealed that two-thirds of respondents view ESG investing as a pivotal driver of sustained, long-term growth. However, 52% expressed dissatisfaction with the current quality of ESG data for informing investment decisions, and 58% emphasized the need for enhanced transparency in the development of ESG ratings.
Legal battles, regulatory fines, and reputational harm stemming from climate-related litigations pose tangible risks akin to the rising sea levels. The tally of ongoing litigation cases exceeds 2,000, with over a quarter filed between 2020 and 2022. Fitch Ratings highlighted that in Europe, lawsuits targeting banks on climate grounds could set precedents, compelling banks to expedite their carbon-neutrality strategies and phase out fossil fuel financing.
The imperative to fortify ESG ratings across all sectors has never been more critical.
Harnessing Artificial Intelligence
The ambiguity surrounding ESG ratings persists due to various factors such as siloed data, overlapping information, incomplete self-reporting, absence of measurement standards, and prevalent greenwashing practices.
While ESG scores are aggregated, decision-makers necessitate granular data and transparency to assess the reliability of their business decisions, especially when faced with outliers. For instance, earlier this year, a tobacco company reportedly obtained a significantly higher ESG score than the electric vehicle manufacturer Tesla.
In evaluating ESG data, research analysts, underwriters, and asset managers should have the ability to delve into the data sources, enabling comparisons with other ratings and research. However, a key challenge lies in the fact that ESG scores rely partly on publicly available data, contingent on companies’ commitment to disclosure and transparency. Addressing gaps in data poses a challenge, given the inherent subjectivity and potential biases of human ESG raters. Fortunately, artificial intelligence (AI) offers a non-biased alternative.
ESG ratings predominantly comprise structured, quantitative data, while sentiments and online chatter constitute unstructured, qualitative data. Leveraging Natural Language Processing (NLP) and Machine Learning (ML) can analyze sentiments, discerning positive, neutral, or negative tones from text, audio, video, and image sources. This analysis aids in validating ESG claims, uncovering discrepancies, and furnishing real-time insights to bolster investment decisions at scale.
Navigating Through Noise and Addressing Geographic Disparities
The unstructured data stemming from ESG-related events often introduces noise into the analysis. Integrating Generative AI with NLP sentiment analysis can refine both positive and negative sentiments, mitigating biases in news reports and enhancing precision.
For instance, an investment bank successfully implemented a real-time ESG sentiment system that aggregates insights from over 70,000 global news sources, empowering traders to swiftly grasp the impact of events on companies’ ESG profiles.
Unstructured data plays a crucial role in bridging gaps where structured data falls short. Financial institutions can leverage unstructured information to create proxy data reflecting ESG activities, enabling comparisons against industry frameworks to unearth inconsistencies and exaggerations in ESG data.
The roadmap ahead is clear: To combat greenwashing, the financial sector must delve deeper into ESG performance to gain robust insights, enabling a proactive assessment of risks and opportunities. This approach not only facilitates the transition towards a low-carbon economy but also ensures that the $130 trillion earmarked for financing this shift is allocated prudently.
Anirban Bose, CEO of Financial Services and Chairman of APAC at Capgemini.
