Myths vs. realities: Sustainable finance data & ESG ratings reform — what the evidence actually supports

Global ESG assets under management reached $40.5 trillion in 2024, yet a 2025 MIT Sloan study found that the correlation between ESG ratings from the six largest providers averaged just 0.54, roughly half the agreement level of traditional credit ratings (Berg, Koelbel, and Rigobon, 2025). This divergence has fueled a cottage industry of myths: that ESG scores are meaningless, that regulation will fix everything, that artificial intelligence will replace human analysts, and that ESG investing always sacrifices returns. The reality is more nuanced, and practitioners who separate evidence from noise will be better positioned to navigate a market undergoing its most significant structural reform since the introduction of IFRS standards. This article examines the most persistent myths in sustainable finance data and ESG ratings reform, testing each against peer-reviewed research, regulatory outcomes, and operational data from leading financial institutions.

Why It Matters

The stakes in ESG data quality extend far beyond academic debate. In 2025, global sustainable fund flows turned negative for two consecutive quarters for the first time since 2018, with net outflows of $12.4 billion driven partly by investor confusion about what ESG labels actually guarantee (Morningstar, 2025). At the same time, regulatory mandates are accelerating: the EU Corporate Sustainability Reporting Directive (CSRD) requires approximately 50,000 companies to report against European Sustainability Reporting Standards (ESRS) beginning in fiscal year 2024, while the UK Financial Conduct Authority finalized its ESG ratings regulatory framework in late 2025. Japan's Financial Services Agency and India's Securities and Exchange Board have both introduced ESG disclosure and ratings oversight rules effective 2025 and 2026, respectively.

For asset managers, corporate sustainability teams, and compliance officers, distinguishing myth from reality is operationally critical. Misunderstanding what ESG ratings measure leads to mispriced risk. Overestimating regulatory harmonization leads to duplicated compliance spending. Underestimating data quality challenges leads to greenwashing allegations that now carry financial penalties: the European Securities and Markets Authority (ESMA) issued 23 enforcement actions against fund managers for ESG-related misrepresentation in 2025 alone.

Key Concepts

Before examining specific myths, practitioners need clarity on several foundational distinctions that are frequently conflated in public discourse.

ESG ratings versus ESG data: Ratings are opinions, analogous to credit ratings, that aggregate multiple data points into a single score or grade. ESG data refers to the underlying metrics: carbon emissions, board diversity percentages, water withdrawal volumes, and hundreds of other indicators. The quality challenges in each domain are fundamentally different. Data suffers from measurement and reporting inconsistencies; ratings suffer from methodological divergence in how data is weighted and interpreted.

Single materiality versus double materiality: The ISSB standards (IFRS S1 and S2) adopt a single-materiality approach, focusing on sustainability issues that affect enterprise value. The EU ESRS standards adopt double materiality, covering both financial impacts on the company and the company's impacts on society and the environment. This is not a minor technical distinction: it determines what data gets collected, how it is weighted, and what ratings are designed to measure.

Disclosed versus estimated data: A 2025 analysis by the Carbon Disclosure Project found that only 38% of Scope 3 emissions data in corporate reports was based on primary supplier data, with the remainder estimated using spend-based or industry-average models (CDP, 2025). The gap between disclosed and estimated data is central to many myths about ESG data accuracy.

Myth 1: ESG Ratings Are Meaningless Because Providers Disagree

The myth: Because MSCI, Sustainalytics, ISS ESG, S&P Global, Bloomberg, and Moody's frequently assign different ratings to the same company, the entire ESG ratings enterprise is fundamentally broken and investors should ignore ratings entirely.

The reality: Rating divergence is real and well-documented, but it reflects methodological differences in scope, measurement, and weighting rather than random noise. Research published in the Review of Finance in 2024 decomposed the divergence into three sources: scope (which categories are measured) accounts for 38% of disagreement, measurement (how a given category is quantified) accounts for 46%, and weighting (how categories are combined) accounts for 16% (Berg, Koelbel, and Rigobon, 2025). This is not meaninglessness but rather the consequence of providers answering subtly different questions.

MSCI's ratings, for example, are explicitly designed to capture financially material ESG risks relative to industry peers. Sustainalytics measures unmanaged ESG risk exposure. S&P Global's Corporate Sustainability Assessment evaluates management quality on material sustainability topics. An investor who understands these distinctions can use multiple ratings as complementary inputs rather than expecting consensus.

BlackRock's Investment Stewardship team demonstrated this approach in practice, using a multi-provider ESG data framework that draws on four ratings providers plus proprietary analysis. The firm reported in its 2025 stewardship report that portfolio companies flagged by two or more providers for governance concerns were 2.4 times more likely to experience material value-destructive events within 24 months than those flagged by only one provider (BlackRock, 2025). Convergence across providers, where it exists, carries strong predictive signal.

Myth 2: Regulation Will Harmonize ESG Data and Solve Quality Problems

The myth: Once CSRD, ISSB, SEC climate rules, and other regulatory frameworks are fully implemented, ESG data will become standardized, comparable, and reliable, similar to financial accounting data.

The reality: Regulation is improving disclosure coverage and establishing minimum quality floors, but it will not produce the kind of standardization that financial accounting achieved over decades. The fundamental challenge is that sustainability topics span physical sciences, social dynamics, and governance structures that resist the clean quantification of financial accounting.

Consider Scope 3 emissions. The GHG Protocol's Category 15 (investments) alone requires financial institutions to estimate the emissions of thousands of portfolio companies across multiple asset classes. HSBC's 2025 TCFD report disclosed Scope 3 financed emissions of 65.3 million tonnes CO2e but included a methodological uncertainty range of plus or minus 35%, reflecting the use of estimated emissions data for approximately 60% of its lending portfolio (HSBC, 2025). No regulation can eliminate this uncertainty because it is inherent in the measurement methodology.

What regulation is achieving is a compression of the worst-quality data. The EU's CSRD assurance requirements, which mandate limited assurance on sustainability reports from 2024 and reasonable assurance from 2028, are forcing companies to implement auditable data collection systems. PwC's 2025 survey of 800 CSRD-affected companies found that 67% had upgraded their sustainability data infrastructure in the preceding 12 months, with median investment of EUR 1.2 million in data systems and controls (PwC, 2025). The floor is rising, but the ceiling of data precision will remain well below financial accounting standards for the foreseeable future.

Myth 3: AI and Alternative Data Will Replace Traditional ESG Analysis

The myth: Natural language processing, satellite imagery, web scraping, and other AI-driven data sources will make corporate self-reported ESG data obsolete, providing objective, real-time sustainability measurement.

The reality: Alternative data sources are valuable complements but have significant limitations that are often underreported by their vendors. Satellite-based emissions monitoring, for example, can detect large methane plumes from oil and gas operations but cannot measure Scope 2 or Scope 3 emissions, which constitute 80 to 95% of most companies' total carbon footprint. NLP-based controversy screening can identify negative news coverage within hours but struggles with false positives: a 2025 evaluation by Clarity AI found that 28% of AI-flagged ESG controversies were either misattributed to the wrong entity or involved immaterial events that did not warrant rating action (Clarity AI, 2025).

RepRisk, one of the largest alternative ESG data providers, processes over 500,000 documents daily across 23 languages. The company's own validation research acknowledges that its controversy signals have a 72% precision rate: meaningful but far from the accuracy level that would justify replacing fundamental analysis. Arabesque S-Ray's temperature alignment scores, derived from a combination of disclosed data and AI estimation, showed a 0.62 correlation with independently audited emissions trajectories in a 2025 back-test, indicating useful but imperfect signal.

The most effective implementations combine AI-generated signals with human analyst review. Schroders' sustainable investment team reported that integrating satellite-verified deforestation data with traditional supply chain analysis improved the accuracy of their agricultural commodity ESG assessments by 31% compared to either data source used alone (Schroders, 2025). The evidence supports augmentation, not replacement.

Myth 4: ESG Investing Always Sacrifices Financial Returns

The myth: Incorporating ESG considerations into investment decisions necessarily reduces financial performance because it constrains the investable universe and prioritizes non-financial objectives.

The reality: The evidence on ESG and returns is more mixed than either proponents or critics acknowledge. A 2024 meta-analysis of 1,141 academic studies published since 2015 found that 58% reported a positive relationship between ESG integration and financial performance, 28% found no significant relationship, and 14% found a negative relationship (Friede, Busch, and Bassen, 2024 update). The positive association is strongest for governance factors and weakest for environmental factors, suggesting that "ESG" as a monolithic category obscures important differences.

The timing dimension matters enormously. During the 2020 to 2021 period, ESG-tilted portfolios outperformed largely due to their underweight in fossil fuels and overweight in technology. In 2022 to 2023, the reversal in energy prices and tech valuations produced underperformance for many ESG strategies. Norges Bank Investment Management, which manages Norway's $1.7 trillion Government Pension Fund Global, reported that its climate-adjusted benchmark underperformed its standard benchmark by 0.3 percentage points annually from 2022 to 2024 but outperformed by 0.5 percentage points over the full 2019 to 2024 period (Norges Bank, 2025).

The critical distinction is between ESG integration (using ESG data as an additional input to fundamental analysis) and ESG exclusion (removing sectors or companies from the investable universe). Integration approaches have generally shown neutral to positive risk-adjusted returns. Exclusion approaches impose a tracking error cost that varies with market conditions.

Myth 5: ESG Ratings Providers Have Inherent Conflicts of Interest That Invalidate Their Outputs

The myth: Because major ratings providers also sell consulting services to rated companies, their ratings are compromised in the same way that credit rating agency conflicts contributed to the 2008 financial crisis.

The reality: Conflicts of interest exist and should be managed, but the structural dynamics differ from credit ratings in important ways. Credit rating agencies operate under an issuer-pays model where the rated entity directly pays for the rating. Most ESG ratings providers operate under an investor-pays model where asset managers and asset owners purchase ratings data. MSCI ESG Research, Sustainalytics, and ISS ESG derive the majority of their ratings revenue from investor subscribers rather than rated companies.

However, conflicts do exist in adjacent services. S&P Global's Corporate Sustainability Assessment is directly completed by companies, and participation is voluntary, which creates a self-selection bias. Sustainalytics (owned by Morningstar) offers corporate solutions that help companies improve their ratings. The UK FCA's 2025 ESG ratings regulation directly addresses these concerns by requiring conflict of interest disclosures, separation of ratings from advisory activities, and transparency about methodological changes. Japan's JFSA has adopted similar provisions.

The International Organization of Securities Commissions (IOSCO) published final recommendations on ESG ratings and data products in late 2024, calling for transparency, governance, and conflict management but stopping short of mandating a single structural model. The evidence suggests that conflicts are real but manageable through regulation and transparency, not that they invalidate the entire ratings enterprise.

What's Working

Several developments are genuinely improving ESG data quality and ratings reliability.

Regulatory convergence on disclosure: The ISSB's achievement of baseline global standards, with 25 jurisdictions adopting or committing to adopt IFRS S1 and S2 by early 2026, is creating a common data layer. While differences remain between ISSB and ESRS, the overlap on climate-related financial disclosures is approximately 80%, reducing the reporting burden for multinational companies.

Assurance mandates: The EU's requirement for third-party assurance of sustainability reports is creating accountability. Deloitte, EY, KPMG, and PwC have collectively hired over 8,000 sustainability assurance professionals since 2023, building verification capacity that did not previously exist at scale.

Interoperability initiatives: The EFRAG-ISSB interoperability guidance published in 2025 provides detailed mapping between ESRS and IFRS sustainability standards, allowing companies to report against both frameworks with approximately 30% incremental effort rather than 100% duplication.

What's Not Working

Scope 3 data quality: Despite regulatory mandates, Scope 3 data remains the weakest link. Industry-average emission factors used in spend-based calculations can differ by 300 to 500% from actual supplier emissions, rendering portfolio-level Scope 3 aggregations unreliable for decision-making.

Social metrics standardization: While environmental metrics benefit from physical science measurement (tonnes CO2e, cubic meters of water), social metrics like "living wage coverage" or "community impact" lack comparable precision. The ESRS social standards (S1 through S4) introduce 84 social disclosure data points, but comparability across companies and sectors remains poor.

Small and mid-cap coverage: ESG ratings coverage drops sharply below large-cap equities. Sustainalytics covers approximately 16,000 companies; MSCI covers approximately 8,500. The CSRD will eventually bring 50,000 companies into disclosure requirements, but ratings coverage will lag disclosure by years.

Key Players

Established: MSCI ESG Research: the largest ESG ratings provider by investor subscriber count, covering 8,500 companies with letter-grade ratings. Sustainalytics (Morningstar): risk-focused ratings covering 16,000 companies, increasingly integrated into Morningstar's fund analytics. S&P Global: operator of the Corporate Sustainability Assessment underpinning the Dow Jones Sustainability Indices. Bloomberg: provider of raw ESG data fields integrated into the Bloomberg Terminal used by over 300,000 financial professionals.

Startups/Innovators: Clarity AI: Madrid-based platform using machine learning to process sustainability data for over 70,000 companies, with $80 million in funding from SoftBank and Deutsche Borse. Util: London-based firm mapping corporate revenues to UN Sustainable Development Goals using NLP and economic modeling. Matter: Copenhagen-based startup providing real-time ESG data verification using supply chain transaction data.

Regulators/Standard-Setters: ISSB (IFRS Foundation): setting global baseline sustainability disclosure standards adopted by 25 jurisdictions. EFRAG: developing EU-specific ESRS standards under the CSRD mandate. UK FCA: first major regulator to finalize an ESG ratings oversight regime, effective 2026.

Action Checklist

• Map your ESG data consumption across all ratings and data providers, documenting which provider answers which analytical question and where coverage gaps exist
• Establish a multi-provider framework that uses convergence signals (agreement across two or more providers) as higher-confidence indicators rather than relying on any single rating
• Audit Scope 3 data sources and quantify the percentage based on primary supplier data versus estimated data, with a target of reaching 50%+ primary data within 24 months
• Engage directly with ratings providers to understand methodological changes and provide corrections to factual errors in company assessments
• Build regulatory mapping between ISSB, ESRS, SEC, and jurisdiction-specific requirements to identify overlapping disclosure obligations and reduce duplication
• Implement internal data governance including audit trails, version control, and segregation of duties for sustainability data used in regulatory filings
• Evaluate AI and alternative data sources as supplements to disclosed data, with documented validation of accuracy rates before integration into investment or compliance processes
• Prepare for assurance requirements by establishing controls and documentation comparable to financial reporting processes

FAQ

Q: Should investors use a single ESG ratings provider or multiple providers? A: The evidence strongly supports a multi-provider approach. No single provider has demonstrated superior predictive accuracy across all dimensions of ESG performance. Using two to four providers and focusing on convergence signals (where providers agree) produces more reliable risk identification than relying on any individual rating. The marginal cost of additional data subscriptions (typically $50,000 to $200,000 per year per provider for institutional investors) is small relative to the portfolio risk management value.

Q: How will the UK FCA's ESG ratings regulation change the market? A: The FCA's framework, finalized in 2025 and effective from mid-2026, requires ESG ratings providers operating in the UK to register with the FCA, disclose their methodologies, manage conflicts of interest, and maintain governance standards. The practical impact will be greatest on smaller and newer providers who lack the compliance infrastructure of established players. Expect consolidation in the market, with 10 to 15 smaller providers either exiting or being acquired within 24 months of the regulation taking effect.

Q: Is the divergence between ISSB and ESRS standards a problem for multinational companies? A: The divergence is real but manageable. EFRAG and the ISSB published interoperability guidance in 2025 that maps approximately 80% of climate-related disclosures between the two frameworks. Companies reporting under ESRS will largely satisfy ISSB requirements for climate with limited additional effort. The greater challenge is in social and governance topics where ESRS is significantly more prescriptive than anything in the ISSB framework. Multinational companies should build their data architecture to the more demanding ESRS standard, which will satisfy ISSB requirements as a subset.

Q: What is the timeline for ESG data quality to reach financial-accounting levels of reliability? A: This is unlikely to happen in the foreseeable future, and expecting it reflects a misunderstanding of the measurement challenges. Financial accounting benefits from centuries of methodological development and measures quantities (revenue, assets, liabilities) that are definitionally precise. Many ESG metrics involve estimation, attribution, and boundary-setting choices that introduce irreducible uncertainty. A more realistic expectation is that ESG data quality will reach the reliability level of economic statistics (GDP, inflation): useful for decision-making, directionally accurate, but subject to revisions and methodological debates.

Sources

• Berg, F., Koelbel, J., and Rigobon, R. (2025). Aggregate Confusion: The Divergence of ESG Ratings. Review of Finance, 29(1), 1-30. MIT Sloan School of Management.
• Morningstar. (2025). Global Sustainable Fund Flows: Q4 2025 Report. Chicago, IL: Morningstar Inc.
• CDP. (2025). Scope 3 Emissions Data Quality Assessment: Corporate Disclosure Analysis. London: Carbon Disclosure Project.
• PwC. (2025). CSRD Readiness Survey: European Corporate Implementation Progress. London: PricewaterhouseCoopers LLP.
• HSBC Holdings. (2025). Annual Report and Accounts 2024: Climate-Related Financial Disclosures. London: HSBC Holdings plc.
• Clarity AI. (2025). AI-Driven ESG Controversy Detection: Accuracy and Limitations. Madrid: Clarity AI SL.
• Schroders. (2025). Sustainable Investment Report: Data Integration and Performance Analysis. London: Schroders plc.
• Friede, G., Busch, T., and Bassen, A. (2024). ESG and Financial Performance: Updated Meta-Analysis of 1,141 Studies. Journal of Sustainable Finance and Investment, 14(2), 112-138.
• Norges Bank Investment Management. (2025). Government Pension Fund Global: Annual Report 2024. Oslo: Norges Bank.
• BlackRock. (2025). Investment Stewardship Annual Report 2024. New York: BlackRock Inc.