The Divergence Record
The Berg, Kolbel, and Rigobon (2022) study, published in the Review of Finance, Volume 26, is the definitive quantification of ESG rating divergence. The researchers analyzed data from six prominent ESG rating agencies: KLD (now part of MSCI), Sustainalytics, Moody's ESG (formerly Vigeo-Eiris), S&P Global (formerly RobecoSAM), Refinitiv (formerly Asset4), and MSCI. The correlations between overall ESG ratings ranged from 0.38 to 0.71, with an average of 0.54. The highest agreement was between Sustainalytics and Moody's ESG at 0.71. The lowest was between KLD and Refinitiv at 0.38.
The dimensional breakdown is revealing. The environmental dimension has the highest average correlation at 0.53 — still far below what would constitute agreement. The social dimension averages 0.42. The governance dimension, which deals with board composition, executive compensation, audit procedures, and shareholder rights — factors that should be among the most objectively measurable — has the lowest average correlation at 0.30. Two raters evaluating the same company's governance can produce ratings that are essentially uncorrelated.
The study decomposed the sources of divergence into three categories. Measurement divergence — raters measuring the same attribute differently — contributed 56% of the total disagreement. Scope divergence — raters including different attributes in their assessment — contributed 38%. Weight divergence — raters assigning different importance to the same attributes — contributed just 6%. The dominant source of disagreement is not that raters value different things, but that they measure the same things and arrive at different conclusions.
Perhaps most consequentially, the researchers identified a "rater effect": a rater's overall impression of a firm influences its measurement of specific ESG categories. If a rater has a positive overall view of a company, that positive view bleeds into specific category ratings. This is the halo effect applied to environmental accountability — the structural analog of a credit rating agency whose overall impression of a borrower's creditworthiness influences its assessment of specific financial metrics. In credit rating, this pattern contributed to the 2008 financial crisis. In ESG rating, it produces a systematic upward bias for companies with strong overall reputations, regardless of specific environmental or social performance.
The Industry Structure
The ESG ratings industry reached an estimated $50 billion in market value by the mid-2020s, encompassing ratings, data provision, index construction, and advisory services. The major providers — MSCI, Sustainalytics (owned by Morningstar), S&P Global, ISS ESG, and Refinitiv (owned by the London Stock Exchange Group) — were consolidated through acquisitions that transformed what began as social responsibility research into financial market infrastructure.
The business model creates a structural conflict of interest that differs from but parallels the credit rating agency model documented in the 2008 financial crisis. In credit rating, the issuer pays for the rating — the entity whose debt is being rated is the customer. In ESG rating, the relationship is more complex: some revenue comes from investors who subscribe to ESG data, but a growing share comes from the rated companies themselves, who pay for detailed assessments, advisory services, and the data that supports their ESG disclosures. The rated company is simultaneously the subject of assessment and a revenue source for the assessor.
The competitive dynamics reinforce divergence rather than convergence. Each rater differentiates its methodology as a proprietary competitive advantage. MSCI emphasizes financially material ESG risks. Sustainalytics emphasizes ESG risk exposure. S&P Global emphasizes corporate sustainability assessment. If the methodologies converged — if they measured the same things in the same way and produced the same ratings — the products would become interchangeable commodities, destroying the premium each rater charges for its proprietary approach. Divergence is a feature, not a bug, of a competitive market for ratings whose value derives from their differentiation.
Credit ratings converged to 0.99 correlation because the market penalized disagreement — a bond that defaulted after receiving a high rating destroyed the rater's credibility. ESG ratings diverge at 0.54 because the market has no equivalent penalty — a company with a high ESG rating that pollutes faces no mechanism that feeds back to the rater's credibility.
High ESG Scores, Poor Environmental Records
The divergence between ESG ratings and environmental outcomes is not theoretical. Documented cases reveal a systematic pattern in which companies receive high ESG scores from one or more raters while maintaining environmental records that contradict those scores. The pattern is structural rather than anecdotal: the ESG rating measures what the company reports about its environmental management, not what the environment experiences as a result of the company's operations.
The fossil fuel industry provides the most striking examples. Major oil and gas companies — including some of the largest emitters in history — have received above-average ESG ratings from certain agencies based on their environmental management systems, disclosure practices, and transition plans. The rating reflects the quality of the company's ESG reporting infrastructure, not the quantity of its emissions. A company that emits 100 million tonnes of CO2 annually but has excellent disclosure, ambitious targets, and a well-staffed sustainability department can score higher than a company that emits 10 million tonnes with less sophisticated reporting.
Fast fashion brands have received favorable ESG scores for supply chain transparency initiatives while their production models depend on high-volume, low-durability manufacturing that generates approximately 10% of global carbon emissions and is the second-largest consumer of water worldwide. Technology companies have scored well on governance while maintaining data collection practices that systematically undermine user privacy. Financial institutions have received high ESG marks for their ESG investment products while simultaneously financing the largest fossil fuel expansion projects. In each case, the ESG rating measures the company's ESG management infrastructure rather than its ESG impact — the input rather than the outcome.
This is not a failure of measurement precision. It is a structural feature of a system that rates what companies do about ESG rather than what effect companies have on ESG outcomes. The distinction is fundamental: a pharmaceutical company's ESG rating reflects its governance structure for managing drug safety, not whether its drugs are safe. An energy company's ESG rating reflects its climate risk disclosure, not whether its emissions are declining fast enough to meet the Paris Agreement targets. The Ratings Architecture converts environmental accountability from an outcome metric (atmospheric concentration, ecosystem health, community impact) to a process metric (policy existence, disclosure quality, management commitment). The process metric is within the company's control. The outcome metric is not.
The Anti-ESG Backlash as Political Theater
Beginning in 2022, a political backlash against ESG investing emerged primarily in the United States, driven by Republican state officials who characterized ESG as a political agenda that prioritized "woke" social objectives over fiduciary duty to investors. Texas, Florida, and more than a dozen other states enacted or proposed legislation restricting state pension funds from using ESG criteria in investment decisions. The SEC's Climate and ESG Task Force, established in 2021, was eliminated in 2025. BlackRock CEO Larry Fink stopped using the term "ESG" in public communications.
The anti-ESG movement correctly identified that ESG ratings are not delivering the environmental accountability they promise. It incorrectly diagnosed the cause. The political critique frames ESG as an overreach of environmental regulation into financial markets — the imposition of social goals on fiduciary processes. The structural critique, documented in this paper, is precisely the opposite: ESG ratings fail because they are not accountable enough, because they measure process rather than outcome, because the rated companies fund the raters, and because the ratings diverge so fundamentally that they cannot function as reliable signals of environmental or social performance.
The political backlash serves a specific structural function: it converts the legitimate critique of ESG's captured accountability mechanism into a partisan issue, preventing the structural reforms that would make ESG ratings actually function as accountability tools. If the anti-ESG movement succeeds in eliminating ESG requirements, the industries with the worst environmental records lose the weak accountability mechanism that ESG represented — without gaining any stronger one. If the pro-ESG movement succeeds in defending ESG requirements without reforming the underlying divergence and capture, the accountability mechanism remains captured. Both outcomes serve the same structural interest: the industries whose environmental impact ESG was designed to measure.
The anti-ESG backlash is political theater that obscures the structural critique by converting it into partisan identity. The structural critique is not that ESG is too aggressive in its environmental demands. It is that ESG is too captured to be aggressive at all. Rating divergence at 0.54, process-over-outcome measurement, issuer-funded assessment, and the absence of feedback mechanisms when high-rated companies produce poor environmental outcomes — these are the features of a captured accountability mechanism, not an overreaching one.
The Ratings Architecture — Named
The structural condition in which an environmental and social accountability mechanism has been captured by the industries it was designed to evaluate, producing ratings that measure process rather than outcome, that diverge fundamentally across raters (0.54 average correlation), that are influenced by halo effects from a rater's overall impression of the rated company, and that face no feedback mechanism when high ratings correspond to poor environmental or social outcomes. The Ratings Architecture is the ESG-specific instance of a broader pattern documented across the Market saga: accountability mechanisms that are funded by the entities they evaluate, staffed by people who move between the evaluating and evaluated institutions, governed by methodologies whose proprietary differentiation is a competitive advantage for the evaluator, and politically defended against structural reform by a partisan backlash that converts the legitimate critique of capture into a culture war issue. The Ratings Architecture does not fail by accident. It succeeds at what it is structurally designed to do: produce the appearance of environmental accountability at a level of rigor that does not threaten the business model of the entities being rated. The 0.54 correlation is not noise in a system that will converge with maturity. It is the equilibrium output of a system whose revenue model depends on the ratings remaining different enough to justify the existence of multiple providers, but similar enough to maintain the collective legitimacy of the rating enterprise.
The ESG ratings industry's structural capture follows the same pattern documented in the Compliance Theater series (Saga VI): the inspection surface is constructed by the regulated entity, the evaluation methodology measures what the entity agrees to make visible, and the accountability mechanism becomes a legitimation tool rather than a constraint. The difference is scale and financial consequence. ESG ratings influence trillions of dollars in investment allocation. Index funds that track ESG benchmarks — including MSCI ESG Leaders, FTSE4Good, and S&P 500 ESG — automatically allocate capital based on ratings whose divergence means the same company may be included in one ESG index and excluded from another.
The reform path is structurally clear and politically blocked. ESG ratings would become meaningful accountability tools if they measured outcomes rather than processes, if raters were funded by investors rather than by rated companies, if methodologies were standardized enough to produce convergence, and if there were consequences for raters when high ratings corresponded to poor outcomes. Each of these reforms would reduce the ESG ratings industry's revenue, reduce rated companies' control over their scores, and reduce the political utility of the anti-ESG backlash. The Ratings Architecture persists because every stakeholder in the current system — raters, rated companies, asset managers, and the political movements that oppose ESG — benefits from its captured form.
The Offset Architecture (CL-004) documents the same structural capture in the carbon offset market — accountability mechanisms that validate phantom reductions because the buyer and the verifier share an interest in the certificate's existence rather than in the reduction's reality. The Stranded Asset Problem (CL-005) traces the financial foundation that makes the entire Climate Architecture necessary: the $100 trillion in fossil fuel assets whose valuation depends on the accountability mechanisms remaining captured.