ESG Ratings Disagree on Almost Everything — Here's What Independent Data Shows Instead
Take any company in the FTSE 100. Check its ESG rating from MSCI. Then check Sustainalytics. Then Refinitiv. Then ISS. You will get four different answers. Sometimes radically different.
This is not a fringe observation. It is one of the most well-documented problems in sustainable investing, responsible investing, and socially responsible investing (SRI) — and it affects every portfolio decision that relies on ESG data.
A 2022 study by MIT Sloan and the University of Zurich found that correlations between major ESG rating providers average around 0.54. For context, credit ratings from different agencies correlate above 0.99. When Moody's says a bond is investment grade, S&P almost always agrees. When MSCI says a company is an ESG leader, Sustainalytics may classify it as a laggard.
The question is not just why they disagree. The question is: what happens when you stop relying on any of them and look at what independent sources — court filings, regulatory actions, investigative journalism — actually show?
The Scale of the Disagreement
The divergence is not subtle. Here are documented examples where major ESG providers reached opposing conclusions about the same companies:
Tesla. Routinely rated as an ESG leader by providers emphasising environmental metrics (electric vehicles, clean energy mission). Simultaneously flagged by others for governance concerns (board independence, executive compensation) and labour practices (NLRB complaints, factory safety incidents). Depending on which provider you check, Tesla is either a sustainability champion or a governance risk.
Amazon. Some providers score Amazon favourably on environmental commitments (Climate Pledge, renewable energy investments). Others weight labour conditions more heavily (warehouse injury rates, union suppression allegations, delivery driver classification). Same company, opposite conclusions, depending on methodology.
Meta. Strong governance scores from some providers. Severe negative scores from others on data privacy, content moderation, and societal impact. The Cambridge Analytica scandal moved some ratings and barely registered in others.
ExxonMobil. Some providers rate ExxonMobil positively on governance and disclosure quality. Others weight environmental impact heavily and rate it poorly. The gap between ExxonMobil's self-reported transition investments and its documented regulatory history is precisely the kind of divergence that makes ESG ratings unreliable.
Coca-Cola. High ESG scores from multiple providers on the basis of sustainability reporting, water stewardship programmes, and corporate governance. Simultaneously the subject of documented regulatory actions and investigative reporting on public health impact, plastic pollution, and labour practices in its supply chain.
The pattern is consistent: companies that invest heavily in sustainability reporting tend to receive higher ESG ratings, even when independent evidence raises questions about aspects of their actual conduct.
Why Do They Disagree?
Three structural reasons:
1. Different Definitions
There is no standard definition of what "ESG" means. Each provider defines its own categories, selects its own indicators, and decides what matters. MSCI uses 35 key issues across 10 themes. Sustainalytics uses a different framework entirely. They are not measuring the same thing — they are each measuring their own version of "sustainability."
This is like asking five doctors to rate your health, but each doctor defines health differently. One measures cardiovascular fitness. Another measures mental health. A third measures bone density. They all call it a "health score." They all disagree.
2. Same Source, Different Weight
Even when providers look at similar data, they weight it differently. One provider might weight carbon emissions at 30% of the total score. Another might weight it at 5%. A company with high emissions but strong governance could score 80/100 with one provider and 40/100 with another — using the same underlying facts.
3. Corporate Disclosure as the Primary Input
This is the fundamental problem. The majority of ESG data comes from companies themselves: sustainability reports, CDP questionnaires, GRI disclosures, TCFD frameworks. Companies choose what to report. They choose what to emphasise. They choose what to omit.
When your entire rating system is built on voluntary self-disclosure, you are not measuring corporate behaviour. You are measuring corporate communication.
What Independent Data Shows Instead
Mashinii scores companies using sources that companies do not control: court records, regulatory enforcement actions, investigative journalism from major outlets, and NGO field research. Every score links to the specific source that generated it.
Here is what this approach reveals about some of the companies where ESG providers disagree most. The scores below reflect Mashinii's methodology — based on independently sourced evidence, not corporate self-reports:
Tesla (TSLA)
| Dimension | Score |
|---|---|
| Respect for Cultures & Communities | +25 |
| Safe & Smart Tech | 0 |
| Fair Pay & Worker Respect | -40 |
| Honest & Fair Business | -50 |
Tesla scored -50 on Honest & Fair Business — reflecting $99 million in documented regulatory penalties, SEC fines for investor protection violations, and a whistleblower complaint alleging improper warranty cost categorisation. It scored -40 on Fair Pay & Worker Respect, driven by NLRB rulings on union activity and OSHA citations. Yet it scored +25 on Respect for Cultures & Communities. A single ESG score cannot hold these contradictions. Dimension-level scoring can.
View Tesla's full score breakdown
Amazon (AMZN)
| Dimension | Score |
|---|---|
| Better Health for All | +30 |
| Safe & Smart Tech | +10 |
| Fair Pay & Worker Respect | -30 |
| Fair Trade & Ethical Sourcing | -40 |
| No War, No Weapons | -40 |
| Honest & Fair Business | -50 |
Amazon scored -50 on Honest & Fair Business — reflecting a $2.5 billion settlement over alleged deceptive Prime subscription practices and a $5.9 million fine for California labour law violations. It scored -40 on both Fair Trade & Ethical Sourcing and No War, No Weapons. Yet it scored +30 on Better Health for All and +10 on Safe & Smart Tech. Pledges do not appear in court filings or regulatory actions — because pledges are not events that independent bodies track. Actions are.
View Amazon's full score breakdown
ExxonMobil (XOM)
| Dimension | Score |
|---|---|
| Kind to Animals | 0 |
| No War, No Weapons | -10 |
| Planet-Friendly Business | -30 |
| Zero Waste & Sustainable Products | -30 |
| Honest & Fair Business | -40 |
| Fair Pay & Worker Respect | -40 |
| Better Health for All | -40 |
ExxonMobil scored negatively on 9 out of 11 dimensions — with -40 on both Fair Pay & Worker Respect and Better Health for All, and -30 on Planet-Friendly Business. The independent record includes decades of documented environmental regulatory actions, occupational safety citations, and community health litigation. ESG providers that credit ExxonMobil's transition disclosures are scoring its communications. Our methodology scores its documented conduct.
View ExxonMobil's full score breakdown
Coca-Cola (KO)
| Dimension | Score |
|---|---|
| Safe & Smart Tech | 0 |
| Fair Money & Economic Opportunity | 0 |
| Honest & Fair Business | -10 |
| Fair Pay & Worker Respect | -30 |
| No War, No Weapons | -30 |
| Planet-Friendly Business | -50 |
| Better Health for All | -60 |
Coca-Cola scored -60 on Better Health for All — the lowest score in this comparison — and -50 on Planet-Friendly Business, despite extensive sustainability reporting and water stewardship programmes. The independent record includes documented regulatory actions on public health impact, plastic pollution litigation, and supply chain labour concerns. Coca-Cola routinely scores well on ESG ratings because it reports well. Independent sources tell a different story.
View Coca-Cola's full score breakdown
The Gap Between Self-Reports and Independent Evidence
The pattern across these examples is consistent:
-
Companies that invest in sustainability reporting score well on ESG ratings. Reporting quality is rewarded, regardless of whether the report accurately reflects company behaviour.
-
Independent sources capture events that reports omit. Court filings exist because someone was harmed. Regulatory fines exist because a rule was broken. Investigations exist because something was wrong. These are not opinions — they are documented facts.
-
Dimension-level scoring reveals what aggregation hides. A company can be genuinely excellent on environmental innovation and genuinely terrible on labour rights. A single score averages these away. Showing both lets investors make informed trade-offs.
This is the core of the ESG rating vs actual conduct gap — and it is not a gap that better ESG methodologies can close, because the problem is the data source, not the methodology.
What This Means for Investment Decisions
If you are an investor relying on ESG ratings to align your portfolio with your values, the disagreement problem is not academic. It directly affects your holdings.
For individual investors: The fund labelled "sustainable" was screened using one provider's methodology. A different provider might have excluded companies your fund holds, or included companies your fund excluded. The label reflects a methodology choice, not an objective standard. Check what is actually inside your funds.
For financial advisors: When a client asks "is my portfolio ethical?" and you check an ESG data provider, the answer depends on which provider you check. Independent verification gives you an answer you can stand behind — one backed by cited evidence rather than proprietary methodology. Read the advisor's guide to independent ESG verification.
For compliance officers: The FCA's Anti-Greenwashing Rule requires that sustainability claims be "fair, clear and not misleading." Relying on a single ESG rating provider's assessment — when that provider disagrees with others — is a risk. Independent verification from adversarial sources provides a defensible evidence base.
Stop Comparing Ratings. Start Checking Evidence.
Mashinii does not produce another ESG rating. We score companies across 11 dimensions using court filings, regulatory actions, investigative journalism, and NGO reports. Every score is cited. Every citation is linked. The data comes from sources companies do not write and cannot edit.