Why Traditional ESG Ratings Often Miss the Full Picture
Trillions of dollars flow through funds labelled "sustainable" on the basis of ESG ratings that most investors never interrogate. A single letter grade decides whether a company enters a responsible investment portfolio or gets screened out. The assumption is that this grade reflects reality.
It often does not.
The agencies producing these ratings face structural problems that cannot be patched with better algorithms. The data sources are controlled by the companies being rated. The scoring compresses contradictory signals into a single number. And the business model creates financial ties between rater and rated. None of this is secret. Most of it is simply inconvenient to talk about.
This is not a case against ESG as a concept. It is a case for better data. Investors who rely exclusively on traditional ESG ratings are making decisions on an incomplete picture -- and the gaps are predictable.
Companies Grade Their Own Homework
Traditional ESG ratings draw primarily from corporate sustainability reports, ESG questionnaires, policy documents, and voluntary disclosures. MSCI, Sustainalytics, S&P Global, and other major agencies collect this self-reported data, apply proprietary weightings, and produce a score.
The entity being evaluated controls the evidence used in the evaluation.
The distortions are predictable. A company with a 50-person sustainability communications team and a 200-page annual report will outscore a company that publishes nothing -- even if the silent company has a cleaner actual record. Disclosure is rewarded. Conduct is secondary.
This is not a fringe critique. A widely cited 2022 study from MIT Sloan found that ESG ratings from different agencies correlate at only about 0.54. Credit ratings from different agencies correlate at roughly 0.99. When two agencies look at the same company and routinely disagree, the inputs deserve scrutiny, not just the weightings.
One Number Cannot Capture 11 Dimensions of Conduct
The most consequential flaw in traditional ESG is the compression of complex corporate behaviour into a single score. A company gets one grade. That grade determines inclusion or exclusion from ESG funds. The trade-offs underneath it vanish.
Tesla: Strong on Waste, Weak on Governance
Tesla is the clearest illustration. Its core product gives it a structural advantage on environmental metrics. On our Zero Waste & Sustainable Products dimension, Tesla scores +40, reflecting independently verifiable achievements including 90% manufacturing waste recycled in 2023, with its Shanghai facility reaching 94%.
But Tesla also scores -50 on Honest & Fair Business and -50 on Fair Pay & Worker Respect. Public records show over $99 million in regulatory penalties spanning financial offences, government contracting violations, and competition issues across multiple jurisdictions. The company's labour record includes 77 incidents since 2008, with 80% being OSHA citations.
Tesla's average across all 11 dimensions is -10.5. An investor who cares about environmental manufacturing would see the +40 and feel reassured. An investor who cares about worker safety would see -50 and reconsider. A blended ESG grade gives neither investor the information they actually need.
S&P removed Tesla from its ESG index in 2022. Many other agencies continue to rate it moderately, because its environmental product story offsets its governance record in composite scoring.
View Tesla's full score breakdown ->
The Public Record Tells a Different Story
Corporate sustainability reports are marketing documents. Produced by the company. Approved by the company. Designed to present the company at its best. This does not mean they contain lies. It means they contain a curated selection of facts, framed favourably.
Independent sources operate under different incentives. Court filings are adversarial -- plaintiffs' lawyers are not trying to flatter the defendant. Regulatory penalties are imposed, not volunteered. Investigative journalists serve their readers, not the companies they cover.
When companies are scored against this independent record -- as Mashinii does across 11 ethical dimensions -- the picture shifts.
Amazon publishes extensive sustainability data: its Climate Pledge, 100% renewable energy matching, 500+ wind and solar projects. These are real. Amazon scores +30 on both Planet-Friendly Business and Better Health for All in our analysis.
The same company scores -40 on Fair Trade & Ethical Sourcing, reflecting a documented $5.9 million penalty in California for violating warehouse quota laws. It scores -30 on Safe & Smart Tech, reflecting a 32 million euro GDPR fine in France for employee surveillance. And -30 on No War, No Weapons, reflecting defence contracts in public procurement records.
Amazon's average: -10.0. A data provider drawing primarily from Amazon's own disclosures will see the environmental strengths clearly and the labour and privacy weaknesses dimly, if at all.
View Amazon's full score breakdown ->
The Business Model Problem
There is a structural issue the ESG industry rarely confronts: most rating agencies earn revenue from the companies they rate.
Not always through direct payments for ratings, though some agencies do charge for advisory services. Companies pay for access to ESG frameworks, benchmarking tools, and consulting on how to improve their scores. The agency rating a company's performance may also advise that company on scoring better next time.
This mirrors the dynamic that plagued credit rating agencies before 2008. The issuer-pays model creates incentives that do not always align with investor interests. An agency that consistently rates its paying clients poorly will eventually have fewer paying clients.
Independent public data has no such incentive structure. Court records exist whether or not anyone pays for them. Regulatory penalties are public. The independence of the source is what gives it integrity.
What Multidimensional Scoring Reveals
When you replace a single grade with separate scores across distinct ethical categories, patterns emerge that blended ratings suppress.
Microsoft: AAA Rating, -60 on Defence Exposure
MSCI assigns Microsoft AAA -- the highest possible ESG grade. The company's carbon-negative pledge, governance framework, and sustainability reporting justify that in disclosure-based analysis.
Mashinii scores Microsoft at +5.5 overall. Positive, but far from the perfection AAA implies. The breakdown shows why:
| Dimension | Score |
|---|---|
| Respect for Cultures & Communities | +50 |
| Better Health for All | +30 |
| No War, No Weapons | -60 |
Microsoft's community engagement and health contributions are genuine. But the company scored -60 on No War, No Weapons -- among the most negative scores on any dimension for any company in our database -- reflecting significant defence contract relationships documented in public procurement records.
An investor screening for community impact would find Microsoft excellent. An investor screening for military exposure would find it among the most problematic in the S&P 500. Both positions are supported by verifiable evidence. A single AAA communicates neither.
View Microsoft's full score breakdown ->
Three Patterns That Traditional Ratings Obscure
Scoring companies against independent sources rather than self-reported disclosures reveals consistent structural patterns.
Environmental Performance Does Not Predict Ethical Performance
Google scores +40 on both Planet-Friendly Business and Zero Waste & Sustainable Products. Its environmental record, verified through independent data, is genuinely strong.
Google also scores -40 on Honest & Fair Business, driven by antitrust proceedings across multiple jurisdictions. It scores -30 on Safe & Smart Tech, reflecting the scale of its data collection practices as documented in regulatory findings. Overall average: +2.3 -- barely positive, despite environmental scores that would rank it among top performers in any conventional framework.
View Google's full score breakdown ->
Data Privacy Is Systematically Underweighted
Traditional ESG treats data privacy as a governance footnote. The independent record suggests it deserves standalone treatment.
Meta scores -50 on Safe & Smart Tech and -40 on Honest & Fair Business. Multiple jurisdictions have imposed fines related to data handling. The company's overall average is -14.5. Yet Meta has appeared in numerous ESG-labelled funds, partly because its renewable energy investments score well in disclosure-based assessments.
When a company scores -50 on data privacy and still enters ESG portfolios, the methodology is weighting something other than what most investors would consider central to a technology platform's ethical profile.
View Meta's full score breakdown ->
The Worst Scores Belong to the Most Widely Held Stocks
In our analysis of 40 S&P 500 companies, 72.5% scored negative on average ethical integrity. The mean was -6.2. The companies that scored worst -- Exxon Mobil (-25.0), Boeing (-20.9), Procter & Gamble (-19.1) -- sit in retirement accounts and index funds held by millions of people who have never examined their ethical exposure.
Exxon Mobil scored negative on every single dimension we measure. Its best score on any dimension is 0, on Kind to Animals, where insufficient independent evidence exists to assess either way. On every other dimension, the independent record is negative.
View Exxon Mobil's full score breakdown ->
Five Questions to Ask Your ESG Data Provider
If you use ESG ratings to make investment decisions, the quality of those ratings determines the quality of those decisions.
1. Where does the underlying data come from? If the answer is primarily corporate disclosures and ESG questionnaires, the data reflects what companies choose to say about themselves.
2. Can a company improve its score by publishing a better report? If yes, the rating measures disclosure quality, not conduct.
3. Is the score multidimensional? A company that performs well on environment and poorly on governance is a fundamentally different holding from one that performs well on both. A single grade destroys that distinction.
4. Can you verify the evidence behind each score? If the methodology is proprietary and the sources are undisclosed, you are trusting the agency's judgement without the ability to check it.
5. Does the agency have financial relationships with the companies it rates? This does not automatically invalidate the output. But it is a conflict investors should know about.
How Mashinii Differs
Mashinii was built on the premise that the independence of the data source matters more than the sophistication of the model.
Adversarial sources, not corporate disclosures. Scores are derived from court filings, regulatory actions, investigative reporting, and NGO assessments. Companies cannot influence their scores through voluntary disclosure.
11 separate dimensions. Every company is scored on each of 11 ethical dimensions independently. A +40 on environment and a -50 on governance both show. The trade-off is visible.
Cited evidence. Every score is backed by referenced sources that users can verify. We show the evidence and let you draw your own conclusions.
No issuer-pays model. Companies do not pay for ratings, advisory services, or improvement consulting. Revenue comes from investors who use the data.
Continuous updates. Scores reflect the latest public record, not annual review cycles.
See how the scoring methodology works ->
The Bottom Line
Traditional ESG ratings are not worthless. They capture useful information about corporate sustainability commitments and governance structures. For investors who want a rough directional signal, they serve a purpose.
But for investors who want to understand what companies actually do -- not just what they report -- traditional ratings fall short. The data sources are too controlled, the scoring too compressed, and the financial incentives too entangled.
The independent public record exists and it is not going anywhere. Court filings, regulatory penalties, and investigative reports provide a parallel dataset that is adversarial, verifiable, and free from corporate influence. Using it does not always produce a worse picture. But it nearly always produces a more complete one.
See What Independent Data Says About Your Holdings
If your portfolio includes ESG-labelled funds or individually selected stocks, the underlying holdings may not match what the independent record reveals. Mashinii scores every company across all 11 dimensions -- not a single blended grade, but 11 separate assessments backed by cited evidence.