The carbon market deserves tough, honest, independent scrutiny

The carbon market deserves tough, honest, independent scrutiny. Carbon Market Watch’s Verified but not trusted: Do audits ensure carbon credit quality?  Is not that.

In September 2026 Carbon Market Watch (CMW) published Verified but not trusted: Do audits ensure carbon credit quality?  It examined 15 of the largest projects in the voluntary carbon market. It found that all 15 lacked integrity and that every one of the 94.3 million credits they produced came from projects that “lacked integrity”. The conclusion of the report was that “fundamental questions about the role of the VCM in climate action need to be further scrutinised”.

For our full line-by-line rebuttal of the report [Click here]

Deep investigative scrutiny of how carbon projects are audited is important. Similarly, fully independent and unbiased organisations that use rigorous transparent methodologies should be a feature taken seriously in all markets, the voluntary carbon market included. However, CMW’s result of 15 out of 15, where every case fails, should count as an immediate red flag against the intentions behind the test.

Using the outward appearance of scientific enquiry (science-washing), CMW’s report combines cherrypicked findings, allegations that the authors admit they could not verify and poorly constructed methodologies to reach conclusions its authors and the organisations that they work for held before they started. If buyers act on the conclusion, the result will not be a better market, it will mean less money reaches initiatives that are actively protecting threatened environments.

 

A verdict in search of a method

The report presents a five-part assessment: a review of Verra’s rules for auditors, audit histories, ratings-agency scores, a methodology assessment, and a review of “claims of harm or wrongdoing” against developers. Each part is seemingly built so that it can return a negative result and cannot return a positive one.

• CMW chose the three largest projects by retirements for each of the five largest auditors. Larger projects tend to attract the most journalism and attention. A test that counts accumulated allegations will find most of them in the largest projects, regardless of their quality or accuracy. The approach may have been more valid if CMW had a) included both positive and negative media of projects b) independently tested the validity of the journalism. CMW references negative media only and concedes that it took the negative media at face-value rather than testing its validity.
• The ratings agencies are the only part of the assessment that looks at individual projects with project data. Where they are negative, CMW treats them as evidence. Where two agencies rate a project AA (Katingan Mentaya) or A (Kuamut), CMW concludes it failed anyway “due to shortcomings that were not captured in the ratings from agencies”. The shortcomings it then lists are allegations from 2020 and 2021 campaign and media reports, not findings of its own (these allegations lack integrity, but remain live links and so CMW counts them as evidence).
• CMW scores projects on Carbon Credit Quality Initiative (CCQI) ratings. The CCQI says its scores “do not apply to individual projects” (CCQI). CMW’s own Digital Annex describes this as an assessment of “the strengths or weaknesses of the seven methodologies associated with the 15-project sample”, yet the report counts it toward a verdict on each project (Annex). The same CCQI factsheet scores VM0048, the REDD+ methodology the ICVCM approved for its Core Carbon Principles label in 2024 (ICVCM), at 1 out of 5 on additionality and permanence. Under CMW’s thresholds, a new project using the most up-to-date approved REDD+ methodology would fail. No project using any of the REDD+ methodologies the CCQI has scored could pass this test.
• The “independent expert review” was chosen by the authors. The Annex says only that sources were “systematic studies in academic or grey literature” and gives no rule for which studies count. CMW’s wider sources include a report by the same two authors. CMW also lists a global evaluation of REDD+ projects as a source, but does not report its main finding, that deforestation in project areas was 47% lower than in matched controls (Guizar-Coutiño et al. 2022). It does not cite a 2026 Cambridge-led synthesis, which found that most projects did protect forest, even though too many credits were issued in aggregate (Swinfield et al. 2026).
• Allegations count without being checked. On its developer-conduct test, CMW writes: “We also did not have any realistic way to assess the veracity of the claim.” CMW’s own table note says the claims “can be associated with any project the developer was linked to”. A response to a public consultation by Permian Global is counted in the same category of “claims of harm or wrongdoing” as allegations of evictions, human rights abuse and sexual abuse against other developers.
• Risk is reported as failure. The report describes its test: if a methodology “fails the test for any one of these fundamental elements … then the whole methodology fails.” But what it feeds into that test are risk scores for categories, not findings about projects. A risk that some projects in a category over-credit is not evidence that a given project did.
• There is no published decision rule and no control. The Annex does not say how the five assessments combine into PASS or FAIL. The method was never tried on a case its authors expected to pass. A test that cannot return “pass” is not a measurement.

In June 2025, the same two researchers published Built to Fail?  for Corporate Accountability (report). It looked at the 47 largest projects using similar data and classed four as not problematic. One of the four was the Katingan Mentaya Project. Fifteen months later the project is the same. Its 2021–2023 monitoring period has since been verified, and credits were issued in July 2026 with the approval of Indonesia’s Ministry of Forestry. What has changed is the authors’ method, and with it the verdict.

CMW sets out to show that audits of carbon projects are not objective. Its own method is no more objective than the audits it criticises, and arguably less. Auditors test projects against rules that were published in advance.

 

The bias argument, applied evenly

CMW’s central case against auditors is about incentives. Auditors are paid by the developers they audit, so they “will be inclined (even if unconsciously) to produce results that will please their client”. The report calls this “unconscious, self-serving bias”. It is a fair point, and the market is working on it.

But the point should apply to everyone. CMW is a not-for-profit association with a declared 2025 budget of €2.94 million. Its declared funders include the Quadrature Climate Foundation, the European Climate Foundation, the ClimateWorks Foundation, the European Commission and the EU LIFE programme (EU Transparency Register via LobbyFacts). CMW says some of these funders support its work on voluntary carbon markets specifically. Its published position is that tonne-for-tonne offsetting “does more harm than good” (CMW, 2023). Both authors wrote Built to Fail?  for Corporate Accountability, which also campaigns against offsetting. The lead author, Rachel Rose Jackson, is its Director of Research and International Policy: Climate (Corporate Accountability).

A report finding that the largest projects in the market had passed a fair test would have contradicted the stated positions of both CMW and Corporate Accountability. It is an institutional interest, the same kind of interest CMW identifies in auditors. The report ends: “Neither the authors, experts nor Carbon Market Watch have any conflict of interest to disclose.” By the standard the report applies to others, that is not true.

 

Masquerading as science / science-washing

CMW’s report takes on the outward forms of the scientific system: a methodology, a digital annex, colour-coded tables, footnotes and an “independent expert review”. But it does not accept the disciplines that give those forms their value.

• It does not say it was peer reviewed.
• Its pass/fail rule was not published in advance, or at all.
• It was not tested against a case expected to pass.
• It admits allegations it says it could not verify.
• It did not put its findings to the developers it named before publication.
• It states its conclusion before its evidence. The first chapter is titled “Setting the scene: Carbon credits are not solving the climate crisis”.

It is not wrong to criticise carbon markets; criticism is partly how the market has improved. The problem is that the report presents campaigning in the clothing of inquiry. Readers who see the tables and the annex will assume the conclusion was reached by the method. The method was built to reach the conclusion.

 

Lessons from decades of subversion of climate science

In 2022 Steve Zwick, writing for Ecosystem Marketplace, warned that coverage of climate solutions was “repeating the same mistakes that derailed coverage of climate science itself for decades” (Zwick, Feb 2022). In the third part of his series he borrowed John Cook’s FLICC framework for identifying science denial: fake experts, logical fallacies, impossible expectations, cherry-picking and conspiracy theories (Zwick, Feb 2022). Four of the five fit CMW’s report closely.

Impossible expectations. The report says auditors “should greenlight only project with guaranteed high-integrity projects” [sic], and treats as disqualifying the “legitimate risk” that climate benefits “cannot be absolutely guaranteed”. No counterfactual claim can be guaranteed. It can be estimated, tested, checked against what actually happens and corrected. As Zwick put it: “Science isn’t about absolutes. It’s about a preponderance of the evidence, the willingness to be wrong, and the concurrence of experts.”

Cherry-picking. Negative ratings are used and positive ones set aside. Studies are cited for their findings on over-crediting; their findings that projects reduced deforestation are not reported.

Logical fallacies. Guilt by association: a developer with a contested 2020 baseline debate is listed in the same column as developers accused of abuses.

Auditor findings are treated as project failures: “88 reports and over 1,600 findings” are presented as failures, even though the report concedes that most findings were resolved. It does not say what number of findings would indicate a sound audit.

Conspiracy. In CMW’s account almost everyone is compromised. Standards profit from issuance; auditors please their clients; developers have “motivations and ethics” that call for scrutiny; buyers are greenwashing; ratings agencies have “limits”; the ICVCM is suspect because its chair’s law firm advises oil companies; and policymakers should not rely on third-party verification. The inference is that only CMW, its sources and its interviewees have no conflict of interest and can be trusted.

A further point is that methodologies “have evolved through 45 years of trial, error, and adjustment” (Zwick, Feb 2022). CMW dismisses revisions as “tweaks” that were “minimal in scope and impact”, rather than treating them as the system correcting itself.

 

From critique to delay

In 2020 a group of climate researchers led by William Lamb published “Discourses of climate delay” (Global Sustainability). They described a family of arguments that “accept the existence of climate change, but justify inaction or inadequate efforts”. They grouped them into four strategies: redirect responsibility, push non-transformative solutions, emphasise the downsides, and surrender.

Three of the discourses they identify appear in CMW’s report.

• Policy perfectionism rejects a measure because it is imperfect. CMW asks for guaranteed integrity before any credit is issued, supermajority consent, a community veto where fundamental issues are identified, auditors with a “duty of care” beyond their contracts, and a “radical overhaul” of the system. Each of these may have merit. But required together and in full before anything else can proceed, as CMW demands, they amount to a bar that no conservation mechanism operating in the real world is likely to ever meet.
• Appeal to social justice treats a conservation approach’s possible harms as a reason to abandon it rather than improve it. Real community concerns appear in the report as reasons to doubt the whole market, not as problems to fix. It is worth adding here that CMW does not consider CCB certification in the projects it assesses. For CCB (for which Katingan and Kuamut have Triple Gold), auditors test FPIC, community safeguards, theory of change, benefit sharing and grievance mechanisms.
• Change is impossible, one of the “surrender” discourses, holds that the obstacles cannot be overcome. CMW calls the market’s conflicts of interest “insurmountable” and its design an “irreconcilable flaw”, and says they “cannot be rectified by the limited and targeted ‘fixes'” because “they are inherent in the design”.

The report’s advice to buyers is direct: “Stop greenwashing your image with carbon credits” and “Corporations must stop offsetting their emissions.” Elsewhere, CMW has argued for “contribution claims”, under which companies fund projects without claiming to offset (CMW, 2023). That is a legitimate idea, and some buyers use it. But CMW presents no evidence that the contribution finance will replace, at anything like the same scale, the money its work discourages. Nor does it explain how CMW’s perceived failings of the current approach will be treated under the contribution claims model.

If CMW’s recommendations were adopted in full, the market for forest carbon would be stopped rather than reformed. Carbon finance is one of very few private sources of money that pays to keep standing forest standing. Aid, philanthropy and public budgets cover only a small part of the cost. When that money stops, the pressures on the forest will continue.

We do not believe it is CMW’s intention to cause deforestation or forest degradation, but its method guarantees a conclusion that, if acted on, leads to less forest protection. Currently, the sector is $XX bn per year, smaller than the global chewing gum market, which begs the question, why so much philanthropic money spent on attacking it.

 

What good-faith critique looks like

Permian has worked with critics for more than a decade, and the market is better for scrutiny from academics, journalists, ratings agencies and NGOs. Criticism that identify genuine problems and exposes issues that ultimately help to improve the market tends to share some features. We would welcome a CMW report that met them.

1. Publish the decision rule before applying it e.g. state in advance how the individual tests combine into a final pass or fail verdict, and publish that rule with the results
2. Test the method on a case its authors expect to pass, and report the result.
3. Judge projects on project evidence. Use methodology-level scores to direct attention, not to deliver verdicts.
4. Verify allegations before publishing them, or leave them out.
5. Keep risk and failure separate.
6. Give named organisations the chance to respond before publication.
7. Declare institutional interests, not only financial ones.
8. Say what evidence would change your conclusion.
9. When recommending that buyers withdraw, say where the replacement finance for forests will come from.

Our own documents are public: project descriptions, baseline papers, monitoring reports, legal opinions and audit reports are on the Verra registry. We answer specific technical questions in detail, as we have done for this report. We will keep doing so.

Criticism that improves the carbon market is welcome. A test designed so that nothing can pass does not improve anything. It moves money away from the forests that need it most.