Why the VCM's credibility crisis
is an opportunity -
if registries hold the line

Let me begin with something that is uncomfortable to say from inside the industry: the journalists, researchers, and regulators who spent the last two years scrutinising the voluntary carbon market were right to do so.

The Guardian, Zeit, and SourceMaterial investigations into Verra's REDD+ portfolio - which found that the vast majority of credits in their study represented no real emissions reductions - were not a hit job. They were a fair reading of a structural failure in how additionality and baseline scenarios were being set. The Science paper by West et al. that followed made the same case with greater rigour.

The South Pole scandal, where one of the largest project developers sold credits from a project that was not sequestering the volumes it claimed, was not an isolated incident of fraud. It was a symptom of a verification regime that had been captured by the economic interests it was meant to police.

I say this not to be self-flagellating about a market I have spent my career working to improve, but because the first step towards rebuilding credibility is an accurate diagnosis of the problem. And the diagnosis is this: the VCM grew faster than its integrity infrastructure could keep up with. It prioritised volume and accessibility over rigour. And it created perverse incentives at every level of the supply chain.

The market contracted sharply as a result. Voluntary carbon credit prices fell by more than 50% in 2023 for nature-based credits. Corporate buyers paused programmes. Several high-profile net-zero commitments were quietly revised. This was painful. It was also, in the long run, probably necessary.

It is worth being specific about the failure modes, because the instinct in industry responses has been to generalise - to say that "confidence was dented" or that "communication was poor." This understates what actually went wrong.

Baseline inflation. The most pervasive problem was the systematic overstatement of what would have happened in the absence of the project. Avoided deforestation methodologies allowed projects to set baselines against theoretical worst-case scenarios that never materialised - not because of the project, but because they were never realistic to begin with. The forest was never going to be cleared at the rate the model assumed. The credits were largely fictional.

Verification without independence. The third-party verification system, as designed, contained a fundamental conflict of interest: verifiers were hired and paid by the project developers whose claims they were assessing. The economic logic created pressure to approve. Verifiers who were too stringent lost clients. This is not a new observation - it was visible in the structure of the system from the beginning - but the market chose to ignore it because the costs of fixing it fell on the industry and the benefits accrued to credibility, which is a public good.

Methodology lag. The science of carbon removal moved faster than the methodologies governing it. Soil carbon accounting improved significantly between 2018 and 2022, but methodologies approved before those improvements were still being used to issue credits against older, less accurate baselines. There was no systematic mechanism to require retrospective recalculation.

Co-benefit washing. Biodiversity and community co-benefit claims were frequently aspirational rather than verified. Projects that had conducted superficial community consultations claimed FPIC compliance. Projects that had planted monocultures of fast-growing exotics claimed biodiversity co-benefits. The secondary claims were not subject to the same rigour as the primary carbon accounting, even though they were central to how credits were marketed.

These failures compound. A project with an inflated baseline, verified by a non-independent auditor, using an outdated methodology, and claiming unverified co-benefits is not simply a project with four separate problems. It is a project where none of the normal error-correction mechanisms functioned. The integrity of the system depends on each layer catching what the previous layer missed. In these cases, none did.

Underlying all the specific failures is a structural one: the VCM is not a regulated market. It is a system of private governance, in which standard-setting bodies, registries, verifiers, and project developers operate under frameworks that they have largely designed themselves, with limited external accountability.

This is not inherently fatal. Private governance can work well when the reputational stakes for participants are sufficiently high, when the information environment is transparent enough for reputational damage to follow bad behaviour, and when there is no gap between who bears the cost of failure and who captured the benefits of the practice that failed.

In the VCM, none of these conditions fully held. Standard bodies and registries had strong reputational incentives, but their reputations were not sufficiently granular: a scandal in one methodology category damaged the reputation of the whole registry, creating incentives to manage perception rather than substance. Transparency was limited: the baseline assumptions, monitoring data, and verification reports that would have allowed external scrutiny were often not publicly available in a useable form. And the people who bore the cost of credibility failures - communities whose forests were supposedly being protected, future generations, the atmosphere - were not the people who captured the economic upside of the credits.

The incentive structure of registries themselves deserves particular attention. A registry that charges per-tonne fees on credits it approves has a direct financial interest in approving more credits. This is obvious in retrospect. It may not have been designed with this intention, but the structural incentive is there regardless of intent - and structural incentives tend to shape behaviour over time more reliably than stated values.

Here is the alternative reading of the past two years: it produced exactly the kind of information environment that allows private governance to self-correct. The investigations made the failures visible. They created reputational costs for specific practices. They shifted buyer behaviour. And they created political space for more ambitious reform.

The question is whether the institutions that fill the space left by discredited practices are genuinely better, or whether they are performing improved credibility while maintaining the structural incentives that produced the original failures. This is not a cynical question. It is a design question.

There are at least three genuine opportunities in the current moment:

None of these opportunities are automatic. They require registries to make specific design choices that impose short-term costs - on methodology development timelines, on the range of projects that can qualify, on the attractiveness of the platform to developers seeking the path of least resistance. The opportunity is real, but it has to be deliberately taken.

"Holding the line" is a phrase that sounds passive - a matter of not doing bad things. In practice, it requires active institutional choices that create friction and impose costs. Let me be concrete about what these look like.

Saying no to projects that are on the edge. Every registry operates a grey zone: projects where the methodology technically applies but where the additionality case is weak, the monitoring is thin, or the community consent process was superficial. The correct response to these is not to find a way to make them work. It is to decline them. This is commercially painful in the short term - it means turning away revenue - but it is the only way to maintain the integrity of the approved pool.

Publishing rejection decisions. One of the least visible but most important integrity mechanisms is the publication of declined applications and the reasons for them. This communicates the actual standard being applied, creates accountability for decisions, and provides learning for the developer community about what the registry requires. Most registries do not do this. They should.

Separating verification from commercial relationships. The conflict-of-interest problem in third-party verification is solvable. It requires either that verifiers are selected and paid by the registry rather than the developer, or that verifiers are subject to meaningful rotation and cannot derive more than a defined proportion of their revenue from any single developer. Both mechanisms exist in other regulated contexts - auditor independence in financial reporting, for example - and there is no principled reason they cannot apply here.

Treating methodology versioning as an integrity mechanism. When the science improves, existing credits should not be automatically grandfathered against old methodologies in perpetuity. A mechanism for periodic recalibration - with appropriate notice periods and transition arrangements, but with real teeth - is necessary for long-run credibility. Credits issued under a methodology later found to be significantly overestimating removal should be annotated accordingly in the registry record.

None of this is revolutionary. These are standard elements of institutional design in other high-stakes verification contexts. The VCM has been slow to adopt them because they impose costs on a market that grew up in an era of low scrutiny. The era of low scrutiny is over.

I should be transparent about the institutional perspective I am writing from. Teravent is a new registry, building its reputation and methodology stack from scratch, in an environment where buyer scrutiny is higher than it has ever been. We have obvious incentives to position ourselves as the high-integrity alternative to incumbents who have been damaged by the investigations of the past two years. You should read this article with that in mind.

What I would ask in return is that you hold us to the commitments we make publicly. We have designed our fee structure to avoid per-tonne approval incentives - our revenue is subscription-based, not volume-dependent. We have committed to publishing declined application decisions in anonymised form. Our methodologies are peer-reviewed, versioned, and subject to public comment periods before approval. Our verification body selection process separates commercial relationships from technical assessment.

These are claims. The test of whether they are more than claims is whether they hold under commercial pressure - when a large developer submits a project on the edge, when a methodology review suggests we have issued some credits at the top of their uncertainty range, when a verifier relationship becomes commercially convenient to maintain beyond what the rotation policy permits.

I cannot give you certainty about how we will behave in situations we have not yet faced. What I can tell you is the structural logic we have tried to build in: that the costs of approving something we should not approve are institutionally larger than the costs of declining it. Whether that logic holds will be visible in our public record over the next few years. We will be publishing that record in full.

The VCM's credibility crisis is real, and the damage to the projects and communities who depend on carbon finance is also real. The contraction in credit prices and buyer participation has cut funding to forest conservation, soil restoration, and blue carbon work that would not otherwise be funded. This is a genuine cost.

But the crisis is also information. It has told the market - more clearly than any previous communication - what the consequences of inadequate integrity infrastructure are. And it has, belatedly, created the conditions in which registries that invest in genuine rigour can be rewarded for it commercially rather than just morally.

The opportunity is narrow. It requires the registries, buyers, and standard-setters who want a high-integrity voluntary carbon market to hold their positions even as commercial pressure pushes in the other direction. It requires buyers to continue doing due diligence rather than reverting to price-based purchasing when the scrutiny cycle moves on. It requires NGOs and researchers to continue publishing critical analysis even when the subject of that analysis is a project or registry they broadly support.

The voluntary carbon market does not need to be perfect to be useful. It needs to be honest about what it can verify, rigorous in the standards it applies to what it claims to know, and transparent enough that its failures can be identified and corrected. That is a high bar. It is the right bar.