First-generation REDD+ projects, the flagship forest-protection credits traded on the voluntary carbon market, claimed 10.7 times more avoided deforestation than independent evaluations can justify. That’s not a rounding error. That’s an order-of-magnitude gap between what was sold to buyers and what actually happened in the forest.
Why it matters
REDD+ (Reducing Emissions from Deforestation and Degradation) credits have been the single largest category of nature-based carbon credits on the voluntary market for years. If the credits overstate their climate benefit by a factor of nearly 11, then companies and governments that retired those credits believing they offset real emissions were, in effect, paying for phantom tons. This paper, published in Nature Communications, doesn’t just confirm the problem. It pinpoints why it happened, and that diagnosis matters far more than the headline number.
The diagnosis: it’s the counterfactual, not the forest data
A team led by Tom Swinfield and Andrew Balmford at the University of Cambridge synthesized six independent, after-the-fact evaluations covering 44 REDD+ projects. The key finding is that most of these projects did reduce deforestation. They weren’t outright frauds. But the amount of deforestation they claimed to have prevented was wildly inflated. Here’s what’s interesting: the inflation wasn’t driven by disagreements over which satellite data to use for measuring forest cover. The choice of forest data turned out to be a secondary factor. The real culprit was twofold. First, selection bias in control areas. REDD+ projects need a “baseline” scenario: what would have happened to this forest without the project? To build that baseline, project developers chose comparison areas, places they argued were similar to the project site but unprotected. The problem is that developers had wide latitude to pick control areas that showed high deforestation rates. If your comparison group is losing trees fast, your project looks like a hero by contrast, even if the threat to your specific patch of forest was never that severe. Second, modeling approaches. The statistical methods used to project future deforestation were flexible enough that developers could, intentionally or not, generate inflated baselines. When independent researchers applied more rigorous counterfactual methods, the numbers shrank dramatically. Put simply: the projects were graded on a curve, and the developers got to draw the curve.
What the researchers say needs to change
The paper acknowledges that recent reforms in the voluntary carbon market are moving in the right direction. Transferring credit assessment to unconflicted third parties (rather than letting project developers mark their own homework) is critical. So is restricting the methodological flexibility that allowed cherry-picked baselines. But Swinfield, Balmford, and their co-authors argue these steps are insufficient. Their core recommendation: ex post certification against credible counterfactuals. In plain language, that means credits should only be fully certified after the conservation has actually been measured against a believable alternative scenario, not before, and not based on the developer’s own projections. This is a significant ask. Most REDD+ credits have been issued on an ex ante basis, meaning they’re sold based on projected future benefits. Shifting to after-the-fact verification would slow down credit issuance and reduce upfront revenue for project developers. But the researchers’ data makes a hard-to-ignore case that the current approach has produced credits that don’t represent real, causal reductions in deforestation.
Who benefits, who loses
If the field moves toward ex post verification with independent counterfactuals, a few things follow. Buyers get more confidence that a credit represents a real ton of avoided CO₂. Project developers who are genuinely protecting high-threat forests benefit because their credits become more credible and potentially more valuable. Developers who relied on inflated baselines lose out. The broader CDR and carbon removal community should pay attention here too. REDD+ is an avoidance credit, not a removal credit, but the credibility crisis around forest offsets has cast a shadow over all voluntary carbon markets. If nature-based avoidance credits can’t demonstrate additionality (the proof that the climate benefit wouldn’t have happened anyway), it erodes trust in the entire system, including for legitimate CDR purchases. It’s also worth being explicit: forest protection credits are not a substitute for cutting emissions. They address a real problem, tropical deforestation, but they should never be used as a license to delay phasing out fossil fuels. The moral hazard is real, and an order-of-magnitude over-crediting problem makes it worse.
Caveats
This study covers first-generation REDD+ projects. Newer methodologies and registries have already tightened some of the rules that allowed the worst abuses. The 10.7x figure is an aggregate across 44 projects; individual projects varied, and most did achieve some real deforestation reduction. The paper is also an early-access, pre-final-edit manuscript, so minor details could shift before final publication. And the six underlying evaluations it synthesizes used somewhat different methods, which the authors acknowledge. The big open question: will the voluntary market actually adopt ex post verification at scale, or will the economic incentives to issue credits quickly win out? The data in this paper suggests the answer to that question will determine whether forest carbon credits ever earn back the trust they’ve lost.
Source: nature.com