In 2023, a joint investigation by The Guardian, Die Zeit, and SourceMaterial revealed that more than 90% of forest carbon credits issued by Verra, the world’s largest certifier, corresponded to no actual avoided deforestation. Hundreds of millions of tons of CO₂ supposedly offset existed only on paper. Multinational corporations had purchased these credits to display flattering carbon balances. The edifice cracked under the weight of evidence.

Two years later, the forest carbon credit market is not dead. It approaches $23.5 billion in 2026, even though the voluntary market had actually suffered a decline of approximately 61% in 2023 according to Bloomberg, at the height of the scandals. This paradox is only apparent. The scandals did not kill the market; they redefined what it was supposed to sell. What is being rebuilt is not the same thing as what collapsed.

The Essentials

  • The forest carbon credit market reaches approximately $23.5 billion in 2026, despite the collapse of several flagship projects
  • Verra’s new VM0048 methodology, published in 2023, standardizes baselines at the jurisdictional scale via third-party satellite data, removing developers’ freedom to define their own reference scenarios
  • The collapse of the Rimba Raya project in Indonesia showed that a certified project can be annulled by sovereign decision of a state, leaving buyers without recourse
  • Brazil is advancing an alternative approach based on restoring degraded ecosystems rather than protecting forests that might never have been cut
  • The unresolved question: who sets the quality standard, and who controls that issued credits correspond to a physical reality?

Ten Years of Credits Sold on Scenarios That Never Happened

To understand what is changing, one must understand what was broken. The REDD+ mechanism (Reduction of Emissions from Deforestation and Forest Degradation) rests on an apparently simple logic: if a forest that was supposed to be cut down remains standing, the difference between what would have been emitted and what actually is emitted constitutes a sellable carbon credit. Avoided deforestation has market value.

The problem lies in that “was supposed to.” To issue credits, one must construct a counterfactual scenario: at what rate would the forest have been destroyed without the project? This scenario, called the “baseline,” is almost impossible to verify. And it creates a powerful perverse incentive: the more the project promoter paints a dark future for the forest without his intervention, the more credits he can issue. Overestimating the threat means overestimating the benefits.

The 2023 investigations documented this mechanism at scale. On projects certified by Verra in the Amazon basin, Kenya, and Colombia, deforestation tracking algorithms showed that protected forests were no more threatened than control zones nearby. The threat was imaginary; the credits, real.

The Rimba Raya case added a different dimension to this picture. This Indonesian project, one of the world’s largest REDD+ projects with several million credits issued, saw its concession revoked by the Indonesian government in 2023 following a lengthy conflict over exploitation rights. Credits already sold to Japanese, American, and European companies found themselves without physical counterpart. The forests remained standing, but the legal architecture that allowed them to be monetized had collapsed. A risk that virtually no buyer had factored in: sovereign risk.

VM0048: Standardizing Baselines to Reduce Bias

Verra’s response to the trust crisis came in two stages. First, recognition of the problem, forced by journalistic investigations. Then the publication, in late 2023, of the VM0048 methodology, which deeply restructures how REDD+ credits are calculated.

The central change is a overhaul of how reference scenarios are constructed. VM0048 retains the counterfactual logic — projects continue to compare their results to a hypothetical “without project” scenario — but it now standardizes baselines at the jurisdictional scale. These references are calculated by a third party (Verra itself) using satellite data, instead of being freely defined by each project developer. The method also imposes geographic “buffer zones” that compare the project to similar forests located in similar socio-economic contexts, further reducing room for maneuver in constructing the baseline.

Concretely, this means that VM0048 projects generally issue fewer credits than their predecessors for the same area. A project that would have delivered one million tons under the old methodology might deliver 600,000 under the new one. For project developers, this is a commercial blow. For buyers who want to be sure their ton of CO₂ corresponds to a real ton, it is a credibility premium.

The adoption of VM0048 is not yet universal. Several existing projects still operate under older methodologies, including some of the most generous ones. The coexistence of different standards on the same market creates a segmentation that resembles organic food labeling in the early 2000s: labels proliferate, buyers get lost, and the quality premium remains difficult to establish transparently.

The Brazilian Bet on Restoration Rather Than Protection

One of the most interesting contributions to the debate comes from Brazil, which has assumed since 2023 a role as a political laboratory for forest carbon credits. Under the impetus of Lula’s Environment Ministry and the work of IBAMA, an alternative approach is taking shape: paying to restore degraded ecosystems rather than to protect forests that might perhaps have remained standing without the project.

The argument is solid. Restoring a hectare of degraded Atlantic forest or recovering a former extensive ranching area in the Cerrado produces verifiable additional carbon: trees growing absorb CO₂ measurably year after year. The counterfactual is not a hypothetical deforestation scenario; it is simply the absence of planted trees. The measurement is physical, the incremental benefit clear.

Brazil has a considerable area of degraded lands. According to MapBiomas, the combined degradation of the Mata Atlântica and Cerrado lies between 30 and 67 million hectares depending on the scenario chosen — a range that underscores both the magnitude of the potential and the measurement uncertainties. If a fraction of this area were engaged in credited restoration programs, the sequestration potential and associated revenues would be substantial for rural communities and Amazon federal states.

This logic joins questions explored elsewhere on ecosystem valuation. In other domains, we observe the same tension between a defensive protection approach, difficult to measure, and an active restoration approach, more verifiable but more costly. The debate on protecting ocean abysses illustrates an analogous dynamic: arbitrages between different forms of natural capital produce displacement effects that certifications struggle to capture.

The limit of the Brazilian model is permanence. Restoring a forest takes decades; the carbon credit is issued upon planting. If the forest burns, is illegally deforested, or abandoned in fifteen years, the credit sold today becomes retroactively fictitious. Guarantee mechanisms, called “buffer pools,” that set aside a fraction of credits to cover this risk remain calibrated on assumptions that have never been tested at the scale of a major climate crisis.

Who Controls the Controllers?

The central political question is not technical. It is institutional. Who decides the standards, and who verifies that issued credits correspond to a physical reality?

The voluntary carbon credit market is today a private space regulated by private organizations. Verra, Gold Standard, American Carbon Registry: these certifiers are financed by the project developers they certify. The relationship creates a structural conflict of interest that no one really contests, but that no one has yet resolved institutionally. The 2023 investigation showed that Verra was aware of methodological problems on REDD+ projects several years before journalistic investigations exposed them.

International governance of the voluntary market has advanced slowly. The Integrity Council for the Voluntary Carbon Market (ICVCM), created in 2021, published its “Core Carbon Principles” in 2023, which define minimum quality criteria. But the ICVCM has no sanctioning power; it labels methodologies, not individual projects. A project can operate under a methodology approved by the ICVCM and produce poor-quality credits if on-the-ground implementation is deficient.

COP29 in Baku in late 2024 took a step forward by activating Article 6.4 of the Paris Agreement, which creates a UN carbon mechanism with interstate supervision. Under this framework, credits traded under UN auspices would be subject to public governance. But the mechanism remains embryonic, and the private actors who built their business model on the voluntary market are not eager to see it superseded by a multilateral regulator.

Host states of projects, for their part, have strengthened their position. The Rimba Raya affair highlighted a reality investors preferred to ignore: a sovereign government can unilaterally annul a concession. Zimbabwe, Zambia, and Kenya have all adopted since 2022 regulations that impose a share of carbon revenues on states and local communities, sometimes exceeding 50%. This movement is legitimate; it profoundly modifies the geography of risk for developers and buyers.

Buyers Learn to Distinguish

On the demand side, something has changed. Large companies that were buying REDD+ credits in bulk to feed net-zero carbon declarations have become more cautious. Several suspended their purchases after the 2023 investigations. Others have begun to distinguish, in their sustainability publications, between avoidance credits (not cutting) and sequestration credits (planting, restoring), implicitly acknowledging that the former are less reliable.

Some institutional buyers have adopted a more demanding approach: independent verification by third parties not funded by the project, requirement for real-time satellite data, restitution clause in case of credit invalidation. These practices remain minority, but they sketch out a high-end market segment where transparency creates a premium.

Demand is also restructured by regulation. The European Union, with its CSRD directive and rules governing the use of carbon credits in sustainability reports, has tightened the conditions under which a company can display forest carbon compensation. Credits used for claims of “neutrality” or “net-zero emissions” face increasing scrutiny from competition regulators, particularly in France, the Netherlands, and Germany.

What the Market Can and Cannot Resolve

Optimism about market reconstruction must be tempered by a structural limit. Forest carbon credits can finance conservation and restoration. They cannot, alone, solve the causes of deforestation. In the Brazilian Amazon, deforestation is driven by expansion of cattle ranching and soy, two sectors integrated into global value chains. A REDD+ project protects a perimeter; it does not change the prices of agricultural commodities nor the land policies that determine pressure on the forest.

This is where complementarity with public policies becomes indispensable. Brazil reduced deforestation in the Amazon by approximately 83 to 84% between 2004 and 2012 — falling from 27,772 km² to 4,571 km² according to INPE/PRODES data — primarily through regulatory means: soy moratorium, mapping of protected areas, strengthening IBAMA capacity. Carbon credits contributed little to this decade of success. Their resurgence under Lula relies on this recovered regulatory foundation, not as a substitute.

The question of who captures revenues remains unresolved. Studies published in Conservation Biology and Nature Reviews Biodiversity converge in documenting that indigenous communities and local populations living in forests capture an infinitesimal share of revenues generated by REDD+ projects. Project developers, intermediaries, and certifiers absorb the majority of flows. This imbalance is both an injustice and a fragility: the most effective guardians of the forest are not those the market rewards best.

Experiments in more equitable revenue sharing exist. In Ecuador, projects developed in direct partnership with Waorani nations return more than 70% of revenues to communities. In Peru, the Amazonia Sur cooperative integrates community governance mechanisms into credit management. These models are more costly to operate and more difficult to certify, but they produce social durability that purely market-based projects struggle to achieve.

Toward a Two-Speed Market No One Wants

The forest carbon credit market will likely segment further in the coming years. On one side, a premium segment: VM0048 or equivalent projects, community governance, real-time satellite data, independent verification, measurable restoration credits. Higher unit prices, lower volumes, serious institutional buyers. On the other, a discount segment: older projects under generous methodologies, questionable baselines, light verification, buyers seeking the cheapest credit to meet obligations or improve their balance.

This segmentation is better than nothing. It creates a space where quality can differentiate itself and command value. But it allows the low-end segment to persist, which continues to offer carbon tax shelters with no physical equivalent to those who do not look too closely.

Public regulation is the only mechanism capable of structurally solving this problem. Not by replacing the market, but by setting a floor below which no credit can be sold under a climate claim. Article 6.4 of the Paris Agreement offers this framework in theory; its operationalization will remain the central political challenge of climate negotiations in the coming years.

In the meantime, the forest continues to grow, burn, and be cut down according to logics that markets capture imperfectly. The forest carbon credit is a real, imperfect, and perfectible tool. Its reconstruction on more solid foundations is a progress that the 2023 scandals made possible by forcing self-criticism on a sector unaccustomed to it. The next step belongs to public regulators and host states who, for the first time, hold part of the keys.

Sources

  1. Chaperon et al., “Quantifying additionality in forest carbon projects”, Conservation Biology / Nature Reviews Biodiversity, 2025 — https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12309627/
  2. Verra, VM0048 methodology — https://verra.org/methodologies/vm0048-reducing-emissions-from-deforestation-and-forest-degradation/
  3. Integrity Council for the Voluntary Carbon Market, Core Carbon Principles — https://icvcm.org/the-core-carbon-principles/
  4. MapBiomas, Annual Deforestation Report, 2024 — https://mapbiomas.org
  5. The Guardian / Die Zeit / SourceMaterial, “Revealed: more than 90% of rainforest carbon offsets by biggest provider are worthless”, January 2023 — https://www.theguardian.com/environment/2023/jan/18/revealed-forest-carbon-offsets-biggest-provider-worthless-verra-aoe
  6. Paris Agreement, Article 6.4, COP29 Baku decisions, November 2024 — https://unfccc.int/process-and-meetings/the-paris-agreement/the-paris-agreement
  7. VM0048 Verra v1.0 — official document — https://verra.org/wp-content/uploads/2023/11/VM0048-Reducing-Emissions-from-Deforestation-and-Forest-Degradation-v1.0.pdf
  8. Rimba Raya license revocation — S&P Global — https://www.spglobal.com/commodity-insights/en/news-research/latest-news/energy-transition/051324-rimba-raya-reignites-regulatory-and-transparency-concerns-for-carbon-projects
  9. Amazon deforestation reduction 2004-2012 — Nature Food — https://www.nature.com/articles/s43016-020-00194-5
  10. MapBiomas — Cerrado and Mata Atlântica degradation — https://brasil.mapbiomas.org/en/2024/07/05/ate-25-da-vegetacao-nativa-do-brasil-pode-estar-degradada/
  11. EP Carbon — VM0048 retains counterfactual logic — https://www.epcarbon.com/what-is-activity-data-and-how-is-it-used-in-vm0048/
  12. Article 6.4 COP29 — Carbon Direct — https://www.carbon-direct.com/insights/cop29-article-6-4-a-new-chapter-in-global-carbon-markets
  13. Nature Reviews Biodiversity 2026 — carbon markets and communities — https://www.nature.com/articles/s44358-026-00150-4