Forestry & Offsets: The Credibility Problem

An Offset Is a Claim About a Counterfactual

Every other asset in this series is a physical thing: a turbine, a pipe, a capture unit. A carbon offset is different in kind. It is not a tonne of anything. It is a claim that a tonne was avoided or removed that otherwise would not have been — a statement about a world that did not happen. And a world that did not happen cannot be measured. It can only be estimated, argued, and rated.

This is why the offset market’s recurring “scandals” are not really the problem, or rather, they are a symptom of a deeper one. The problem is structural and epistemic: the core quantity being sold — additionality, the difference between the actual world and the counterfactual baseline — is inherently unfalsifiable. You cannot go and check what would have happened. Two honest analysts can look at the same forest and disagree by a factor of three about how much of it was really at risk, and neither can be proven wrong. The credibility problem is not that some credits are fraudulent. It is that credibility itself has to be manufactured around an object that cannot supply it on its own.

Why Forestry Sits at the Soft End

Forestry and land-use credits are the largest category in the voluntary market — roughly 37% of all retirements in 2025, with avoidance-based REDD+ alone at about 25% — and also the most contested. That is not a coincidence. Nature-based credits compound the counterfactual problem with two failure modes the engineered credits mostly avoid.

The first is impermanence. A tonne locked in a tree is locked only until the tree burns, is logged, or dies of drought or pest — and in a warming climate those reversals are correlated with the very thing the credit is meant to address. The second is leakage: protect one forest and the logging pressure often moves to the next valley, so the accounting boundary determines the answer. Layer these on top of a baseline that is a prediction about deforestation that never happened, and forestry credits carry the softest claim in the market. The table below is the whole asset class arranged by how checkable its claim is.

Read top to bottom and the market’s entire recent behaviour becomes legible. It is sorting credits by durability and verifiability, paying up steeply for the bottom of the table and derating the top — not because the top is fraudulent, but because its claim is the one that cannot be checked.

The Market Repriced Along the Durability Axis

The result is not a market that shrank; it is a market that split. Buyers stopped treating credits as interchangeable tonnes and started paying for the attributes they can defend — additionality, permanence, and measurement — which has opened an enormous price spread between the soft and durable ends.

Analysts describe the current market as “smaller and sharper” — consolidating around a handful of high-credibility projects rather than recovering in aggregate. Quality now drives demand. And the sorting is fine-grained: within forestry, buyers rotated toward Improved Forest Management, whose retirements roughly doubled into 2025 — even though formal integrity certification for that category only arrived in December 2025. Buyers moved before the quality stamp existed, which tells you the repricing is being led by fear of reputational exposure, not by the standards bodies.

Credibility Was Manufactured, Not Found

Because the object cannot supply its own credibility, an entire industry has grown up to manufacture it around the object. This is the most investable observation in the piece. The response to the integrity crisis has not been to fix offsets — it has been to build a credibility stack on top of them: standards that set a quality threshold (the Integrity Council’s Core Carbon Principles), claims codes for buyers (VCMI), independent ratings agencies that score projects, digital measurement and monitoring that cuts verification cost, and a nascent carbon-credit insurance market that underwrites reversal and invalidation risk.

Each of these is a business selling trust into a market that cannot generate its own. That is a durable role precisely because the underlying problem is permanent: as long as the unit is a counterfactual, someone has to certify, rate, monitor and insure it. The credibility infrastructure is, in an important sense, a better asset than the credits it rates — it earns whether the credits appreciate or derate, and its relevance rises with every scandal.

The Incentives Point at Over-Crediting

Follow the money through the market and the credibility problem stops looking like an accident. Every party to a credit has an incentive that points the same way — toward issuing more tonnes than the atmosphere actually saw. The buyer wants the cheapest defensible claim against its target. The project developer is paid per tonne, and gets to propose the baseline that sets how many tonnes exist. And, critically, the verifier is usually paid by the party it is verifying — the same conflict that sat under credit-rating agencies before 2008. When everyone at the table is paid more if the number is bigger, the number tends to be bigger.

The opacity compounds the incentive problem. When most retirements are anonymous, demand signals cannot be verified, corporate progress cannot be audited, and the market cannot build the reputational feedback loop that would discipline quality. This is the mirror image of the “greenhushing” that has buyers going quiet to avoid criticism — and a market where the buyers hide is a market where the standards bodies, not the customers, have to do all the disciplining. That is a fragile way to build trust.

Where the Value Migrates

Put the structural read together and the flows resolve. Value is leaving the soft, counterfactual-heavy end of the market and migrating to three places: durable removal, the credibility infrastructure, and the compliance frontier.

Durable removal — engineered capture, BECCS, biochar — wins because its claim is checkable and its supply is scarce; the largest technology buyers have already made it the benchmark, with a single hyperscaler accounting for the majority of durable removal purchased in 2025. The credibility infrastructure — ratings, digital MRV, registries, insurance — wins because it is paid to certify a permanently uncertifiable object. And the compliance frontier matters most of all: as voluntary credits get pulled into regulated systems — Article 6 of the Paris framework, aviation’s CORSIA, the EU’s carbon-border and green-claims rules, India’s move from a voluntary to a compliance Carbon Credit Trading Scheme, and California’s disclosure statutes — a mandatory demand floor forms under the credits that qualify, and vanishes under those that do not.

Reading It Through the Frameworks

Where does policy become the cash flow? Increasingly, it is the whole story. A voluntary offset’s value is a reputational judgement and can evaporate with a single investigative report. A compliance credit’s value is a legal obligation to surrender it, which does not. As Article 6, CORSIA, CBAM-adjacent rules and national schemes mature, the credits that gain a regulatory use-case acquire a demand floor and a defensible price, while purely voluntary avoidance credits keep trading on sentiment. The migration from voluntary to compliance is the single most important repricing event in this market.

What kind of risk is it? An offset is a short position in its own credibility — the holder is exposed to the day the counterfactual is re-examined and found wanting. That risk is uncorrelated with the project’s physical performance and highly correlated with scrutiny, which is why it clusters and re-rates in waves. The discipline is to buy the attribute you can defend — permanence, measurement, a compliance use-case — and to treat a cheap avoidance credit not as a bargain but as an unhedged reputational liability sitting on the balance sheet.