A biodiversity credit is a receipt. Someone measured a change in a defined patch of nature, converted that change into units, and sold you a claim on the units. The measurement can be careful, the register public, the buyer acting in complete good faith. It is still a receipt — and a receipt travels.
That one structural fact explains most of what confuses people about this market. The unit is transferable. The place is not.
what a biodiversity credit actually is
A biodiversity credit is a quantified, verified, tradable unit of biodiversity uplift — produced against a published methodology, recorded in a register, sold to a buyer who wants to say they paid for it. BloombergNEF draws the line the market has mostly settled on: credits are tradable units of uplift, while offsets compensate for loss a permitted development is going to cause anyway.
The same instrument travels under several names. Biocredit, biodiversity certificate, nature credit, and in some jurisdictions a statutory biodiversity unit all describe a measured quantum of nature that changes hands — product dialects, not four asset classes. Credits are named explicitly in Target 19 of the Kunming-Montreal Global Biodiversity Framework. Two shapes circulate: ex post, issued after a change is measured on the ground, and ex ante, sold against a projected one.
Now the part most explainers skip. A credit is not a lien, a covenant, an easement, or a title, and it does not encumber the land. Statutory schemes usually do put a management commitment alongside the unit — England's habitat banks carry a 30-year obligation, US wetland mitigation banks long-term site protection instruments — but it lives in the land agreement, not in the unit you bought. The unit is the receipt for an arrangement held elsewhere, by someone else.
Which is not a design flaw. A thing that stays attached to one hillside cannot clear a market.
credit, offset, contribution, certificate
Four words get used interchangeably in the same meeting, and they describe four different purchases.
| what it is | the claim it supports | what has to be true | how it fails | |
|---|---|---|---|---|
| credit | a tradable unit of measured biodiversity uplift | "we funded a measured gain" | metric, baseline, attribution and permanence all hold | the baseline or attribution doesn't hold and the unit was noise |
| offset | a unit compensating for a permitted loss elsewhere | "the damage we caused is balanced" | the gain is genuinely equivalent to the loss | the loss already happened and the gain is not like-for-like |
| contribution | money that funds nature without buying a neutrality claim | "we paid for this work" | the money reached the work | the money is real but the buyer's own impacts continue |
| certificate of ensurance | a position issued one-to-one against an agent standing for a named place | "we hold a funded position in this place, now" | that place's condition is observable now | condition degrades and the funding was too small or too late |
Two of those rows carry a claim about somewhere else. Two do not. That distinction, not integrity grading, is the useful axis when deciding what to buy.
The contribution row is the salvage most serious critics have landed on. Barbara Haya's work at Berkeley moved from fixing offset quality to arguing money should be given as a contribution rather than sold as a compensating unit, with the buyer keeping full responsibility for its own impacts. A worse product. A more truthful one.
how the unit gets made
Reading the pipeline is what lets you judge a methodology without a consultant.
- Choose a metric — species richness, habitat area times condition score, an integrity index, a basket. This decides everything downstream.
- Quantify it — survey the site, score the habitat, model what you cannot survey.
- Detect change and attribute it — show the metric moved, and that your intervention is why, against a counterfactual that did not happen.
- Adjust issuance — deduct a buffer for uncertainty and permanence, deduct for leakage, issue units.
Every step is a judgment call with a defensible range, and the ranges compound. Wauchope and colleagues reviewed a sample of these methodologies for Proceedings of the Royal Society B in 2024 and found deep uncertainty — the kind that cannot be quantified away — through the whole chain: ecological measurement error is large, most metrics swing between years for reasons unrelated to management, and short crediting windows therefore risk what they call commodifying noise. Their conclusion is not a rejection: credits can help in a contribution model, where they fund measurable work, and much less easily where they support a "net" claim about harm somewhere else.
That is a review of methodologies, not an accusation against the people writing them. The problem is hard because biodiversity is not a tonne.
the useful half, stated properly
The credit industry has built two things nature finance badly needed. A register — a public list of who created what, where, against a published method — is infrastructure, and conservation spent decades unable to say that much about its own outcomes. A payor appears wherever a rule requires a unit; England's mandatory biodiversity net gain regime is the clearest case, and how far a statute gets you has its own page.
Once a rule names a payor, a unit, and a consequence, ordinary capital shows up — as land finance rather than as conservation. In September 2026 Atom Bank launched loans of £100,000 to £2 million at up to 65% loan-to-value against English habitat-bank land, after funding three sites that registered 480 biodiversity units with Natural England, enough for roughly 2,410 homes. US wetland mitigation banking has been a real land market under the Clean Water Act for decades. These are land deals with registers attached, and they behave like land deals.
Voluntary demand is the honest disappointment. BloombergNEF's Biodiversity Finance Factbook, COP16 edition of October 2024, estimated less than $1 million of biodiversity credits purchased. Larger cumulative figures circulate for this market; they count different things over different periods, so treat any single headline number with care.
Ask buyers directly and the answer is not "no," it is "not like that" — a 2026 study of German food companies by the Global Nature Fund with GIZ and Food for Biodiversity found 89% required a clear connection to their own supply chains and no respondent chose credits alone. They want the work done in the landscapes they depend on, which is a request for a hold described in credit vocabulary.
what the receipt cannot do
You may be reading this as "credits are a scam." That is not the claim, and it is a lazy position that gets people out of thinking. The claim is narrower: fungibility is the feature that makes a unit clear a market, and the same feature lets the claim travel away from the site that produced it. Once the receipt is in a sustainability report and the money is spent, nothing in the instrument requires anyone to keep watching the hillside. Whether the place stays in condition depends on the land agreement, the steward, and the next twenty-nine years of somebody's budget — none of which the buyer holds.
A credit is a receipt for a unit. It is not a hold on the living system.
A second failure mode belongs in any honest definition. Units are created on land, and land has occupants and rights holders. Consent, tenure, and who receives the revenue are live questions wherever credits are generated on Indigenous or community territory — raised by those communities and their own organisations, who do not need us to speak for them.
None of this makes a credit useless. It makes it a claim about a gain, priced and traded, one layer away from the place. If a gain claim is what your obligation needs, buy the credit and read the methodology properly. If you need the place itself to keep working, you are shopping for something else.
the people already paying
The money is already moving — for the failure rather than for the system. The developer buying statutory units pays at the most expensive point in the sequence, after the design is fixed. The food manufacturer restocking a failed sourcing region pays. The utility treating sediment a watershed used to hold pays. The government funding restoration after a permitted loss pays twice for the same hectare. None of them lacks a budget. What they lack is an instrument that funds the present condition of a named living system and books it as something other than charity.
what a hold on a named place looks like
Ensurance is our attempt at that instrument. The claim is deliberately narrow.
A certificate — specific ensurance — is issued one-to-one against an agent, an account standing for a named natural asset. Funding attaches to that place with its condition on the record, and proceeds route to its stewards. A certificate is not a credit: not fungible with another place, not sold as neutrality, no secondary offset book. It buys the present tense — the system is either producing services now or it is not, and you can go and look.
Underneath sits the accounting — ecosystem stocks, the flows they produce, measured condition. That number is a bridge that lets capital see a place. It is never the worth of the place, and a nature product that forgets which is which has stopped being useful.
This is not a claim of arrival. The agents, the places, the coins, and the certificates are live; the volumes are small. Anyone calling a nature finance product finished — ours included — is describing a roadmap.
Absorb, don't replace. A habitat bank with a register and a funded hold on the same parcel is in better shape than either alone. The credit industry is doing the unglamorous work of making nature legible to a procurement committee. The column still missing is who pays to keep the place in the condition the unit assumed.
frequently asked questions
what is a biodiversity credit?
A biodiversity credit is a quantified, verified, tradable unit of biodiversity uplift, produced against a published methodology and recorded in a register. An offset, by contrast, compensates for a permitted loss elsewhere. Neither obliges the buyer to keep the site in condition — that obligation sits in a separate land agreement.
how do biodiversity credits work?
A project chooses a metric, quantifies it on a defined site, shows the metric changed and that the project caused the change, deducts buffers for uncertainty and leakage, then issues units — sold after the measured change (ex post) or against a projected one (ex ante). The buyer gets a claim, the register an entry, the landowner the management commitment.
how is a biodiversity credit different from ensurance?
Where the claim points. A credit is a transferable unit of measured gain, so it can be spent somewhere other than the place that produced it. A certificate of ensurance is issued one-to-one against an agent representing a named place and funds that place's present condition — non-fungible, no neutrality claim, no offset equivalence. One is a receipt for a gain, the other a hold on a living system. See what ensurance is.
the series
Six posts on credit words for living systems — what each instrument is, what the unit cannot reach, and who funds the place between transactions.
- what a biodiversity credit actually is — the definition and the receipt trap (this post)
- the carbon template biodiversity is copying — offset evidence as the honesty layer
- a habitat bank is a land deal, not a vibe — mitigation banking as a real land market
- nature credit vs carbon credit — same receipt shape, different thing counted
- an offset is permission to destroy nearby — the mitigation hierarchy and where it ends
- what you hold if you want the place to stay — the alternative purchase, stated plainly
Adjacent: demand by statute on how a rule manufactures a buyer, and capital doesn't invest for nature on why policy produces commitments and only a product produces a transaction.