Carbon credit risk analysis · risk ratings
Carbon credit risk analysis: the four tests behind every risk rating
Carbon credit risk analysis asks one question: is this a real, additional, durable tonne?
Every carbon risk rating answers it through the same four tests — additionality, the baseline, permanence and leakage — scored by each agency in its own way.
This page takes each test in depth, then sets out how BeZero, Calyx Global and Sylvera describe their own assessment.
The framework
Four tests, one question
The four tests are not an invention of any one agency.
Additionality and permanence are two of the “emissions impact” principles in the ICVCM Core Carbon Principles, beside robust quantification and no double counting.
Carbon Market Watch’s comparison of the rating agencies found the agencies differ most in how they score exactly these questions: additionality, double counting, leakage, buffers and permanence.
For a UK reporter, the same tests sit behind UK SRS S2 paragraph 36(e)(iv), which asks for any factors needed to understand the credibility and integrity of the credits, with permanence as its example.
How that disclosure works is set out in carbon credits in sustainability reporting.
| Test | The question | What goes wrong |
|---|---|---|
| Additionality | Would it have happened anyway? | Credits paid for a project that needed no credit revenue. |
| Baseline | Was the counterfactual realistic? | An inflated baseline issues more credits than tonnes. |
| Permanence | Will the carbon stay stored? | Fire, disease or land-use change reverses a removal. |
| Leakage | Did the activity just move? | Emissions displaced outside the project boundary. |
Test 1 · Additionality
Would it have happened anyway?
A reduction is additional only if it would not have happened without the revenue from selling the credit.
A project that fails this test is over-credited from the day it is registered, however well it is run afterwards.
On 6 August 2024 the ICVCM refused the CCP label to eight renewable-energy methodologies, as “insufficiently rigorous in assessing whether the projects would have gone ahead without the incentive of carbon credit revenues”.
Those methodologies covered about 236 million unretired credits, 32% of the voluntary carbon market at the time.
The position moved on 30 April 2026, when the ICVCM approved a revised methodology, VMR0017, for new issuances subject to a benchmark analysis.
It said in the same decision that “no historical issuances will be CCP-Approved”, so neither “renewables can never carry the label” nor “the old credits are rehabilitated” is right.
The renewable-energy ruling
- 6 Aug 2024Eight methodologies refused
About 236 million unretired credits, 32% of the voluntary market at the time.
- 30 Apr 2026VMR0017 approved for new issuances
A revision, subject to a benchmark analysis. No historical issuances will be CCP-Approved.
Source: ICVCM release, 6 August 2024; ICVCM assessment status.
Test 2 · Baseline and over-crediting
Was the counterfactual realistic?
Every credit is measured against a baseline: what would have happened without the project.
Assume a higher deforestation rate than the real one, or credit against an unrealistically high historical emissions rate, and the project issues more credits than tonnes.
Calyx Global groups this with project emissions and leakage under one heading, over-crediting, in its GHG rating approach.
The largest single example is structural.
Verra did not submit its older REDD+ methodologies — VM0006, VM0007, VM0009, VM0015 and VM0037 — for CCP assessment, and its programme notices moved projects towards the newer VM0048.
The ICVCM records that those legacy methodologies produced the majority of REDD+ credits on the market and around a quarter of all carbon credits retired in 2023, and that credits under them can never receive the CCP label.
“Verra is CCP-Eligible” is true and tells you almost nothing about a given Verified Carbon Unit: always ask which methodology, and which version.
Test 3 · Permanence
Will the carbon stay stored?
Forestry and soil carbon carry reversal risk from fire, disease, drought and land-use change that geological storage largely avoids.
Even “permanent removal” spans a wide range: Puro.earth’s durability classes run from 100+ years for terrestrial biomass to 1,000+ years for geological storage.
Programmes manage reversal risk with buffer pools: a share of credits held back to cover losses across the portfolio rather than any one project.
The UK’s Woodland Carbon Code keeps its buffer in a single account held in the UK Land Carbon Registry and managed by Scottish Forestry.
It also issues predicted carbon first as Pending Issuance Units — in the Code’s words “a promise to deliver a Woodland Carbon Unit in future, but they are not guaranteed” — which convert only after verification.
UK SRS S2 names permanence as its example of what a company should explain about the credits it plans to use.
Test 4 · Leakage
Did the activity just move?
Leakage is emissions displaced rather than avoided: logging pushed to the next valley, or emissions-intensive production relocated.
It happens outside the project boundary, which makes it the hardest of the four to observe from project data.
Carbon Market Watch lists leakage among the areas where the agencies’ approaches differ.
Calyx Global counts it inside over-crediting, beside baselines and project emissions.
A fifth risk sits beside the four: the same tonne claimed twice, which the ICVCM calls double counting and Calyx Global assesses as overlapping claims.
How the agencies assess them
Three agencies, described from their own material
A factual summary of what each agency publishes about its own approach, read on 30 September 2026 — not an audit of any of them, and not a ranking.
| BeZero Carbon | Calyx Global | Sylvera | |
|---|---|---|---|
| What the grade expresses | Its "current opinion on the likelihood that a given credit achieves a tonne of CO₂e avoided or removed". | Greenhouse gas integrity. | Project quality, AAA the highest grade and D the lowest. |
| Factors it names | Risk factors unchanged since its 2023 scale change; broader risks such as fraud or default considered only as they affect carbon outcomes. | Additionality; over-crediting (baselines, project emissions, leakage); permanence; overlapping claims. | Carbon score, additionality and permanence. |
| Scale | Eight points, AAA to D. | Eight points, AAA to D, since 21 January 2025. | AAA to D. |
| Kept outside the grade | Co-benefits such as biodiversity or social impacts. | SDG impacts, rated separately +1 to +5. | Co-benefits, scored separately 1–5. |
| Stated limit | Classifying a project as "failed" is for standards bodies, not the rating. | Moved scale to make comparison across agencies easier. | May be unable to rate where there are fundamental red flags or missing data. |
Sylvera says it builds frameworks per project type rather than one generic template, so a removal project and an avoidance project are not scored alike.
BeZero says its 2023 move to eight points changed the scale, not the assessment, in its scale note.
Where agencies disagree on a project, that reflects different methodological choices applied to the same evidence, not proof that one is wrong.
The scales themselves, and what a letter means on each, are compared in carbon credit ratings.
Reading the ICVCM table
“Withdrawn” means undecided
The ICVCM’s assessment status table uses a word that is misread in both directions.
A programme may withdraw a methodology or version during assessment, and the ICVCM says that in that case “no Decision will be taken”.
So Withdrawn is not rejected, and it is not approved-and-archived either: no decision exists.
The table listed 44 CCP-Approved methodologies when it was updated on 12 August 2026.
Isometric DAC v1.0 is marked Withdrawn.
Isometric DAC v1.1 is CCP-Approved.
Versions 1.2 and 1.3 were still in assessment on 12 August 2026.
A “CCP-approved DAC credit” means nothing without the version number.
Two meanings
A credit’s risk rating is not a company’s
“Carbon risk rating” is used for two different products.
A carbon credit risk rating is the subject of this page: an opinion on whether a credit delivers its tonne.
A company’s carbon or ESG risk rating assesses the company; MSCI’s ESG Ratings, for example, are industry-relative and run on a seven-band scale from AAA to CCC.
Both are unregulated opinions today.
The ESG Ratings Order 2025 requires FCA authorisation for certain ESG ratings from 29 June 2028, where the rating is likely to influence a decision to make an investment specified in Part 3 of the Regulated Activities Order.
A spot voluntary credit is not named as a specified investment there, though a future or contract for difference over one may be, so whether the regime reaches carbon credit ratings is an open question.
That separate perimeter question is set out in carbon credit regulation in the UK.
The standards the disclosure is written to are set out in the UK Sustainability Reporting Standards reference library.
To talk through a project or portfolio question, you can book a free 15-minute call.
Frequently asked
Questions people ask
What is carbon credit risk analysis?
The assessment of whether a carbon credit is likely to represent a real, additional and durable tonne of CO2e. In practice it comes down to four tests: additionality, the baseline and over-crediting, permanence, and leakage. Rating agencies publish their conclusions as letter grades; a buyer can run the same tests against a project's own documents.
What is a carbon risk rating?
The phrase is used for two different things. A carbon credit risk rating, such as those published by BeZero Carbon, Sylvera or Calyx Global, is an opinion on whether a credit delivers its tonne. A company's carbon or ESG risk rating, such as an MSCI ESG Rating, assesses the company, not a credit. This page covers the first.
What is a carbon risk score?
Usually a rating agency's grade or factor score for a carbon project. The agencies score the same broad risk factors but define and weigh them differently, so a score from one agency cannot be read on another agency's scale.
What are the main risk factors for a carbon credit project?
Additionality (would it have happened anyway?), the baseline and over-crediting (was the counterfactual realistic?), permanence (will the carbon stay stored?) and leakage (did the activity simply move elsewhere?). Calyx Global adds overlapping claims — the same reduction being claimed twice.
How do carbon credit ratings compare across Sylvera, BeZero and Calyx?
All three use a letter scale from AAA to D. BeZero rates the likelihood that a credit achieves a tonne. Sylvera rates on carbon score, additionality and permanence. Calyx Global rates greenhouse gas integrity across additionality, over-crediting, permanence and overlapping claims. The same letter from two agencies is two different opinions.
Can renewable energy credits carry the CCP label?
Historical ones cannot. In August 2024 the ICVCM refused the label to eight renewable-energy methodologies as insufficiently rigorous on additionality. On 30 April 2026 it approved VMR0017, a revision, for new issuances subject to a benchmark analysis, and said no historical issuances will be CCP-Approved.
Does "Withdrawn" on the ICVCM table mean a methodology was rejected?
No. The ICVCM says a programme may withdraw a methodology or version during assessment, and that no decision will be taken on it. Withdrawn means undecided, not rejected and not approved.
Are carbon risk ratings regulated in the UK?
Not today. The ESG Ratings Order 2025 requires FCA authorisation for certain ESG ratings from 29 June 2028, where the rating is likely to influence a decision to make an investment specified in Part 3 of the Regulated Activities Order. Whether that reaches ratings of carbon credits is an open question.
Sources
Primary sources
Every figure, date and status on this page traces to the instrument’s owner. Secondary commentary is never the source for a number.
- ICVCMThe Core Carbon Principles
Additionality, permanence, robust quantification and no double counting among the ten.
- ICVCMAssessment status (table updated 12 August 2026)
44 CCP-Approved methodologies; "Withdrawn" means no decision; VMR0017 approved 30 April 2026.
- ICVCMCarbon credits from current renewable energy methodologies will not receive the CCP label (6 August 2024)
About 236 million unretired credits, 32% of the voluntary market.
- VerraVerified Carbon Standard programme — methodology notices
Legacy REDD+ methodologies inactivated or transitioned to VM0048.
- BeZero CarbonThe BeZero Carbon Rating scale explained (13 March 2023)
Eight points AAA–D; risk factors unchanged.
- BeZero CarbonThe BeZero Carbon Rating: definition and rating scale
What the rating does and does not assess.
- Calyx GlobalA new GHG rating scale (21 January 2025)
Four GHG dimensions; over-crediting includes baselines, project emissions and leakage.
- SylveraCarbon credit ratings
Frameworks per project type.
- SylveraCarbon credit ratings terminology (31 January 2022)
Carbon score, additionality and permanence; projects that cannot be rated.
- Scottish ForestryWoodland Carbon Code Standard v3.0 (clarifications January 2026)
The pooled buffer, and Pending Issuance Units converting on verification.
- Puro.earthDefining carbon removal durability
Durability classes by removal method.
- legislation.gov.ukFSMA 2000 (Regulated Activities) (ESG Ratings) Order 2025, SI 2025/1349
Authorisation from 29 June 2028; the Part 3 investment test.
- MSCIMSCI ESG Ratings methodology
A company-level rating on a seven-band AAA–CCC scale — not a carbon credit rating.
- Department for Business and TradeUK SRS S2 Climate-related Disclosures, ¶36(e)(iv)
Credibility and integrity factors, with permanence as the example.
Continue reading
Read next
Carbon credit ratings
What a grade from each agency says, and why the same letter differs.
Carbon credits in sustainability reporting
UK SRS S2 ¶36(e), ESRS E1 and the offset-claim rules.
Carbon credit regulation in the UK
Is trading a carbon credit a regulated activity? Spot, derivatives and allowances.