A credit-quality rating is not a framework. It does not set obligations — it scores how convincingly a single, specific project meets the obligations the frameworks already define. Where ISO 14068-1 and the ICVCM Core Carbon Principles describe what additionality, permanence, and accurate carbon accounting are, an independent rating answers a narrower and more practical question: for this project, this vintage, this batch — how likely is it that the claimed tonne was actually delivered? This page explains how the credit-ratings model works, why it sits across all seven frameworks rather than alongside them, and where its limits lie.
A carbon credit rating assesses the likelihood that the credits issued by a specific project have delivered on their claim to avoid or remove one tonne of carbon dioxide equivalent (tCO₂e). It is a forward-looking, probabilistic judgement about a single project at a single point in time — closer in spirit to a sovereign or corporate debt rating than to a certification standard. What S&P and Moody’s ratings are to debt markets, credit-quality ratings are intended to be for carbon markets: a comparable, independent signal of quality that lets a buyer weigh one issuance against another.
That distinction matters for how this page sits in the framework library. The seven frameworks are normative: they impose obligations a credit, claim, or programme must satisfy. A rating imposes nothing. It is an external assessment of how well a given project meets the quality concepts the frameworks define — additionality, permanence, conservative accounting, leakage control. A rating cannot make a credit conformant with ISO 14068-1, and conformance with a framework does not guarantee a high rating. The two operate on different axes: frameworks set the bar; ratings estimate how far over it a particular project clears.
A rating is also, explicitly, an opinion. It is not a statement of fact, not investment or financial advice, and not a recommendation to buy or refrain from buying. It does not address market value, price volatility, or any individual purchaser’s objectives. A buyer using a rating is expected to exercise independent judgement and conduct their own evaluation before acting. Treating a letter grade as a guarantee is a category error of the same kind as treating a credit framework as a rating.
A credit-quality rating resolves to a single letter grade on an AAA–D band. That headline grade is built from three core pillars — carbon accounting, additionality, and permanence. A fourth dimension, co-benefits, is scored on its own and kept out of the grade on purpose.
Confidence that the project has reported its activities accurately. 100% means measured activity matched the claim; below 100% flags a reporting-driven over-crediting risk.
A blended view of two questions: would the activity have happened anyway, and has its impact been correctly quantified? Carries the over-crediting risk assessment.
Likelihood the avoided or removed GHGs persist over the claimed period. Natural reversal risks and people-related risks are assessed and combined additively.
Biodiversity and community impact, mapped to the UN SDGs. Reported alongside the grade but excluded from it, so strong co-benefits can never inflate a weak carbon project.
The three core pillars determine the grade. Co-benefits are reported separately and do not change it.
The single most important mechanical feature of a credible rating is what it refuses to do: average. A strong pillar cannot buy back a weak one. This is what separates a genuine quality signal from a marketing score.
The limiting-factor principle
If the three pillars were averaged, a project with flawless carbon accounting but no credible additionality case would still score respectably — the strong pillar would mask the fatal one. A sound rating treats certain pillars as strictly limiting. Additionality is the clearest example: a project may report its activities perfectly and still warrant a low grade, because a tonne that would have been avoided anyway is not a tonne the buyer can claim, however accurately it was measured. The grade is governed by the binding constraint, not the flattering average.
Why accurate-but-not-additional still fails
The pillars answer sequential, non-substitutable questions. Carbon accounting asks: did the reported activity actually happen? Additionality asks: would it have happened without the credit revenue, and was its impact correctly quantified? Permanence asks: will the benefit last? A failure at any one of these undermines the claim regardless of performance on the others. Averaging would imply these are interchangeable. They are not — which is why a defensible methodology combines them through scoring logic and matrices that encode how scores constrain one another, rather than a weighted mean.
Bottom-up, not top-down
A grade derived only from top-down project characteristics — the certification methodology, the project type, the registry — is an insufficient signal, because quality varies widely within any single methodology. A rigorous rating is built bottom-up from project-specific evidence: primary performance data, independent measurement, and project documentation read in full. Top-down context is used to frame that analysis, not to substitute for it.
Worked illustration
Carbon accounting: 95%. The project measured and reported its activity almost exactly as claimed. On its own, this looks excellent.
Additionality: 2 / 5. Common-practice analysis suggests the activity was likely to have happened without carbon finance, and the baseline appears inflated — a material over-crediting risk.
Resulting grade: low. An average would have rewarded the accounting. The limiting-factor rule does not: accurately counting tonnes that were never additional does not produce a credit a buyer can stand behind.
Each pillar maps onto criteria already defined across the seven frameworks — but a rating goes further, scoring how convincingly a single project satisfies them rather than whether the box is ticked.
Carbon accounting — is the reporting accurate?
The carbon pillar verifies whether a project has accurately reported the activities that underpin its avoidance or removal claim, by benchmarking those reports against independent data. For nature-based activities — afforestation, avoided deforestation — developer-reported figures are tested against independent measurement such as satellite imagery and machine-learning classification of forest cover and change. For technology-based activities, third-party benchmark data is used where available. The output is a confidence percentage: the degree to which reporting is not a source of over-crediting risk. It maps to the “real” and “measurable” concepts in ISO 14068-1 §11.2 — but it scores the strength of the evidence, not merely its presence.
Additionality — would it have happened anyway, and was it counted correctly?
The additionality pillar blends two distinct sub-questions into one score:
Additionality of activities — the likelihood the project’s activities would not have been implemented without carbon-credit revenue, assessed through financial additionality, the policy and regulatory landscape, and common-practice analysis.
Over-crediting risk — whether the volume of credits issued is justified by the accounting: baseline quantification (is the counterfactual realistic, or inflated?), the conservatism of accounting parameters, leakage, and project-type-specific factors such as boundary definition and pre-project land-cover change. This maps to §11.2 “additional” and the conservative-baseline requirement in §11.2(c) — and is where most quality is won or lost.
Permanence — will the benefit last?
The permanence pillar assesses whether the avoided or removed GHGs are likely to be maintained over the period claimed — typically up to 100 years. Two risk families are assessed and then combined additively. Natural risks: historical and forecast exposure to fire, drought, flood, pests, and storms, including project-specific vulnerability under different climate scenarios. People-related risks: land-tenure and credit-issuance rights, free, prior and informed consent of affected communities, the developer’s access to capital, and geopolitical exposure. Because the risks are additive rather than averaged, multiple moderate vulnerabilities can compound into a low permanence score even where no single risk is catastrophic.
Co-benefits — scored, but held separate
Co-benefits capture biodiversity and community impact, mapped to the UN SDGs and assessed alongside species richness and the quality of activities that reduce pressure on biodiversity. The score is reported on a 1–5 scale — and deliberately excluded from the headline grade. The reasoning is structural: the grade exists to estimate the likelihood that the claimed tonne was delivered, and a generous co-benefits profile must never be allowed to lift the grade of a project that is underperforming on carbon, additionality, or permanence. Co-benefits can command a price premium or satisfy philanthropic criteria; they cannot substitute for carbon quality.
A defensible rating is the product of a two-stage process. First, a project-type-specific assessment method is developed and stress-tested. Second, that method is applied to an individual project through a structured, evidence-led workflow — and then kept under review.
Stage one — the method
Before any project is graded, a rating method is built for the relevant project type. It is rooted in the applicable crediting methodologies and the technical characteristics of that project type, developed against sample projects, reviewed by subject-matter experts and an external review process, and only then signed off for production use. The reason for type-specific methods rather than one generalised model is simple: different project types involve different activities and incentives, and a single blunt framework would miss the nuances that determine real quality.
Stage two — the rating workflow
A rating is only as good as the independence behind it and the honesty about its limits. Both deserve to be stated plainly, because both are routinely overstated in the market.
Independence is the whole asset
The value of a rating collapses the moment the rater has a stake in the outcome. A credible rating provider does not sell the credits it rates, is not paid by developers to rate their projects, and earns nothing by connecting buyers to supply — and is, ideally, held to account by an external governance body. This is the same structural principle that separates an independent verification body from the standard it assesses against: the party forming the opinion must not profit from the opinion being favourable. A grade issued by a party that also sells the credit is not a rating; it is marketing with a letter on it.
Transparency over the black box
A trustworthy grade is legible. The user can see the pillar scores, the sub-scores beneath them, how those scores are weighted and constrained, and the tests used to reach them — together with commentary explaining the rationale. A single letter with no visible reasoning invites exactly the misplaced confidence that ratings are supposed to dispel. The grade is the headline; the rationale is the product.
The limits, stated plainly
A rating is an opinion at a point in time, built substantially on information the rater cannot fully validate, and it can change. It does not address market value or price risk, and it is not a substitute for the buyer’s own due diligence, judgement, and — where a formal claim is being made — independent verification against a standard. Used well, a rating narrows the field and sharpens the questions a buyer should ask. Used badly, as a guarantee, it reintroduces the false confidence it was built to remove. The 360° Impact Portfolio treats the pillars a rating scores — carbon accuracy, additionality, permanence — as exactly the dimensions credits are screened on before they enter the portfolio, with verification and the disclosure pack doing the work a letter grade cannot.
A rating is not an eighth framework. It is the assessment layer that scores how well a project meets the obligations the frameworks define. Each pillar maps onto criteria already compared in the framework library.
Defines the quality bar
The CCPs define additionality, permanence, robust quantification, and no double counting at programme level. A rating scores how convincingly an individual project clears those bars — the pillars are the project-level expression of the same concepts.
Real · additional · measurable · permanent
§11.2 sets the credit criteria a rating’s pillars assess directly: the carbon score speaks to “real” and “measurable”, additionality to “additional”, and the permanence pillar to “permanent”. The standard sets the obligation; the rating estimates the strength of the evidence.
Durability and removals
The permanence pillar operationalises the durability question the Oxford Principles raise — scoring reversal risk project by project, which is the practical input to the portfolio-composition shift toward higher-durability project types over time.
Every credit in the 360° Impact Portfolio is screened on the same dimensions a rating scores — carbon accuracy, additionality, and permanence — with independent verification and a complete disclosure pack doing the work a single letter grade cannot. The 30-minute discovery call is the starting point.