The review approach examines the evidence behind each material value in the file. It looks at what exists, whether sources corroborate one another, where they conflict and what remains missing.
The references below separate what the law requires from how WEETRA structures evidentiary review and from controls that WEETRA adds for robustness. The objective is not to produce an impression of overall quality, but to make the evidentiary reasoning explicit enough for another professional to follow and challenge.
Version references, Evidence Manifests and SHA-256 integrity references are WEETRA controls. The Regulation does not prescribe those specific tools.
Official sourceRegulation (EU) 2023/956 · Art. 7(5)-(6) · Annex V · Art. 19(2)
Records supporting embedded-emissions calculations must be sufficiently detailed for verification and later review, and retained until the end of the fourth year after the relevant declaration year.
WEETRA tests whether the reviewed corpus preserves the source, fact and period relationships needed to reconstruct the evidentiary state later.
Version references, Evidence Manifests and SHA-256 integrity references are WEETRA controls. The Regulation does not prescribe those specific tools.
Official sourceImplementing Regulation (EU) 2025/2546 · Art. 5
The verifier applies 5% materiality levels to total specific embedded emissions and total specific embedded free allocation, while also applying expert judgement to other or smaller misstatements and non-conformities.
WEETRA may identify contradictions and inconsistencies upstream. A WEETRA finding is not a verifier’s materiality conclusion.
Official sourceImplementing Regulation (EU) 2025/2547 · Arts. 13-15
For complex goods, precursor reporting periods and source installations can affect the embedded-emissions calculation. Evidence may be required to establish the actual precursor period or installation subset.
WEETRA examines whether precursor values can be traced to the relevant producer, installation, period and evidentiary basis.
Official sourceRegulation (EU) 2023/956 · Art. 9(1)-(3)
A reduction based on carbon price paid requires records demonstrating the carbon price, relevant rebates or compensation, independent certification and evidence of actual payment, with retention through the fourth year.
WEETRA examines whether the file supports the carbon-price treatment being claimed and records documentary limitations where the evidentiary chain remains incomplete.
Primary legal texts prevail. Commission guidance is explanatory and is identified separately from binding law.
On this page
On this pageA value is not yet a chainFrom documents to DeterminationThree conditionsFact and sourceControls, not scoresProfessional reviewReconstructibilityA regulatory file arrives as workbooks, certificates, invoices, monitoring reports, correspondence and system exports. The regulatory position itself does not rest on the mere presence of those documents. It depends on the facts they establish and on the relationships between those facts.
A document may contain a number, a declaration or a statement. That alone does not establish what goods it concerns, who produced them, at which installation, during which period, where the information originated, what calculation or review approach produced it, whether another source contradicts it, or whether the source is capable of supporting the specific claim being made.
WEETRA treats documents as evidence carriers, not as conclusions. The review distinguishes what the reviewed corpus establishes, what conflicts within it, what remains unsupported, and what can only be concluded within defined source limits.
The CBAM page describes this problem where it is currently most acute. This page describes the reasoning used to examine that evidence, independent of any single regulation.
A supplier workbook states an emissions intensity of 1.84 tCO₂e per tonne. The number exists and can be extracted. It may also be arithmetically correct.
None of that establishes that it can support a specific regulatory position for particular goods over a particular period. Between the cell in the workbook and the claim it is being used for sits a series of relationships, and each one is either established by evidence or it is not.
Illustrative. The relationships that apply depend on the regulatory scope, the mandate and the evidence available.
A review that stops at the number has checked arithmetic. A review that follows the chain has examined whether the number can be relied upon. These are different exercises, and they produce different kinds of confidence.
The review approach moves through defined stages. Each stage narrows what is being asserted and records what supports it, so that the reasoning remains legible after the fact.
Public review approach. Which stages apply, and at what depth, follows the Review Level and the mandate.
Files and records received inside the defined review corpus. What lies outside the corpus lies outside the review, and that boundary is recorded rather than assumed.
Relevant evidence is identified and associated with provenance, version and integrity references according to the applicable process.
Statements, values and relationships are identified from the evidence, or recomputed where the mandate permits.
Facts relied upon by the review are represented in a controlled structured form while remaining linked to their provenance.
Controls examine defined facts, relationships, documentary conditions or calculations according to applicability.
Conflicts, missing support and bounded source conditions remain explicit. They are carried forward rather than smoothed away.
Where professional judgement is required, an authorised reviewer examines the relevant evidence and unresolved conditions.
The Determination reflects the reviewed evidentiary state within mandate.
The versions, relevant reasoning and review state relied upon are preserved according to the applicable process and Review Level.
This is the public shape of the review approach. The internal review rules, reasoning rules and review structure are not published.
Much of the value of a review lies in refusing to blur three situations that a generic assessment tends to merge into one.
Necessary evidence for a fact or relationship has not been identified in the reviewed corpus. Illustrative example. An installation is named, but no reviewed evidence connects the relevant goods to it.
Two or more relevant sources support incompatible states affecting the reviewed issue. Illustrative example. A supplier workbook identifies Installation A while production records identify Installation B for the same goods and relevant period.
The available evidence or authoritative source material permits only a bounded conclusion. Illustrative example. A document establishes the supplier but does not establish the producing installation.
These are different evidentiary conditions with different consequences and different remedies. Missing evidence may be obtainable. A contradiction has to be resolved before either side can be relied upon. A source limitation may be irreducible, in which case the honest output is a bounded conclusion rather than a confident one. Collapsing all three into a single risk score destroys precisely the information that makes the review useful.
Consider a structured fact. INSTALLATION = TR-001. Read in isolation, it is close to useless from an evidentiary perspective. It asserts a state without carrying anything that would allow that state to be tested.
Where the point is relevant, the review should be able to answer
Which document or record established the value.
Which version of that document, where more than one exists.
Which reporting period the value belongs to.
Whether the value was declared, extracted from evidence, or recomputed.
Whether another source in the corpus supports a different state.
Whether later evidence displaced the earlier state, and from when.
Structured facts must remain connected to the evidence capable of supporting them. A fact detached from its provenance can still be reported, but it becomes difficult to defend. Once challenged, it is little more than an assertion.
WEETRA uses controlled review positions to examine defined evidentiary relationships, requirements or calculations according to applicability. A control is not an impression of whether a document looks compliant. It is a defined question asked of a defined state.
Some inconsistencies appear only when data families are read against one another rather than in isolation. A production volume incompatible with reported energy consumption, or an emissions intensity incompatible with the declared physical characteristics of an industrial process, will not surface in a document-by-document check. Controls exist to reach that class of inconsistency.
WEETRA has established approximately 140 points of control across ten areas of examination. The figure describes the breadth of WEETRA’s review, not a fixed checklist applied identically to every engagement. Applicability depends on the regulatory scope, the mandate, the Review Level, the evidence available, the source conditions and the facts actually relevant to the position under review.
The catalogue itself, the individual positions, their decision rules and their internal parameters are not published.
A structured review may involve several kinds of examination
Where professional judgement is required, an authorised reviewer examines the relevant evidence, unresolved conditions and implications before WEETRA issues any Determination. The responsibility is substantive. The reviewer decides what the examined record supports, including where it supports less than the file appears to claim.
The point is not the mechanics of any individual calculation. It is the responsibility attached to the conclusion. WEETRA remains accountable for the Determination it issues within the agreed mandate.
Reviews are usually produced well before the moment they need to withstand scrutiny. By then the people involved have moved on, systems have been replaced, and the working context has evaporated. What remains is the file.
A controlled review state should allow an authorised later reviewer to reconstruct, according to the Review Level and the mandate
A conclusion without a reconstructible basis is weaker than a conclusion whose evidentiary history can be followed. The second can be examined. the first can only be believed or doubted. This is why the reviewed state is preserved rather than summarised away.
Explore Documented Evidentiary Due Diligence →
WEETRA’s Determination complements these roles by documenting what the reviewed evidence supports within the agreed mandate. Accredited verification, where required, remains the independent statutory function of the verifier. Mandate & Boundaries →
This review approach is also translated into professional training and executive briefings. Professional training →
