What an environmental figure on a passport is actually counting

Nothing in EU law makes you publish one today

A negative finding, stated with what was checked, because the reverse is asserted constantly.

The first row is worth one more sentence, because the power it describes is specific rather than general. Annex I of the framework lists the parameters a requirement may be built on, and three of them are exactly this subject: the environmental footprint of the product, expressed as a quantification of life cycle environmental impacts in one or more impact categories or an aggregated set of them; the carbon footprint of the product; and the material footprint of the product. Article 7 then allows an information requirement to cover performance in relation to those parameters, and names a carbon footprint and an environmental footprint among its examples. So the instrument that would require a figure is drafted and waiting. What is missing is the delegated act that would use it, and for textiles that act does not exist. Read the two together and the position is clearer than either alone: the figure is not required, the power to require it is explicit, and the shape it would take is already described.

So a brand publishing a carbon figure today is making a voluntary statement about its product. That is a reasonable thing to do, and it changes the kind of scrutiny the statement attracts, because a voluntary claim is judged as a claim rather than as a compliance field.

What is genuinely coming, and the reason to understand this before it does, is that the preparatory work behind a textile act proposes an environmental indicator with a very specific shape.

Four numbers, four different questions

A brand's environmental data is usually four unrelated things stored in one place. They fail in different ways and only one is testable on the object.

The water row is the cleanest example of a word doing two jobs. The international standard on water footprint assessment is a standard about an assessment rather than about a quantity. Its own catalogue entry states that the result of a water footprint assessment is a single value or a profile of impact indicator results, and the document adds that where the impacts have not been comprehensively assessed, the term may be used only with a qualifier.

A litres figure is an inventory quantity. It is a real number about real water and it is not a water footprint in the sense the standard defines. The two are routinely printed under the same heading, and a scarcity-weighted result and a volume are not comparable in either direction.

The same catalogue entry carries a second fact worth having, because it repeats a pattern. Communication of water footprint results, in the form of labels or declarations, is outside that standard's scope. Communication of a product carbon footprint is likewise outside the scope of the standard governing how one is quantified, which moved that guidance to a separate document. Twice, the standards that tell you how to produce these numbers decline to tell you how to say them. The rules that do govern saying them are consumer protection rules rather than technical ones, which is why the last third of this page is about wording.

The boundary is most of the number

If one idea from this page survives to tomorrow, it should be this one. Two brands can produce carbon figures for the same garment that differ by an order of magnitude without either doing anything wrong, because the figures answer different questions. The boundary is the question.

The preparatory work informing the textile decision makes this concrete. Its proposal for the footprint indicator is that the rules in the category rules for apparel and footwear be followed as applicable to lifecycle stage 2 only, the manufacturing stage. The reason given is not a policy preference. It is that the datasets available for the several types of fibres use inequivalent system boundaries, which prevents their fair comparison.

Read that twice, because it does two things at once. It proposes a carbon figure that excludes the production of the fibre the garment is made of. And the reason given for excluding it is that the underlying fibre data is not comparable, which is a statement about the state of the evidence rather than about the importance of the stage.

Three cautions belong in the same breath. That proposal sits in preparatory work, which is pre-decisional by design and is not law. It was read at the published summary of a preparatory milestone, and the later study that would carry the current proposal truncated before its proposal sections on every route this estate attempted, so nothing here is attributed to that document. And a preparatory study exists to be argued with: the act it informs may take it up, vary it or set it aside.

What follows for a business is a habit rather than a decision. Whenever a figure is produced, record the boundary beside it, in the same field, in words a stranger could act on: which stages, which of the garment's components, and which version of which method. A figure whose boundary is not recorded cannot be compared with next season's figure, cannot be defended when somebody asks, and cannot be corrected, because nobody can tell what it was answering.

Company-specific data, secondary data and a default are three different things

After the boundary, the thing that decides a number is where its inputs came from. Three tiers are routinely flattened into one word, and the flattening is what turns a modelled figure into something a reader takes for a measurement.

Company-specific data is measured or collected from a facility. Secondary data is a dataset standing in for a process nobody measured for this product. A default value is a number the method itself supplies so that a calculation can complete at all.

The same preparatory work is explicit that the approach it proposes allows the use of secondary datasets and default values to facilitate the calculation in the absence of primary data. That is a sensible design decision for a rule that has to work for every apparel business in Europe. It also means a compliant figure and a measured figure are not the same object, and that two products with identical published figures may rest on entirely different amounts of actual knowledge.

The estate's own handling follows the same logic. Modelled figures on a worked passport carry the word modelled beside the state, name the calculator and pin the version it came from, because the answer moves when the version does. Where a required input is missing the whole figure is withheld rather than computed without it, since a sum with a term dropped out of it gets smaller, stays publishable and is wrong.

There is a fourth state that is not a data tier at all, and it is the one worth being strictest about. A figure can be absent because nobody has calculated it. What a business may publish in that position, and the two named instruments that already contain worked patterns for a bounded unknown, is set out on what you are allowed to put when you do not know.

The one score the method will not let you show a consumer

This is the sharpest finding on the page and it runs against what almost everybody assumes.

The category rules for apparel and footwear state that the single score is not allowed for business-to-consumer communication, nor for comparison against the representative products. That is the document's own sentence, at version 3.1, dated 29 April 2025 and stated as valid to the end of 2027, and it was returned identically on three separate reads of the file. The single score is the aggregated figure produced by weighting the sixteen impact categories together. It is also, for obvious reasons, the number a brand most wants to put on a swing tag, because it is the only one that fits.

Comparison itself is not forbidden. The method above the category rules is explicit about when it is allowed, in a sentence in its annex on category rules: comparisons and comparative assertions are allowed only if studies are conducted in compliance with category rules. The category rules for apparel and footwear say the same thing from the other side, in their review statement: studies carried out in compliance with them may be used to make comparisons and comparative assertions under prescribed conditions, and they name the section that sets those conditions out. This estate could not read that section and therefore describes none of its conditions.

So the position that can be stated is narrower than either the permissive or the prohibitive version in circulation. Comparison is available where the studies followed the category rules and met conditions set out in a section we have not read. The aggregated single score is not available for consumer communication in this category at all.

Two consequences follow immediately.

A passport built as a consumer-facing surface and populated with an aggregated environmental score is, for apparel, populated with the one figure that method says must not go there. Our view is that this is a design problem rather than a drafting accident. The method was written to govern comparisons between assessments, and the passport is being built as a display.

And there is a contrast nobody in this category seems to want to name. France's environmental cost scheme, which the government's own information page describes as voluntary to enter and which prescribes the method once entered, is built entirely around displaying a single aggregated figure in impact points to a consumer. The European category rules for the same garments forbid the single score in exactly that setting. Both are real, both are current, and they point in opposite directions on the same question for the same product.

We are not saying either is wrong. We are saying that a brand asking whether it may show a consumer one number has two answers available depending on which document it reads, and that anybody selling a European single-score display for apparel should be asked which of the two they are working from.

Two properties of the sixteen categories are worth knowing before anybody treats the aggregate as complete. There is no biodiversity impact category among them, and there is no microplastics category. Effects of either kind reach the score indirectly, through categories built for something else, or not at all.

From 27 September 2026 the sentence is tested, not the model

A separate instrument decides what may be written beside a number, and it starts applying through national measures on that date. Four described practices move onto the list of commercial practices treated as unlawful in themselves, which removes the question of whether an average consumer was actually misled. Which instruments reach a clothing brand and when is maintained on what already applies today, and this page carries the date only because this section is about it.

Two of the four decide most environmental-figure cases before the evidence is opened.

A claim about the whole product resting on one aspect of it is caught. A garment described as low impact on the strength of a manufacturing-stage figure is a claim about the whole product resting on part of it, and the boundary section above is exactly what makes that visible.

A generic environmental claim is caught where the trader cannot demonstrate recognised excellent environmental performance. The definition of generic is the commercially useful part, and it is defined by what is missing rather than by which words were chosen. A generic environmental claim, at point (p) of the amended definitions article, is an environmental claim not included on a sustainability label and where the specification of the claim is not provided in clear and prominent terms on the same medium.

Turn that round and it stops being a prohibition and becomes an instruction. The same words carrying their specification, in the same place, are not a generic claim.

A code linking to a page is not the same medium. That matters more here than anywhere else in this subject, because a passport is reached through a code, and a specification living at the far end of one has not travelled with the claim.

The gateway to recognised excellent environmental performance, at point (s) of the same article, has three limbs and only two of them are certification schemes: performance meeting the EU Ecolabel regulation, or meeting national or regional type one ecolabelling schemes officially recognised in the Member States, or top environmental performance in accordance with other applicable Union law. The three run as continuous prose joined by "or" rather than as numbered sub-points. The third does not require a certification scheme of any kind, which is why the position here is less absolute than the certification-or-silence version that circulates in summaries.

One term that is not in this directive is worth naming, because assuming it is there is the most common way to misread the definitions. There is no definition of an explicit environmental claim in it. That concept belongs to the separate proposal on substantiating environmental claims, which is blocked and is not law, and importing it produces a reading of the definitions article that does not match the text.

There is also a housekeeping item that will catch a number of businesses quietly. The international standard on self-declared environmental claims had its second edition withdrawn on 24 June 2026 and replaced the same day by a third edition, published under a changed title. A claims policy, a supplier contract or a technical pack citing the withdrawn edition now cites a document that no longer exists, and nothing in the ordinary running of a business would surface it. We have not bought the new edition and nothing here describes what it contains.

Five questions that decide what a figure will support

In the order they can be answered. Four can be settled from the calculation record without asking anybody.

What boundary produced it. Which lifecycle stages, which components, and whether the garment's fibre production is inside or outside. Without this the number cannot be compared with anything, including with itself next season.

Which method, at which version. A method is revised and the same inputs then produce a different answer. A figure without a version is a figure nobody can reproduce.

Which inputs were measured, which were secondary and which were defaults. Three tiers, recorded per input rather than as an overall grade, because an overall grade hides the one input that was a default and that input is usually the one somebody will ask about.

What it is a figure of. A garment, a kilogram of fabric, a model, a batch, or a use. The functional unit is where a great deal of quiet error lives, and a per-use figure and a per-garment figure differ by a lifetime assumption that somebody chose.

Whether the sentence you want to write reaches further than the figure. The only one that cannot be answered from the calculation, because it is a comparison between two objects: the exact words and the exact scope. It is also the one that decides cases.

A figure surviving all five supports a statement about a modelled result, under a named method, at a stated boundary. It does not support a statement that the garment is better than another garment unless the conditions above are met, and it does not support a statement about the whole product.

What to do with the numbers you already have

None of this argues for publishing nothing. It argues for publishing the number with the thing that makes it interpretable, which is cheaper than most of the alternatives.

Record the boundary in the same field as the figure, not in a methodology document that travels separately. The field is what survives an export, a channel feed and a change of system.

Pin the method version and the dataset version. Both move. A figure that cannot name them cannot be defended and cannot be reproduced, which are the same problem seen from two sides.

Keep the tier per input rather than a grade for the whole figure.

Publish the calculator's own answer rather than your arithmetic on top of it. Where a public method exists, use the public one and say whose it is, so anybody can reproduce the number rather than take yours. Where your own figure disagrees with the calculator's, the honest output is the calculator's undivided figure and a recorded disagreement, not a compromise between them.

Read the sentence against the scope before funding more evidence. The comparison costs an afternoon. It is the control with the highest return in this whole subject and it is not a software feature.

Fix the composition and the weight first. The modelled figures in this category are built from the composition and expressed against the mass. A wrong weight produces four wrong figures that still render beautifully. Where those two values live in an ordinary catalogue, and why the weight field usually holds the wrong kind of mass, is on where product data stops meaning the same thing.

What would change this page

A textile act adopting an environmental information requirement, at which point the proposal described here stops being a proposal and this page is rewritten around what was adopted.

The category rules read at their publisher's own host rather than at the third-party copy this page rests on, which would let the communication rule be quoted at full strength and would let the conditions on comparison be described rather than named.

A published position from any authority on whether an aggregated score may be shown to a consumer through a passport specifically, which would settle the contrast in the middle of this page rather than leaving it standing.

A method that makes fibre production datasets comparable, which is the stated obstacle behind the manufacturing-stage-only proposal and the thing that would most change what a textile carbon figure means.

Sources

• Ecodesign for Sustainable Products Regulation (EU) 2024/1781, consolidated text
  In forceRelevant provisions reviewed

  CELEX 32024R1781. In force. Read in full at the Official Journal HTML rendering, 28 August 2026, including Annex I. Read for the proposition that the framework creates the passport and the power to set product requirements without imposing any environmental figure on textiles, and now also for Annex I points (m), (n) and (o), which are the environmental footprint, carbon footprint and material footprint parameters, and for Article 7, which allows an information requirement to cover performance against those parameters and names a carbon footprint and an environmental footprint among its examples. The annex establishes that the power exists and describes its subject matter. It establishes no duty. A delegated act is what turns a parameter into a requirement, and none exists for textiles.

  View official source

• Product Environmental Footprint Category Rules for Apparel and Footwear, version 3.1
  Voluntary schemeRead in full, in a copy hosted by somebody other than its publisher

  Final, 29 April 2025, stated as valid to 31 December 2027. Reached through a third-party copy of the document, not read at the publisher's own host. The publisher's host refused automated retrieval on three attempts, and its published questions page is several versions behind the document, still describing the rules as work in progress. Two propositions rest on the copy that was read: that the single score is not allowed for business-to-consumer communication nor for comparison against the representative products, and that the review statement records that studies carried out in compliance with the rules may be used for comparisons and comparative assertions under prescribed conditions set out in a named section. The first was returned identically on three separate reads. That named section is beyond the extraction limit on this copy, so no condition from it is described here. The data needs matrix and the data quality thresholds were not obtained and no figure from either appears here.

• Commission Recommendation (EU) 2021/2279 on the use of Environmental Footprint methods
  Institutional guidanceRelevant provisions reviewed

  CELEX 32021H2279. Recommendation. Annex I read at a Commission-hosted copy. Two propositions rest on it. The list of sixteen impact categories, and the absence of a biodiversity category and a microplastics category among them. And the sentence, verbatim, that comparisons and comparative assertions are allowed only if studies are conducted in compliance with category rules, which sits in Annex I in the section on category rules rather than in the numbered points of the recommendation.

• Joint Research Centre, textile preparatory study, third milestone, published summary, 12 December 2025
  Pre-decisional researchPublisher's own summary read, the document itself not

  Pre-decisional. Published summary read at source; the full study and the later content study were not read. Four propositions are taken from it: that the footprint indicator is proposed following the apparel and footwear category rules as applicable to lifecycle stage 2 only, the manufacturing stage; that the stated reason is inequivalent system boundaries in the datasets available for the several types of fibres, preventing their fair comparison; that secondary datasets and default values are permitted in the absence of primary data; and that the reporting proposed is voluntary, with an excellence distinction available only where a manufacturer demonstrates performance superior to the category benchmark. The later study on passport content, dated 13 May 2026, truncated before its proposal sections on every route attempted, so no proposal is attributed to it here.

• ISO 14046, environmental management, water footprint, principles, requirements and guidelines
  International StandardPublished abstract read, full text behind a purchase

  First edition, published August 2014, recorded as published and confirmed. Catalogue entry read at the publisher's own catalogue. Published preview read. Full text behind a purchase and not bought. From the catalogue entry: that the result of a water footprint assessment is a single value or a profile of impact indicator results, and that communication of those results as labels or declarations is outside the standard's scope. From the published preview only: that where impacts have not been comprehensively assessed the term may be used with a qualifier. The catalogue entry does not carry that second sentence, so it is stated at the weaker of the two depths.

• ISO 14067, greenhouse gases, carbon footprint of products
  International StandardBehind a purchase, not bought, not read

  Published standard. Catalogue entry and published preview read, full text behind a purchase. Named for one structural fact rather than for any requirement: the current edition covers quantification, and guidance on communicating a product carbon footprint sits in a different standard.

• ISO 14021, self-declared environmental claims, edition withdrawn 24 June 2026
  International StandardOfficial source confirmed, detailed review pending

  Second edition withdrawn 24 June 2026. Third edition published the same day under a changed title. Catalogue entries read; neither edition bought and neither read. Used for a fact about a catalogue rather than about a text. A document citing the withdrawn edition now cites an edition that no longer exists. Nothing here describes what the replacing edition contains.

• Directive (EU) 2024/825 on empowering consumers for the green transition
  Adopted, not yet applicableRelevant provisions reviewed

  CELEX 32024L0825. In force, applying from 27 September 2026. Relevant provisions reviewed at the Official Journal in two renderings, with the point lettering independently checked against the Commission's own questions and answers. Read for the application provision, for the described practices this page uses, for the definition of a generic environmental claim at point (p) and for the definition of recognised excellent environmental performance at point (s). Both definitions are quoted at their point letters, which Phase 1 of this workstream held as an open check and which is now closed: four passes across three documents agree on the sequence and there is no lettering gap. The directive does not define an explicit environmental claim, and this page says so, because assuming it does is what produced the earlier confusion.

• Order of 6 September 2025 on the signage and methodology for calculating the environmental cost of textile clothing products, and the official methodological notice published alongside it
  Government modelRead in a public body's copy rather than in the official gazette

  Government model and official notice. Notice read at its own address; the government information page read; the decree and order behind them refused automated retrieval and were not read. Behind two propositions: that the government's own information page describes display under the scheme as a voluntary approach while the method is prescribed for those who display, and that the figure displayed is a single aggregated value in impact points.