Carbon claims that survive EU greenwashing scrutiny

The era of decorative carbon labels is ending in Europe. The Empowering Consumers for the Green Transition Directive is already in force. The Green Claims Directive is in trilogue and expected to land in late 2026. The Norwegian Consumer Authority's 2022 ruling against Higg MSI-based marketing claims set the enforcement template. Together they create a single, harder question for textile brands: can your published carbon number be defended as a substantiated claim, or is it decoration that will be ordered withdrawn?

This post is a clear-eyed read on what regulators expect, what the Higg ruling actually said, and how to frame a footprint disclosure today so that it survives the regime that's already operational.

The two directives that matter#

Two pieces of EU law shape the rules for environmental claims at product level.

Empowering Consumers Directive (2024/825)#

Already in force as of March 2024. Member State transposition deadline is March 2026; enforcement is operational across the EU now or imminently. The directive amends the existing Unfair Commercial Practices Directive to:

• Ban generic environmental claims without substantiation. "Eco-friendly," "green," "sustainable," "climate-friendly" as standalone terms are prohibited unless backed by demonstrated environmental performance.
• Ban claims based on carbon offset compensation for product-level claims like "carbon-neutral." Brands cannot label a product carbon-neutral on the basis of having purchased offsets equal to its emissions.
• Restrict sustainability labels to those that meet a defined standard or have third-party certification.
• Require future warranty information to be disclosed at the point of sale.

The directive applies to claims at the product level — exactly the level a DPP operates at.

Green Claims Directive (proposal 2023/0085)#

Still in trilogue at time of writing, expected adoption late 2026. The Green Claims Directive doesn't replace the Empowering Consumers Directive; it adds a substantiation regime on top:

• Every explicit environmental claim must be supported by a published assessment following recognised scientific methodology.
• The assessment must be third-party verified before the claim can be made (the burden of proof model — show the working, then make the claim).
• The methodology must use recognised life-cycle assessment standards. The Commission's preferred framing is the Product Environmental Footprint (PEF) method and its product-category-specific rules (PEFCRs).
• For textiles specifically, the relevant PEFCR is the Apparel and Footwear PEFCR, finalised in late 2024 and operational from 2025.

The combined effect: a brand making a product-level environmental claim in 2027 will need a substantiated assessment, methodology-disclosed, third-party-verified, before the claim is published.

What the Norwegian ruling against Higg actually said#

The 2022 Norwegian Consumer Authority (Forbrukertilsynet) action against Norrøna and other brands using Higg MSI scores in consumer-facing marketing is the most-cited enforcement precedent in this space. The ruling is often summarised in caricature; the actual finding is more specific.

The Authority did not rule that the Higg MSI methodology was wrong. The Authority ruled that using Higg MSI scores in consumer-facing marketing to imply environmental performance was misleading because:

• The Higg MSI is a fibre-level dataset, comparing raw materials. Consumers were being shown numbers framed as if they characterised the finished garment, which the methodology does not support.
• The dataset is based on industry-reported data without consistent third-party verification at the input level. The methodology disclaimer notes this; the consumer marketing did not.
• Many of the underlying data points were 2010s-era at the time of the 2022 ruling; the marketing did not surface methodology vintage.
• The framing in marketing routinely compared one fibre favourably to another in ways the underlying methodology does not statistically support at the precision implied.

The Authority's order was not that the data was bad; it was that the data was being used in a way the data didn't support, without the methodology disclosure that would have let consumers calibrate.

The wider implication for textile brands is what regulators now expect by default: methodology disclosure, vintage disclosure, scope disclosure, and framing that matches what the underlying data actually supports.

What "substantiated" means in practice#

Putting the directives and the enforcement precedent together, a substantiated product-level carbon claim in 2026 looks like this:

1. Named methodology#

The claim cites the methodology. "Cradle-to-gate carbon footprint estimate per the Apparel and Footwear PEFCR." Not "carbon footprint." Not "Higg-based estimate" as a marketing line. The named methodology is what makes the claim falsifiable.

2. Disclosed scope#

Cradle-to-gate (raw material extraction through factory exit) is the most common scope for textiles. Cradle-to-grave (including consumer use phase and end-of-life) requires assumptions about consumer behaviour that brands should not be making implicitly. The scope must be visible — typically as a one-line disclaimer near the number.

3. Disclosed data sources#

The factor sources — which emission factors were used for the fibre, the spinning, the dyeing, the freight — should be cited. Public-LCA datasets (ecoinvent, GaBi, JRC's Environmental Footprint reference packages) are acceptable when named. Proprietary aggregations are acceptable when the proprietor and the aggregation method are named.

4. Vintage#

When were the factors collected? Most public datasets carry a vintage. Mixing 2010-era cotton factors with 2024-era polyester factors is not inherently wrong but the dates need to be visible.

5. Estimate framing#

The number is an estimate. "Approximately 8.2 kg CO₂e" — not "is 8.2 kg CO₂e." Cradle-to-gate LCA of a textile garment cannot be measured to the precision implied by an absolute number; treating the calculation as inherently approximate is honest and protects against precision-of-claim misuse.

6. Brand-override disclosure#

If the brand uses its own measured value rather than the calculator's estimate (because the brand has primary data from its supply chain), the basis for the override should be visible. Otherwise the override looks like a marketing decision rather than a measurement decision.

What this looks like in practice on a label#

A carbon claim that meets the substantiation bar reads roughly as follows on a passport:

Estimated cradle-to-gate carbon footprint: ~8.2 kg CO₂e per garment. Methodology: Apparel and Footwear PEFCR, cradle-to-gate scope. Factor sources: public LCA datasets (ecoinvent 3.10, JRC EF 3.1). 2023–2024 vintage. Calculated by TextilePass; brand may override with measured supply-chain data.

That sentence — six lines — is the difference between a defensible claim and a marketing claim that gets withdrawn under enforcement.

What not to put on a label#

Patterns that have triggered enforcement in 2022–2025 and should be avoided:

• "Carbon-neutral product" on the basis of purchased offsets. Banned outright under the Empowering Consumers Directive.
• "Climate-positive" / "carbon-negative" without methodology. These terms imply a measurement that doesn't exist at product level for most textiles.
• A single absolute number with no scope or methodology. A bare "5.2 kg CO₂e" with no context is exactly the framing the Norwegian Authority called out.
• Comparative claims ("50% lower than industry average") without naming the industry average reference. The comparator needs to be a specific, citable benchmark — not a marketing assertion.
• "Sustainable" as a product property. Empowering Consumers Directive prohibition.
• Best-in-class certification labels that aren't independently certified. The directive restricts sustainability labels to certified or standardised schemes.

How TextilePass handles this today#

TextilePass calculates a cradle-to-gate carbon estimate per published product, using public-LCA factors from ecoinvent and the JRC EF reference packages. The implementation:

• Names the methodology in the consumer-facing methodology field on the passport.
• Discloses scope (cradle-to-gate) inline near the number.
• Names the factor sources — visible on the methodology drawer on the passport, not buried in a footer.
• Discloses vintage of the factor library.
• Frames the number as an estimate ("Estimated cradle-to-gate footprint") consistently across the consumer surface, the passport JSON-LD response, and the PDF export.
• Surfaces the brand-override mechanism. Brands with their own measured supply-chain data can override the calculator's value; the override is recorded with the basis, the date, and the responsible user. Consumer surfaces show that the value is brand-measured rather than calculated.

The factor library is documented; the methodology language was revised in May 2026 explicitly to drop the Higg MSI brand-name framing and re-anchor on public-LCA datasets, in response to the regulatory landscape this post describes.

For the broader regulatory picture, see the ESPR compliance post and the DPP guide. For a plain-English walk-through of proving a single claim, see how to prove a "sustainable" clothing claim under the EU's 2026 rules. To see how a substantiated carbon claim renders on a real passport, open the live demo or start free on the Starter plan.

A defensible footprint claim isn't more expensive than a misleading one. It just requires the methodology to be visible. The brands that will look credible in 2027 are the ones that started writing the methodology line into the label in 2026.