Europe Banned Destroying Unsold Clothes in July. The Test We Proposed Can't Be Run.
In our piece on Europe’s textile collection mandate we listed the unsold goods destruction ban as a thing to watch, and proposed a test:
Watch whether deadstock volumes arriving in Accra fall, because that is the cleanest available test of whether the ban works.
The ban took effect for large companies on 19 July 2026. It is now three months old. This is what we found when we tried to run our own test, and the answer is that we cannot.
What actually came into force
The Ecodesign for Sustainable Products Regulation prohibits the destruction of unsold consumer products listed in its Annex VII. The initial list is limited to textiles and footwear — apparel, clothing accessories and shoes. Article 25 lets the Commission extend it to other product groups by delegated act.
It binds large companies from 19 July 2026. “Large” is the standard EU test: exceeding at least two of more than 250 employees, more than €50 million turnover, or more than €25 million in total assets. Medium-sized companies follow on 19 July 2030. Micro and small enterprises are exempt outright.
Destruction remains permitted in narrow cases set out in Delegated Regulation (EU) 2026/296: products that are dangerous, products that are non-compliant and unfit for purpose, and goods proven to infringe intellectual property. Companies need documentation to justify an exception.
One provision worth noticing
Shredding unsold garments purely to recover fibres counts as destruction and is prohibited.
That is a sharper rule than it first appears, and it closes the obvious escape. Without it, a brand holding unsold stock could route it to a fibre-recovery process and book the result as recycling rather than destruction — which, given that fibre-to-fibre recycling remains below 1%, would have been a disposal route wearing a circularity label.
The test we proposed, and why it fails
Our test assumed somebody is counting what arrives in Accra. Nobody is, at least not in a way that would detect a change.
The most-cited volume figure for Kantamanto remains around 15 million garments a week. It is attributed to the Or Foundation and dates to roughly 2024. We could find no newer independent measurement, and no 2026 figure at all.
That matters more than staleness, because of something our original piece already said. We flagged that the Or Foundation’s widely-repeated 40% waste figure is disputed by an order of magnitude by the Ghana Used Clothing Dealers Association, and that all parties to that dispute have an obvious stake in the answer. Using the same source’s volume figure as the baseline for measuring a European regulation inherits exactly that problem. We proposed a test whose instrument we had, in the same article, described as contested.
That is our error, not the regulation’s.
The disclosure that would settle it arrives in 2027
There is a measurement coming, and it is not in Ghana.
Large companies have been subject to disclosure obligations on unsold goods since 2025, but a standardised reporting format applies only from 2 March 2027. Until then, what firms publish is not comparable between them, which means it cannot be aggregated into a number that shows whether destruction fell.
So the sequence is: the ban binds from July 2026, and the data that would show whether it changed anything becomes comparable from March 2027. Eight months in which the obligation exists and its effect is unobservable.
Who is not covered
The ban reaches large companies now and medium-sized ones in 2030. Micro and small enterprises are permanently exempt.
Any volume measured in Accra is therefore the sum of covered firms, uncovered firms, non-EU exporters and genuine second-hand clothing. A fall would not isolate the ban’s effect, and a flat line would not disprove it. Even with perfect import data, the test we named could not have attributed causation.
What to watch instead
Three things, in order of how soon they can tell you anything:
- Exception filings. Companies must document why a destruction falls under Delegated Regulation (EU) 2026/296. If “non-compliant and unfit for purpose” becomes a common justification, that is the loophole opening, and it is visible before any volume data.
- The March 2027 standardised disclosures. The first comparable set is the first real evidence. The question to ask of it is not how much was destroyed but how much was declared — a sharp year-one fall in reported destruction may be reclassification rather than behaviour.
- Whether Annex VII is extended. Article 25 allows other product groups. Electronics would be the significant one.
The uncomfortable part
We named a test that cannot be run, using an instrument we had already described as contested, to measure a regulation whose own evidence base does not arrive until eight months after it binds.
The regulation is better drafted than our test was. Banning fibre-shredding as destruction shows someone thought about the obvious workaround. The staged thresholds are defensible. The weakness is not the rule; it is that compliance becomes observable long after the obligation starts, and that the downstream effect everyone cares about is measured by interested parties or not at all.
This is the pattern from the construction target and America’s missing e-waste number in a third configuration. There the measurement counted the wrong thing, or stopped existing. Here it exists, is well specified, and simply has not started yet.
Sources: Regulation (EU) 2024/1781 (ESPR) Article 25 and Annex VII for the destruction prohibition and its product scope; Delegated Regulation (EU) 2026/296 for the permitted exceptions; published compliance analyses for the 19 July 2026 and 19 July 2030 application dates, the large-company thresholds, and the 2 March 2027 standardised reporting format, where secondary sources differ on the thresholds and we have used the majority reading. The Kantamanto volume figure is the Or Foundation’s and is approximately two years old; its limitations are discussed above rather than in a footnote.