From enabler to gatekeeper: digital product passports and the blind spot in circular trade governance

1 Introduction

When the World Trade Organization (WTO) Committee on Technical Barriers to Trade (TBT) met in July 2026, its agenda included a cluster of concerns raised by members about digital product passports (DPP): the scope of the requirements, the time available to comply, and the recognition of equivalent regulations and of conformity assessment carried out abroad (WTO, 2026a,b). An instrument conceived as an enabler of circular trade had, before its principal rules were even in force, become the object of formal trade concerns. That juxtaposition is the starting point of this article.

The passport is the centerpiece of the Ecodesign for Sustainable Products Regulation (ESPR), the framework instrument of the EU's circular economy agenda, which replaced the earlier Ecodesign Directive and extended its logic from energy-related products to almost all goods placed on the Union market (Regulation (EU), 2024/1781). It will carry the material and lifecycle data on which reuse, repair and recycling decisions depend, and textiles are the first sector in line. The first content specification for textile apparel, issued by the Joint Research Center (JRC) in May 2026, sets out 49 data points across a tiered reporting structure, ahead of a delegated act expected around 2027 (González-Torres and Arcipowska, 2026); the revised Waste Framework Directive, in force since October 2025, has meanwhile placed extended producer responsibility for textiles on the same trajectory (Directive (EU), 2025/1892). Because the obligation attaches to placing a product on the Union market, it reaches every supplier that sells into the Union, wherever located—a reach that commentators have already read as a textbook instance of the “Brussels effect,” the EU's capacity to set de facto global standards through market access rather than negotiation (Bradford, 2020), though the reach of that effect is itself contested, since on some accounts the stringency of EU rules makes it a preference outlier with limited pull over international standards (Young, 2015).

The circular economy seeks to decouple growth from resource use by closing material loops, and international trade determines whether those loops can close across borders or remain confined within them (Yamaguchi, 2021). Reliable information about what a product contains is a genuine precondition for circularity: in textiles, a sector whose globalized structure has steadily eroded supply-chain transparency, traceability is widely identified as a prerequisite for any credible circular transition (Das et al., 2025). The case for the passport, in these terms, is strong. Yet the scholarship on passports has developed largely along technical and managerial lines, data architecture, adoption barriers, standardization, while giving comparatively little attention to their legal and institutional consequences, and in particular to how they intersect with the rules of international trade and whether their design might entrench asymmetries among the actors caught up in them (Yuan, 2025). Where the external effects of the ESPR have been examined, the analysis has tended to remain either at the level of general trade-law compatibility, such as the treatment of “like products” and eco-labeling under WTO rules (Blot, 2025), or at the level of firm-level readiness (Farrukh and Sajjad, 2024). The mechanism connecting the two, how a market-access condition is actually transmitted to the producers who must satisfy it, has not been articulated.

Addressing that omission is the contribution of this article, and the registers on which it draws are worth separating at the outset. That the passport obligation attaches to the placing of a product on the Union market, and falls on the economic operator who does so, follows from Regulation (EU) 2024/1781. That the textile specification is at this stage a technical recommendation rather than a binding requirement is likewise a matter of record. That comparable data obligations have in practice traveled downstream to suppliers is an empirical observation drawn from adjacent regimes (Sela et al., 2025). What is advanced here as original is not the doctrinal point that the TBT disciplines bind measures attributable to Members rather than private instruments, which is a settled limitation. It is the application: tracing a specific transmission mechanism in this regulatory setting, public obligation, contractual condition, upstream capacity constraint, possible exclusion, and showing that the safeguards designed to cushion it do not reach it, because ESPR Article 22 is territorially bounded and the cascading assurances of the Corporate Sustainability Due Diligence Directive (CSDDD) run along contractual privity that stops short of the tiers where the data originates.

By blind spot is meant a categorization gap rather than a gap in enforcement: the exclusion is real and the disciplines are intact, but the instruments producing it, purchase orders, onboarding conditions, audit criteria, are attributable to no Member and answer to none of the definitions on which the disciplines of the Agreement on Technical Barriers to Trade operate. The argument is not that the passport is unlawful, nor that a legal vacuum exists.

The article develops this claim in three steps: it characterizes the passport as a market-access condition, traces the contractual cascade through which that condition is transmitted, and locates the blind spot this creates, before turning to the design choices through which a passport might enable circularity without excluding the producers on whom it depends. In doing so it extends a broader argument that trade policy, reformed with intent, can turn from a barrier into an enabler of circular value chains (Bato and Kaštáková, 2025).

2 The digital product passport as a condition of market access

In its own terms the passport is an enabler of circularity: the data carrier attached to the product is to make material composition and lifecycle attributes legible to the recyclers, repairers, regulators and consumers who handle it after the point of sale [Regulation (EU), 2024/1781]. There is little reason to doubt that this is its purpose.

Its legal form, however, is that of a market-access condition. The obligation attaches to the placing of a product on the Union market, so that a product without a compliant passport may not lawfully be sold, and it applies irrespective of where the product was made [Regulation (EU), 2024/1781]. This is a shift in kind rather than degree. Traceability requirements have historically operated as post-market obligations, verified through surveillance once goods are already in circulation; the passport moves verification upstream, into a pre-market digital condition that must be satisfied before access is granted. For the large majority of textiles consumed in the Union that are imported, the condition falls in the first instance on producers established outside it.

The Commission does not treat this consequence as incidental. It has acknowledged that the requirements will have significant implications for trading partners particularly developing ones (European Commission, 2025), since all goods placed on the Union market, imports included, must comply, and it has committed to assessing the effects on third countries and to attending to the needs of smaller operators in the design of specific requirements. These are meaningful commitments, and the argument that follows does not assume their absence.

It is therefore worth stating plainly what is, and what is not, contested here. The legitimacy of the environmental objective is not in question, nor is the proposition that circularity depends on reliable product information: a passport that no one can populate serves no one. What is at issue is narrower and concerns the instrument's design, the granularity of what must be reported, the sequence in which obligations take effect, and the distribution of the resulting cost across a value chain whose upstream participants had no part in setting the requirement. It is to that mechanism of distribution that the article now turns.

3 The regulatory cascade: from public obligation to private condition

The ESPR allocates the DPP obligation to the economic operator placing a product on the Union market—the manufacturer, the importer, or an authorized representative established in the EU [Regulation (EU), 2024/1781]. The allocation is unambiguous, and in formal terms it falls entirely within the Union's regulatory jurisdiction. The information the passport must carry, however, does not originate there. This is a different problem from the one posed by overlapping traceability regimes across jurisdictions, where the difficulty lies in legal conflict between systems of broadly comparable regulatory capacity (Huang and Nottage, 2026); here the transmission occurs within a single value chain, between parties of markedly unequal capacity. Figure 1 sets out the cascade and the point at which it crosses out of the reach of the trade disciplines.

The first complete content specification for textile apparel, published by the JRC as a recommendation that the delegated act may yet revise, sets out 49 data points across four categories: product identification, producer identification, product information, and compliance documentation (González-Torres and Arcipowska, 2026). Several can be supplied by the brand or importer. Most cannot. Fiber composition, recycled content and the recycling route by which it was obtained, substances of concern reported by name and concentration, facility-level identifiers, and carbon and environmental footprint data are all generated two or three tiers upstream. The JRC proposes a tiered approach to granularity: model-level reporting as the baseline for attributes that do not vary across a production run, batch-level for parameters subject to manufacturing variability, and item-level as a voluntary option. The tiering is a proportionality safeguard, and it matters. But the batch-level subset, facility identifiers, chain-of-custody certification for recycled content, and chemical usage, is composed almost entirely of data generated two or three tiers upstream, in spinning, dyeing and finishing operations and among chemical input suppliers with which the EU operator typically holds no contractual relationship. The graduation is by data type, not by the capacity of the actor who must produce it.

This creates a structural gap between where the obligation is placed and where the compliance capability resides. The gap is bridged not by regulation but by contract—and the recent architecture of EU value-chain law takes that bridging for granted. The CSDDD requires the in-scope company to obtain contractual assurances of compliance from its business partners, and to seek corresponding assurances cascading further up the chain of activities where the partner is unable to give them directly (Directive (EU), 2024/1760). Against that background, Union brands and retailers may be expected to make passport-ready data a condition of supplier onboarding and of contract renewal, so that suppliers unable to furnish verifiable product data would face delisting, failed buyer audits, and exclusion from long-term sourcing programmes (Gereffi et al., 2005). That the burden travels this way is already observable in an adjacent regime: under the EU Deforestation Regulation [Regulation (EU), 2023/1115], the data needed to demonstrate compliance has fallen largely on suppliers, many of them in economies where small and medium-sized firms predominate (Sela et al., 2025).

Three consequences follow, and they are analytically distinct from the familiar observation that regulation imposes compliance costs. First, the incidence of the measure diverges from its formal allocation: the obligation is European, the expenditure is largely extra-European, and assessments of regulatory burden that examine the duty-holder will not detect the displacement. Second, exclusion operates on a different timetable from the legislation. The textile delegated act is expected around 2027, with a transition period before application; contractual conditions, by contrast, are imposed as soon as buyers begin de-risking their own future compliance. A supplier may therefore lose market access well before the measure it failed to anticipate has entered into force. Third, no state decision is taken at the moment of exclusion—a purchase order is not a technical regulation.

These effects matter for circularity, and not only for equity. A passport is worth precisely as much as the data it carries: recyclability scores, recycled content claims and substance declarations are the inputs on which downstream sorting, reuse and recycling decisions depend. Where the reporting burden is transmitted to actors lacking the metering, laboratory and record-keeping infrastructure to generate such data, the response to be expected is not withdrawal from the market but approximation, declarations assembled from supplier assurances rather than from measurement. A passport populated in this way would satisfy the compliance requirement while degrading the informational asset on which circular decision-making rests. There is a second and sharper irony. The passport architecture is organized around the first placement of new goods on the market, whereas reused components, remanufactured parts and second-hand garments generally re-enter trade without one. Absent deliberate design choices, circular goods risk being documentarily disadvantaged relative to virgin ones, reproducing in digital form the very asymmetry that already burdens reclaimed construction materials at the border (Bato and Kaštáková, 2025). The cascade, in short, transmits the effect of a public measure through instruments that the disciplines governing that measure were never designed to reach, while quietly eroding the data quality that justified the measure in the first place.

4 A blind spot in trade law

The argument of this article is not that the DPP violates WTO law. It is that the measure may well be consistent with it, and that this consistency is precisely what should give pause.

The relevant disciplines are those of the Agreement on Technical Barriers to Trade (WTO, 1995). A threshold point should be made explicit, since the analysis assumes it. The passport is capable of qualifying as a technical regulation within the meaning of Annex 1.1 of the TBT Agreement: it applies to an identifiable product, it lays down product characteristics and related administrative provisions, and compliance is mandatory—the three elements the Appellate Body drew from that definition in EC—Asbestos (WTO, 2001) and EC—Sardines (WTO, 2002). The definition expressly extends to marking and labeling requirements, which the data carrier affixed to the product supplies. On that footing the disciplines of Article 2 are engaged.

Article 2.2 requires that technical regulations be no more trade-restrictive than necessary to fulfill a legitimate objective. Environmental protection and resource efficiency are unambiguously legitimate, and the passport's contribution to them is plausible; the contestable element is therefore narrower than the specification as a whole. Where granularity is set at model level, the reporting burden is largely absorbed at the design stage. The necessity question attaches to the batch-level subset, and specifically to whether chain-of-custody certification and per-batch chemical reporting contribute to the circularity objective by a margin that justifies pushing verification into upstream operations whose readiness the JRC's own assessment records as low. Necessity analysis under Article 2.2 is capable of engaging that question; it has not yet been asked. That analysis is relational: it weighs the measure's contribution to its objective against its trade-restrictiveness and against the gravity of the consequences of non-fulfillment and asks whether a reasonably available alternative would make an equivalent contribution while restricting trade less (US—Tuna II, WTO, 2012; EC—Seal Products, WTO, 2014). Whether a less trade-restrictive alternative is reasonably available is the question on which such an analysis would turn, and several candidates present themselves: a compact core of data points with the more demanding elements phased in, model-level rather than batch-level granularity, longer transition periods differentiated under Article 12.3, and recognition of equivalent third-country schemes and of conformity assessment carried out abroad under Article 6.1. Each would have to be shown to contribute to the objective to an equivalent degree; that demonstration has not been attempted, but neither has its impossibility.

Article 2.4 raises a further issue that is analytically open. Members are required to use relevant international standards as a basis for technical regulations, and Article 2.5 attaches a rebuttable presumption of consistency to regulations that do so. The technical architecture of the passport is being fixed through CEN/CENELEC deliverables, whose membership comprises European national standardization bodies (European Commission, 2024). Whether such instruments qualify as international standards within the meaning of Annex 1 and the Committee's Decision on Principles for the Development of International Standards, which sets out openness, impartiality, and a development dimension among its criteria, is genuinely contestable, and the answer determines whether the presumption applies at all. The test is that of Annex 1.2 as read in EC—Sardines: a standard is international where the body that adopts it is open to the relevant bodies of at least all Members, a threshold European standardization deliverable do not meet. Bodies that might: ISO and IEC, UN/CEFACT, and—for the data-carrier layer—GS1; ISO/TC 323 is already developing circularity data instruments—ISO 59040 (ISO, 2025). A model routed through them, unlike a CEN/CENELEC one, could in principle attract the Article 2.5 presumption.

Article 2.12, requiring a reasonable interval between publication and entry into force so that producers in exporting Members may adapt, is where the divergence between formal and effective compliance time becomes visible. An eighteen-month transition, the interval Article 34 of the ESPR sets as the default before ecodesign requirements apply, is adequate for a European brand, for which the passport is a change of format. It is a different proposition for a supplier that must establish measurement, laboratory and record-keeping capability across three jurisdictions and several tiers.

None of these provisions, however, reaches the mechanism described in the preceding section. Article 2 disciplines the technical regulation, Article 5 the conformity assessment carried out by central government bodies, and Articles 11 and 12 the accompanying obligations of assistance and transparency—all of them instruments adopted by a Member and notifiable as such. A purchase order specifying data deliverables is neither. The cascade therefore falls outside the field of view of the disciplines that govern the measure generating it, and outside the reach of the remedies those disciplines provide. The categorization gap identified at the outset is therefore not a failure of enforcement: the exclusion is real and the disciplines are intact, but the instrument producing it answers to none of the definitions on which they operate. It might be objected that the cascade is already caught, since Article 3.1 requires Members to take such reasonable measures as may be available to them to secure compliance by non-governmental bodies with Article 2. The objection does not hold. A non-governmental body is defined in Annex 1 as one with power to enforce a technical regulation, which a brand imposing conditions through its purchase orders does not have; Article 4 reaches non-governmental standardizing bodies, which a buyer's specification is not; Article 8 reaches non-governmental conformity assessment bodies, which a buyer's audit is not. The Committee's own notification formats bear this out: every format presupposes a measure of a government body or a standardizing or conformity-assessment body, none a purchase order (WTO, 2026c). The transmission escapes each provision that might have domesticated it. The difficulty is not novel: the same question has occupied the SPS Committee since 2005 without producing a disciplining instrument (Henson and Humphrey, 2010).

The normative resources for addressing it nonetheless exist. In US—Shrimp, the Appellate Body found that a unilateral environmental measure failed not because of its objective but because of the manner of its application: the absence of serious across-the-board negotiation, and a rigidity that required essentially the same regulatory programme of exporting Members without inquiry into the appropriateness of that programme for conditions prevailing in their territories (WTO, 1998). Read against the passport, the reasoning identifies flexibility and negotiation, rather than substantive relaxation, as the relevant obligations. Article 12.3, which requires Members to take account of the special development, financial and trade needs of developing country Members in the preparation of technical regulations, and Article 11, on technical assistance, remain largely unactivated in this context. The concerns raised at the TBT Committee in July 2026, clarity of scope, adequate implementation time, and recognition of equivalent regulations and of conformity assessment conducted abroad, track these provisions closely (WTO, 2026a,b).

The Union's own safeguards illustrate the categorization problem in miniature. Article 22 of the ESPR obliges Member States to support small and medium-sized enterprises through information points, training and, in some cases, financial assistance, and the Commission has undertaken to provide digital tools. These measures are real. They are also territorially bounded: their beneficiaries are undertakings established in the Member States, whereas the burden the cascade transmits is distributed along the chain. A tier-two dyehouse outside the Union bears reporting costs while falling outside every category of assistance designed to offset them. Protection is defined territorially; burden is defined chain-wide.

That this mismatch is consequential rather than theoretical is suggested by readiness data from inside the single market itself. A survey of 1,078 firms in German industry, construction and industrial services found that two thirds were either unaware of the passport (38%) or did not consider it relevant (27%), with only 4% having taken preparatory measures (Büchel and Neligan, 2025). If preparedness is this limited among firms with access to national information points, sectoral associations and Brussels representation, the assumption that suppliers without such access will adapt in step is difficult to sustain. The same asymmetry attends participation: the consultation on the JRC specification, open until June 2026, was formally available to all, but the capacity to respond to a technical consultation of that kind is not evenly distributed.

5 Discussion

The transition to a circular economy has been described as an institutional lag: trade systems built for linear flows have not adapted to remanufactured goods, secondary raw materials and reused components (Tuerk and Sporysheva, 2022). The digital product passport is often presented as a corrective to that lag, an instrument that finally makes circular attributes legible across borders. The preceding analysis suggests a more qualified reading. The passport does not merely encounter an unadapted trade system; it introduces a new mechanism of its own, in which a public obligation is transmitted down the value chain as a private condition, and it does so in a form that the disciplines governing the obligation cannot see. The contribution here is to name that mechanism and to locate the point at which trade law loses sight of it.

This reframes, rather than contradicts, the existing literature. Work on the legal and institutional consequences of passports has noted that their design may reinforce asymmetries among producers, regulators and consumers, while calling for closer attention to the intersection with trade rules (Yuan, 2025). Analysis of cross-jurisdictional traceability has addressed conflicts between comparable regulatory systems and argued for mutual recognition over centralization (Huang and Nottage, 2026). The cascade sits beneath both: it concerns not the relation between regulators but the transmission of a single regulator's requirement to parties who are not its addressees and who have no standing in its design. It is, in this sense, the intra-chain counterpart to the inter-jurisdictional problem others have described.

Who bears this is not evenly distributed, and the point is worth making concretely. The burden falls least on tier-one exporters with a direct relationship to a European brand and most on the tier-two and tier-three wet processors (spinners, dyers, finishers) where the required data originates and where no contractual relationship to the EU operator exists. Their constraints are specific: whether an accredited laboratory is reachable, and whether the national accreditation body belongs to the international laboratory accreditation cooperation (ILAC) and international accreditation forum (IAF) arrangements without which even a local test is not recognized abroad; the cost of chemical analysis levied per production batch rather than per model; and the working capital to pre-finance compliance before an order is confirmed (UNCTAD, 2026). These constraints vary by country. A supplier in Turkey, in a customs union with the Union and largely aligned on standards, is not placed as one in Bangladesh. Under the resolution currently in force, Bangladesh graduates from least-developed-country status in November 2026 (UN General Assembly, 2021); a three-year extension to 2029 has been recommended by the Committee for Development Policy and referred by the Economic and Social Council to the General Assembly, which had not decided at the time of writing (United Nations, 2026). The uncertainty is itself the difficulty. A supplier must commit capital to traceability infrastructure now, on a timetable set by the delegated act, while the tariff terms on which it will sell into the Union after graduation remain unsettled—two adjustments to the same market, neither of whose timing it controls.

If the difficulty is one of categorization rather than intent, the response is a matter of design, and four directions follow. The first is to extend the graduation the JRC has already proposed. Its tiering differentiates by data type; what it does not do is differentiate by the capacity of the actor required to produce the data. A phase-in for the batch-level subset, keyed to supplier size and country status under Article 12.3 of the TBT Agreement, would carry the same logic one step further. The second is mutual recognition of conformity assessment with accreditation of local verifiers, so that verification need not be routed through European service providers—the channel through which much of the cost is presently incurred (WTO, 2026a,b). This runs through the Union's accreditation infrastructure and the existing ILAC and IAF arrangements, with Article 6.1 inviting it; the trade-off is the trust and enforceability that recognition demands where verifier capacity is uneven. The third concerns capacity support, which the Commission has promised through technology transfer and capacity building (European Commission, 2022) but which runs into the territorial limit of Article 22. Extending it would mean attaching passport readiness to the Union's external trade and development instruments rather than to development assistance alone; Article 11 of the TBT Agreement is the unused legal hook, and the trade-off is additionality and the risk of aid conditionality. The fourth is a multilateral, interoperable minimum data model developed through the World Customs Organization and the International Organization for Standardization rather than through regional standardization alone (Gupta, 2024). These bodies are chosen deliberately: the WCO governs the customs data model and stands at the border where the passport would in fact be checked, and ISO, unlike CEN/CENELEC, is an international standardizing body whose deliverables could attract the Article 2.5 presumption, with ISO/TC 323 already working on circularity data (ISO, 2025). The trade-off is the slower pace of consensus standard-setting against the speed of the EU's regulatory calendar, and the risk that European deliverables establish a de facto lock-in before any international model matures.

None of these measures asks the Union to relax the environmental objective, and none disputes that traceability is a precondition of circularity: circular decisions cannot be made on data that does not exist. They ask instead that the instrument be built so that the actors on whom it depends are able to comply, and that its informational value not be quietly eroded at the point of collection. Trade policy, if strategically reformed, can shift from a barrier into an enabler of circular value chains. The passport shows that it can be both at once. Which tendency prevails will be settled less by the environmental ambition written into the regulation than by the design of its transition periods and by the answer to a distributive question that the text currently leaves implicit: who pays for circular transparency, and who is excluded when they cannot.

Three limits should be stated. The regulatory framework remains in formation: the textile delegated act has not been adopted, and the data set discussed here is a technical recommendation the act may revise. Evidence of passport-related supplier exclusion is not yet available, and the consequences traced above are prospective rather than documented—a claim about where the incentives point, not a finding that the outcome has occurred. The analysis is confined to textiles, whose fragmented and geographically dispersed structure may make it an unusually acute case rather than a representative one, and the reading of the TBT Agreement offered here is analytical: no adjudicative determination on these questions exists.

Generative AI statement

The author(s) declared that Generative AI was used in the creation of this manuscript. The author(s) declare that Generative AI was used (Anthropic's Claude) in the preparation of this manuscript. AI was used to assist with refining language, organizing references and improving the clarity and phrasing of the text. The author(s) independently verified all sources, claims, and legal references, and revised the manuscript accordingly.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Konstantinos P. Tsagarakis, Technical University of Crete, Greece

Reviewed by: Sidra Nazir, Ural Federal University, Russia