The Uyghur Forced Labor Prevention Act (signed 23 December 2021): Public Law 117-78 and the Rebuttable Presumption
This record documents the United States Uyghur Forced Labor Prevention Act in full: passed by the House 428 to 1 on 8 December 2021, signed as Public Law 117-78 on 23 December, and enforced from 21 June 2022 through a rebuttable presumption, an Entity List that reached 190 entries, and CBP detention data reported with its dates and sources.
1. Introduction: an import ban that reverses the burden of proof
On 23 December 2021, President Joe Biden signed the Uyghur Forced Labor Prevention Act in Washington, and the bill became Public Law 117-78. It was the first United States statute devoted specifically to forced labour in China's Xinjiang region, and it created the most demanding import-control regime any country has built around the products of a single region. Its core was not another declaration of concern. It changed who has to prove what at the border. From 21 June 2022, goods mined, produced or manufactured wholly or in part in Xinjiang, and goods produced by entities named on a statutory Entity List, are presumed to have been made with forced labour and are barred from entry under Section 307 of the Tariff Act of 1930. An importer who wants an exception must prove the negative, with clear and convincing evidence that the goods were not made with forced labour. This page records the background, legislative history, enforcement machinery, growth of the Entity List, detention data, reactions and the position as of November 2026, drawing on the statutory text, the congressional record, Federal Register notices and data published by United States Customs and Border Protection.
The significance of the Act lies in that reversal. Before it took effect, CBP worked under Section 307 case by case, issuing withhold release orders against particular companies or products only after the government itself had gathered enough information to name a target. That model was slow and narrow. A named producer could change its corporate name, move a production step, or route goods through a third country, and much of the trade continued untouched. The Prevention Act inverted the sequence. Instead of the government proving, shipment by shipment, that a particular consignment was tainted, the law presumes that goods from an entire region, and from every listed entity, are tainted, and it places the work of rebuttal on the importer. For global supply chains that run through Xinjiang cotton, polysilicon and tomato products, compliance review stopped being a problem for a handful of named firms and became a routine obligation for every exporter to the American market. A provision of domestic trade law thereby became one of the few international responses to Xinjiang with operational teeth.
2. Background: Section 307 and the forced labour allegations
The Act's legal foundation is Section 307 of the Tariff Act of 1930, which had long prohibited the importation of goods made with forced labour. A 2016 amendment closed a longstanding loophole, and CBP thereafter made fuller use of withhold release orders, detaining suspect shipments while importers argued their case. Yet the instrument remained case-specific. Every order needed its own evidentiary basis and its own named producer or product, and enforcement resources spread across worldwide supply chains were always thin. Between 2019 and 2021 CBP issued orders touching several Xinjiang producers of cotton, tomato products and silica-based goods, including measures aimed at the Xinjiang Production and Construction Corps and at silica producers. Those orders showed both what the old system could do and where its ceiling lay: only a few nodes of a long industrial chain were ever named, while unnamed stages shipped as before.
The second line of background is the accumulating evidence about forced labour in Xinjiang. After the mass detention system that began in 2017 came to light, researchers, journalists and Uyghur exiles documented released detainees assigned to factory work, organised transfers of rural workers to factories elsewhere in China, and coercive arrangements in cotton picking and polysilicon production. Reports by research institutes traced transfers of Uyghur labour into supply chains serving global brands, and the assessment issued by the Office of the UN High Commissioner for Human Rights on 31 August 2022 listed forced labour among its concerns. The Chinese government has denied throughout that forced labour exists in Xinjiang, describing employment there as voluntary and lawfully protected, and calling outside allegations fabricated. In Congress, by contrast, a bipartisan consensus formed in 2020 and 2021 that case-by-case orders could not answer a systemic problem, and that legislation should designate Xinjiang itself as a presumption region. That consensus is the direct origin of the Act.
One legislative choice deserves separate record. Congress did not build a free-standing prohibition from scratch. It grafted the rebuttable presumption onto the existing Section 307 ban, so that the presumption's legal effect is to bring covered goods within Section 307's prohibition unless the importer rebuts it. The advantage was that the entire customs enforcement apparatus could be used without rebuilding it. The consequence is that every dispute ultimately returns to a single question: does the importer's supply-chain evidence meet the clear and convincing standard? That standard is higher than the ordinary balance-of-probabilities test in civil cases. It demands evidence that is clear and persuasive, which in practice means complete origin tracing, purchase contracts, production records and audit material that corroborate one another. Many shipments have been denied not because forced labour was affirmatively proved in that consignment, but because the importer could not assemble a chain of documents strong enough to displace the presumption. Keeping that distinction in view is essential to reading the enforcement figures in section 6.
3. Legislative history: 428 to 1 and the December 2021 sprint
The road through Congress was not smooth. Similar bills had been introduced in the previous Congress, where a version passed the House by an overwhelming margin but stalled amid Senate scheduling and differences between texts. Reintroduced in 2021, the measure again produced House and Senate versions of differing strength, and coordination between them stalled for months over the presumption's scope, compliance burdens on business and diplomatic fallout. The deadlock broke in December 2021. Once the two chambers settled on a unified text, the remaining steps moved at unusual speed. On 8 December 2021 the House passed the bill by 428 votes to 1, the sole vote against being cast by Representative Thomas Massie. The Senate then passed an amended text by unanimous consent, the House agreed to the Senate amendment on 14 December, and the bill went to the President. On 23 December 2021 President Biden signed it, and the Act became Public Law 117-78.
The 428 to 1 division is worth recording on its own. In a polarised Congress, few measures of any kind attract 428 affirmative votes, let alone a law that directly burdens trade with China and imposes heavy tracing duties on American importers. The tally shows that Xinjiang forced labour had become one of the rare consensus zones in Washington: Democrats supported the bill from human rights and labour standpoints, Republicans from a hard line on China and from supply-chain security, and the White House treated the Act as a signature element of its human rights policy toward Beijing. The debate was not, however, free of reservations. Business groups warned that complete tracing was technically difficult, that compliance costs and port delays would rise, and that importers needed a longer transition and clearer guidance. Those arguments did not change the outcome, but they shaped implementation, including CBP's staged guidance and question-and-answer documents. This page records the vote and the process because the Act is not a temporary executive measure; it is a durable institution enacted through the full legislative process by a near-unanimous legislature.
The statutory text also built two pieces of machinery that are easy to overlook. First, it required the interagency Forced Labor Enforcement Task Force to deliver, within a set deadline, a strategy for forced labour enforcement involving China, and to maintain the Entity List, with additions and removals published through the Federal Register. Second, it required the executive branch to report periodically to Congress on implementation and to address the diplomatic dimension, including possible sanctions against foreign persons involved in forced labour. Congress, in other words, did not merely set a presumption at the border. It tried to build a connected machine of strategy, listing, reporting and diplomacy. How that machine has performed is the subject of the sections that follow, measured by the pace of listing, the detention figures and the widening list of priority sectors.
4. The core mechanism: how the rebuttable presumption works
The presumption began to apply on 21 June 2022, roughly six months after signature, a period left for CBP to issue implementation guidance and for companies to rework supply chains. It captures two classes of goods. The first is geographic: any goods mined, produced or manufactured wholly or in part in Xinjiang fall under the presumption regardless of who exports them and regardless of further processing in a third country, so long as a Xinjiang stage is present. The second is entity-based: goods produced by an entity on the Entity List are covered even if production took place outside Xinjiang. When covered goods arrive at an American port, CBP may detain them, and the importer has three practical routes: show that the goods are outside the presumption's scope altogether, for example because their inputs have no connection to Xinjiang or to any listed entity; apply for an exception and prove, by clear and convincing evidence, that the goods were not made with forced labour; or re-export or abandon the shipment. In practice the first route, proving the goods are out of scope, is the most travelled, and its core is again complete supply-chain tracing.
The clear and convincing standard is where the Act bites hardest. CBP guidance asks importers for documentation across the whole chain, from raw material to finished product, including supplier lists, information on producers at each stage, purchase orders, invoices, transport records, wage and working-hour records and audit reports, and it stresses that the documents must corroborate each other. A bare declaration from a supplier is nowhere near sufficient, and a third-party audit that could not be conducted independently on the ground in Xinjiang carries limited weight. Because genuinely independent auditing inside Xinjiang is close to impossible, many companies have adopted a different compliance strategy: rather than trying to prove that a Xinjiang input is clean, they remove Xinjiang inputs from their United States-bound supply chains altogether. That source-cutting effect is exactly what the Act's authors intended and welcomed. Its other face is stratification: large firms can rebuild supply chains, while smaller importers may exit entire product categories because tracing costs more than the trade is worth. Part of the changing composition of denied shipments reflects precisely that divide.
A further point is often misunderstood and should be stated plainly. Detention or denial of a shipment does not, in law, amount to a finding by the United States government that a particular company used forced labour. The presumption is a rule allocating the burden of proof: on the basis of a legislative judgement about conditions in Xinjiang as a whole, risk is presumed, and where the importer fails to displace that presumption with high-standard evidence, the goods do not enter. News reports that describe every denied shipment as a proven case of forced labour therefore overstate the record, just as claims that denials have nothing to do with forced labour evade the legislative judgement itself. This page accordingly describes the figures as enforcement outcomes under the presumption, and does not elevate individual outcomes into factual findings against named companies. That restraint is owed both to readers and to the standards of an archive.
5. Enforcement architecture: the Task Force and the Entity List
The Act's enforcement hub is the Forced Labor Enforcement Task Force, chaired by the Department of Homeland Security and bringing together seven agencies, among them the Departments of State, Labor, the Treasury, Justice and Commerce, the Department of Agriculture and the Office of the United States Trade Representative. The Task Force's first statutory task was the enforcement strategy on China, published on 17 June 2022, just before the presumption began to apply, which set priority sectors and working methods for customs enforcement. The strategy's original high-priority sectors were apparel, cotton and cotton products, silica-based products including polysilicon, and tomatoes and downstream tomato products, precisely the industries where Xinjiang's weight in world supply and the forced labour allegations were both greatest. Coverage then widened step by step: aluminium, polyvinyl chloride and seafood were added to the priority list, and the strategy update of August 2025 added caustic soda, copper, jujubes, lithium and steel, bringing the total of high-priority sectors to thirteen. Each expansion extends both CBP's inspection focus and importers' tracing duties into new industrial chains.
The Entity List is the Task Force's other central product. It was first published in June 2022, consolidating entities already subject to withhold release orders and related listings, and covering Xinjiang-linked producers across cotton, silica, electronics and textiles, among others. Additions have followed in successive rounds, each announced by Federal Register notice naming the entities added and the basis for listing. On 14 January 2025 the list grew by 37 entities in a single round, the largest single expansion at that time, reaching 144 entities, a figure confirmed in the August 2025 strategy update. Growth continued thereafter: by August 2026 the list stood at 187 entities, and a Federal Register notice of 29 October 2026 added three more, including BGI Genomics, bringing the total to 190. Moving from the first publication to 190 entries took a little over four years. That pace matters: it shows the list is not a one-off political gesture but a rolling enforcement register that continues to be maintained and expanded.
A note on counting is essential here, because without it the numbers mislead. The way the list is counted changed in 2024, when subsidiaries began to be counted separately, with distinct legal entities under one corporate group listed individually. Totals published in different years, and totals quoted by different institutions, therefore cannot be compared directly without explanation, and growth in the headline number cannot be read as wholly proportional growth in newly identified problem companies. Part of the increase reflects genuine new listings; part reflects finer counting. The figures used on this page, 144, 187 and 190, are each tied to their date and source, and readers comparing across years should keep the 2024 counting change in mind. An archive records numbers, and it also records how the numbers were counted; that is what distinguishes it from a news summary.
6. Enforcement data and reactions: detentions, values and a still-growing regime
After more than three years of enforcement, the data have accumulated to a considerable scale. According to CBP data, as cited in an August 2025 assessment by the Center for Strategic and International Studies, by August 2025 CBP had detained 16,755 shipments with a declared value of nearly 3.7 billion dollars, and more than ten thousand of those shipments were ultimately denied entry, the denied goods being worth roughly 900 million dollars. The distinctions matter: detention opens the process, after which some shipments are released on the evidence submitted, some are denied, and some are re-exported or abandoned by the importer. That denied shipments outnumber released ones is unsurprising given the height of the evidentiary bar and the length of the chains to be documented; a single missing link can doom an entire consignment. By value, electronics account for the largest share, chiefly solar modules and their polysilicon supply chain, which follows directly from Xinjiang's importance in world polysilicon capacity. Apparel and cotton goods contribute large numbers of shipments at lower average values.
The origin pattern in the data is equally instructive. Detained shipments come not only from China but also, in significant numbers, from third countries such as Malaysia and Vietnam. That is not mission creep. It is the supply-chain transfer effect made visible: Xinjiang polysilicon, cotton and other inputs move to third countries, are made into modules or garments there, and are then shipped to the United States. Under the geographic presumption, a Xinjiang stage keeps the finished product inside the law's reach, so CBP's focus on inputs rather than on the last exporting country has defeated simple rerouting strategies. Annual volumes are still rising: CBP detained 7,325 shipments in fiscal year 2025, up from 4,850 in fiscal year 2024. Rising detentions may reflect wider sector coverage and stronger enforcement, and may also reflect changes in import volumes; the shipment count alone cannot prove that forced labour itself is rising or falling, and this page records the figures and their sources without drawing inferences the data cannot carry.
One further caution about sources must be entered. CBP's public statistics dashboard was redesigned in 2026 and its counting basis changed with it, so totals quoted by different sources at different times are not directly comparable. The same period can yield different detention totals under the old and new counting bases, and subtracting one source's total from another's to manufacture an increment is a common misuse. The figures used here, 16,755 shipments, nearly 3.7 billion dollars, and more than ten thousand denials worth about 900 million dollars, are uniformly dated to August 2025 and sourced to CBP data as cited in the August 2025 CSIS assessment. A reader who finds a different total on the current dashboard should first check the statistical date and the counting basis before comparing. This may look like pedantry, but for an official-record page it is the part that cannot be skipped: a wrong figure, once endlessly repeated, hardens into a false archive.
China's opposition has run continuously from the legislative stage into enforcement. The Foreign Ministry has repeatedly called the Act political manipulation and interference in internal affairs under the pretext of human rights. The Ministry of Commerce and the Xinjiang regional government have issued repeated statements of opposition, maintaining that employment in Xinjiang is voluntary, that forced labour does not exist, and that American measures disrupt global supply chains and harm the employment and development rights of Xinjiang's people. Within the framework of the Anti-Foreign Sanctions Law adopted in 2021, Beijing has also said it will counteract conduct that cooperates with foreign sanctions to the detriment of Chinese companies, which is widely understood as pressure on firms that cooperate with American enforcement and tracing demands. These are China's stated public positions, recorded here as such. It should also be recorded that, as of November 2026, this page has found no public record of China bringing a World Trade Organization dispute against the Act; the contest has remained at the level of diplomatic statements and domestic countermeasure legislation, and that absence is stated here as a no-public-record finding within this page's checking scope.
Inside the United States, enforcement has not been uncontroversial either. The solar industry felt the sharpest early impact, with large volumes of modules held at ports, and industry groups warned of project delays and rising costs while pressing CBP for faster reviews and clearer release standards. Textile and apparel interests have continued to question the reliability of cotton-tracing technology and the burden on smaller firms. CBP has responded with sector-specific guidance, published answers and industry engagement, and some detained goods have been released once evidence was completed. Congressional criticism, meanwhile, has mostly come from the opposite direction: members of both parties have argued that listing still moves too slowly and that coverage of key sectors remains incomplete, pressing the Task Force to accelerate. The single-round addition of 37 entities in January 2025, the August 2025 strategy update widening priority sectors, and the list's continued growth to 190 in 2026 can all be read against that sustained congressional pressure. As of November 2026 there is no legislative movement to amend or weaken the Act; the presumption regime remains fully in force and the list is still growing, which is the latest position this page records.
The Act has also had a clear demonstration effect abroad. In 2024 the European Union adopted forced labour product legislation establishing that goods made with forced labour must not enter the Union market, though its design differs from the American geographic presumption. Canada and Mexico operate forced labour import prohibitions within the North American trade framework. Corporate tracing systems built to satisfy several jurisdictions increasingly take the Act's evidentiary demands as their reference point, because documentation capable of passing CBP review will usually satisfy most other regimes as well. An American statute aimed at Xinjiang has thus helped to shape the emerging global standard for supply-chain compliance in practice. Whatever view one takes of the Act, that fact alone earns it a page in the international official record.
7. Documents and procedure index
The core official documents and procedural points for this page are as follows. First, the statutory text: the Uyghur Forced Labor Prevention Act, Public Law 117-78, signed on 23 December 2021, with the rebuttable presumption applied from 21 June 2022. Second, the congressional record: the House passed the bill on 8 December 2021 by 428 votes to 1, the sole dissent being Representative Thomas Massie; the Senate passed an amended text by unanimous consent; the House agreed to the Senate amendment on 14 December. Third, the underlying prohibition: Section 307 of the Tariff Act of 1930. Fourth, the enforcement documents: the Forced Labor Enforcement Task Force strategy on China, published 17 June 2022 and updated in August 2025, with high-priority sectors expanded to thirteen. Fifth, the Entity List: first published in June 2022; expanded by 37 entities on 14 January 2025 to 144; standing at 187 in August 2026; and raised to 190 by the Federal Register notice of 29 October 2026 adding three entities, with each round of additions published by Federal Register notice. Sixth, the enforcement data: CBP detention statistics standing, as of August 2025, at 16,755 shipments and nearly 3.7 billion dollars, as cited in the August 2025 assessment of the Center for Strategic and International Studies. Readers should rely on the official statutory text, the Federal Register notices and the version of the CBP dashboard current at the time of checking.
8. Verification note
The legislative dates, the public law number and the 428 to 1 division on this page follow the statutory text and the congressional record, with the sole dissenter's name checked against the House vote record. The 21 June 2022 application date, the clear and convincing evidence standard and the relationship to Section 307 follow the Act's own provisions. Entity List totals are each tied to their source and date: 144 after the single-round addition of 37 on 14 January 2025 and as confirmed in the August 2025 strategy update; 187 as of August 2026; 190 after the Federal Register notice of 29 October 2026 added three entities. This page states expressly that the counting basis changed in 2024, when subsidiaries began to be counted separately, and that totals from different dates and sources are not directly comparable. The enforcement data are uniformly dated to August 2025 and sourced to CBP data as cited in the August 2025 CSIS assessment; the comparison between fiscal 2025 and fiscal 2024 shipment counts carries the caveat that CBP's dashboard was redesigned in 2026 with a changed counting basis. Chinese reactions are relayed from the public statements of the Foreign Ministry, the Ministry of Commerce and the Xinjiang regional government and are labelled as that government's position. The statement that no WTO dispute over the Act appears in the public record is a finding within this page's checking scope as of November 2026. No single quotation exceeds fifty words.
Our Take
Our Take: the Prevention Act is the first international response to Xinjiang that actually changed corporate ledgers. Joint statements, parliamentary motions and assessment reports all matter, but they change discourse, not the movement of goods. This Act changes the fate of every covered shipment at the border. Its 428 to 1 vote shows it was no passing diplomatic posture of one administration, but a long-term institutional choice made by the legislature almost unanimously. Its effects are already written in the data: 16,755 shipments detained, more than ten thousand denied, and companies cutting Xinjiang inputs out of their supply chains in order to pass review, a source-level severance no condemnatory resolution could achieve. This site also records the Act's limits honestly: a presumption is not a factual finding against a particular company, some denied shipments failed on incomplete paperwork rather than proven forced labour, and compliance costs weigh genuinely on smaller firms. But those limits argue for a debate about evidence and procedure, not for a return to statements alone. Beijing says there is no forced labour in Xinjiang. The most powerful answer to that claim would not be diplomatic language but open doors: independent investigation and on-site auditing that let supply chains withstand the very scrutiny Beijing insists they deserve. While the audit door stays shut, a reversed burden of proof is the only workable self-protection an importing country has. The list will keep growing, the detentions will continue, and so will this record.
9. Supplementary record: from case-by-case orders to a regional presumption
Supplementary record: the Act's place in American policy is clearest when set against the withhold release order system it superseded in practice. Under that system, customs officers first had to lock onto a specific target through intelligence, tips and investigation; a single order often covered one company or one product line; and the process from opening a case to issuing an order could take months or years. A named company could escape by renaming, restructuring or moving a production stage, leaving enforcement permanently behind the supply chain's next shape. The Prevention Act reversed that logic wholesale. The legislature made a judgement about conditions across Xinjiang, the presumption went first, and case-level review moved to the importer's rebuttal. The price of the reversal is a heavy review burden and normalised trade friction. Its return is that enforcement, for the first time, runs ahead of supply-chain reshaping rather than behind it. More than four years of data show the reversal has bedded in: detentions counted in the tens of thousands of shipments, a list that keeps expanding, and priority sectors grown from four to thirteen. Anyone studying this institutional turn should mark 23 December 2021, the signing date, as the line between the case-by-case era and the presumption era in American forced labour import policy.
A second detail worth archiving is the third-country factor. As enforcement proceeded, a growing share of detained goods arrived from Malaysia, Vietnam and other third countries. In archival terms, those figures capture a global supply chain reorganising in real time: Xinjiang inputs did not disappear, but were routed through third-country processing before attempting entry to the United States. CBP's answer was to anchor review in the origin of inputs rather than the country of final export, which extends the Act's practical reach far beyond bilateral United States-China trade, into component plants and garment factories across Southeast Asia. Researchers of economic coercion, supply-chain security or sanctions spillover will all have to work with this dataset. This site preserves the Act's enforcement statistics, dated point by point, precisely so that such research can rest on a checkable original series rather than on a bare conclusion. The archive is preserved and never deleted.