Home / Guides / Forced-Labor Compliance

UFLPA Compliance Checklist: Clearing a Withhold Release Order Stop

Updated 2026-10-21. Every figure below is recomputed from our locally maintained copies of the official UFLPA data pipelines; provenance notes are embedded in the page source.

1. Three instruments, one burden of proof

Importers conflate three distinct legal tools, and the distinction changes what evidence clears the stop. Section 307 of the Tariff Act of 1930 (19 U.S.C. 1307) is the foundation: it prohibits entry of any goods made wholly or in part with forced labor. A Withhold Release Order (WRO) applies that prohibition to named producers, regions or commodities, and lets CBP detain shipments until the importer demonstrates clean production. The Uyghur Forced Labor Prevention Act (UFLPA), Public Law 117-78, signed December 23, 2021 and enforceable since June 21, 2022, goes further: it creates a rebuttable presumption that any goods mined, produced, or manufactured wholly or in part in the Xinjiang Uyghur Autonomous Region — or by an entity on the statutory list — are made with forced labor and barred from entry under Section 307.

The operational consequence: under a WRO you argue about the specific facility; under UFLPA you must trace the entire bill of materials back through possibly several tiers, because partial XUAR content anywhere upstream triggers the presumption. The statute directs the enforcement task force to maintain lists covering entities using forced labor, entities working with government labor-transfer programs, the products they make, and the facilities sourcing material from the region — and it names cotton, tomatoes, and polysilicon as high-priority sectors by law.

2. What the enforcement numbers actually show

We recompute the following table at generation time from our structured extraction of CBP's official UFLPA Enforcement Statistics Dashboard Guide (November 2025 data window). Every figure maps to a named dashboard figure (Fig 2, Fig 8, Fig 20, Fig 29–30); the full per-row citation table lives on our UFLPA enforcement statistics page.

Dispositions recomputed from data/uflpa-enforcement-stats.jsonl; “—” = disposition not stated in that dashboard figure. Stopped = denied + released + pending wherever all four are stated; Thailand-origin pending is such an unstated case (Fig 8 breaks out only stopped/denied/released).

Three readings matter for planning. First, just over 60% of all-time stopped lines were ultimately released (39,829 of 65,707 = 60.6%), so a stop is expensive but survivable with evidence. Second, outcomes are industry-specific: FY2024 electronics releases ran at a 70.0% rate (2,163 of 3,088), while the FY2025 aviation-and-aerospace slice under HTS 8806 saw 5,344 of 5,906 stopped lines denied — a reminder that aircraft-parts supply chains were hit hardest. Third, third-country routes are firmly inside the enforcement perimeter, with 1,268 Thailand-linked stops recorded through the cutoff.

The 1,663 still-pending lines are the quiet warning: reviews take months, cargo sits in bonded storage, and buyers cancel during exactly that window. Budget the carrying cost before you commit to a Xinjiang-exposed input, not after CBP commits you to it.

3. Know your counterparty: the entity-list census

The statutory entity list is the fastest-moving risk surface. Our ledger, parsed from the Federal Register notices themselves, tracks 208 rows; the current roster stands at 189 entities, added by notice 2026-15628 with an effective date of August 3, 2026. Sibling pages may print 188 instead — the same roster after merging alias spellings into one legal entity. The remaining rows are historical bookkeeping: renames, consolidations and superseded notices that still trip naive name-matching.

• Clause distribution matters. 132 of the 189 current entries sit in clause (v) — the labor-transfer provision — with 36 in clause (ii), 11 in both, and a small tail in clause (i) and in combined designations. Clause text tells you what conduct put the entity on the list, which shapes what evidence will rebut the presumption for downstream goods.
• Aliases are the practical trap. 74 of our 208 rows carry alias records. Screening on the exact legal name without alias folding is how banned suppliers slip through vendor-onboarding questionnaires.
• Statuses are additive. The raw notices mark 156 rows as “listed” and 52 as “added”; our pipeline reconciles both into the current-roster flag rather than trusting either label alone.

Browse the full directory with per-record provenance in our UFLPA entity list browser, run instant name-and-alias lookups in the free UFLPA Entity List search tool, or screen your whole vendor list through the sanctions and entity screening tool.

4. The ten-point rebuttal checklist

1. Map the supply chain to the raw-material tier. Tier-one invoices do not rebut anything. Trace cotton-to-yarn or silica-to-polysilicon depth, because the presumption attaches at any XUAR touchpoint.
2. Screen every supplier against the current roster and its aliases. Re-run at each Federal Register update — 189 entities today, more likely next quarter. Automate it; manual quarterly checks are how detentions happen.
3. Classify precisely before quoting landed cost. Enforcement intensity varies sharply by HTS family (compare HTS 8806's 5,344 denials with electronics' 70.0% release rate). Our HS code lookup and the Section 301 stack map belong in the same workflow.
4. Demand transaction-level records, not attestations. Under UFLPA the exception standard is clear and convincing evidence — a signed questionnaire is neither clear nor convincing on its own.
5. Keep origin declarations consistent across documents. CBP reconciles entry summaries, commercial invoices and producer letters; discrepancies convert a routine review into a full detention.
6. Pre-stage the rebuttal file. Assemble the tracing dossier, affidavits and supporting testimony before entry, indexed by purchase order, so a stop at the port becomes a document hand-off rather than a scramble.
7. Model the cash-flow exposure. Detained cargo accrues storage and per-diem while evidence is reviewed; price that carrying cost into quotes with the landed cost calculator.
8. Watch transshipment partners. Goods routed through Malaysia, Thailand or Vietnam stay in scope if inputs originated in XUAR — the 1,268 Thailand-linked stops prove the point.
9. Track the data vintage. The 2026 dashboard revision counts shipments at the line level and supersedes earlier series; never benchmark your exposure against pre-2026 aggregates.
10. Escalate genuinely clean chains to binding guidance. Where tracing shows no regional nexus but CBP disagrees, the remedy is administrative — not another attestation letter. Document, submit, and iterate.

5. Frequently asked questions

What is the difference between UFLPA and a Withhold Release Order?

UFLPA is a statute (Public Law 117-78) that attaches a region-wide rebuttable presumption to goods with any Xinjiang nexus, while a Withhold Release Order is a narrower Section 307 instrument CBP issues against specific producers, regions or commodities. Both block entry; UFLPA shifts the burden of proof to the importer by statute.

How many shipments has CBP stopped under UFLPA?

Through the November 2025 data cutoff of CBP's enforcement dashboard, 65,707 shipment lines were stopped for review, of which 24,215 were denied entry, 39,829 were released and 1,663 remained pending. Fiscal year 2024 alone saw 11,778 stops.

How many entities are on the UFLPA Entity List now?

The latest roster added by Federal Register notice 2026-15628, effective August 3, 2026, carries 189 entities; our tracking ledger keeps 208 rows including renamed and superseded records. Most entries sit in clause (v), the labor-transfer provision, with 132 of the 189.

What does clear and convincing evidence mean in practice?

It is the statutory standard an importer must meet to defeat the presumption: comprehensive supply-chain tracing down to the raw-material tier, transaction records, and affidavits or testimony that together show no forced-labor input. Clean customs paperwork alone does not reach this bar.

Do UFLPA rules apply if my goods ship from a third country like Thailand or Vietnam?

Yes. The presumption follows goods mined, produced or manufactured wholly or in part in Xinjiang regardless of where they are exported from, and CBP's own statistics show third-country routing under review — 1,268 Thailand-linked shipment lines stopped through the data cutoff.

Which sectors should expect the most scrutiny?

The statute itself names cotton, tomatoes and polysilicon as high-priority enforcement sectors, and the dashboard shows FY2025 apparel-footwear-textile stops from China at 6,637 shipment lines plus 5,906 automotive-and-aerospace stops under HTS 8806.

Where can I screen my suppliers against the UFLPA Entity List?

Use our free sanctions screening tool to match supplier names and aliases against the entity list we maintain from the Federal Register, then browse the full entity directory for provenance details on every record.

6. Recap

• UFLPA flips the burden of proof: a Xinjiang nexus anywhere in the chain presumes forced labor unless clear and convincing evidence says otherwise.
• Enforcement is large and measurable: 65,707 stopped lines, 24,215 denials, and a conservation identity (denied + released + pending = stopped) that our datasets reproduce wherever the guide states all four dispositions.
• The entity list moves in Federal Register increments — 189 current entities as of the August 3, 2026 roster — and alias-aware screening is the only screening that works.
• The ten-point checklist above converts a detention from a crisis into a document hand-off; start at mapping, not at the port.