The merits are over and the refund route is not self-executing
Learning Resources held that IEEPA does not authorize the President to impose tariffs. The Supreme Court tied that result to Congress's taxing power, the absence of inherent peacetime tariff authority, and the absence of express tariff or duty language in IEEPA. The Court vacated the Learning Resources district-court judgment and remanded with instructions to dismiss for lack of district-court jurisdiction, because the claims fell within the Court of International Trade's exclusive jurisdiction under Section 1581(i), while affirming the V.O.S. Selections judgment in the CIT track, where the Court of International Trade had held that IEEPA did not authorize the challenged worldwide, retaliatory, or trafficking tariff orders.
That settles the tariff-authority defense. It does not make refunds automatic. A tariff can be unlawful while a refund still turns on importer of record, entry number, liquidation date, protest posture, and the remedial form available in court or at CBP. The merits premise is now common to every importer. The route to money is not. In practical terms, a finally-liquidated entry is one where the ordinary CBP reliquidation and protest routes may no longer be open, which leaves the importer dependent on litigation-specific relief, class treatment, or whatever survives appellate review.
The implementation record has already started with CAPE and the CIT refund orders
The refund dispute is no longer waiting for a first implementation signal. The signal has arrived. After the Federal Circuit issued its mandate to the CIT on March 2, 2026, Judge Richard Eaton ordered CBP in a test case, Atmus Filtration, Inc. v. United States, No. 26-1259, on March 4, 2026, to liquidate unliquidated entries and to reliquidate liquidated entries for which liquidation is not final, in each case without regard to the IEEPA duties. Styled as a universal injunction, that order reached all such entries whether or not the importer had sued. On March 27, 2026, Judge Eaton amended the order to cover all IEEPA duties on any entry regardless of liquidation date and status, which extended it to finally liquidated entries.
CBP built a claims module to carry out the recent-entry refunds. Its Consolidated Administration and Processing of Entries process, known as CAPE, went live on April 20, 2026 for claims on most entries within 80 days of liquidation. In its filing in Euro-Notions Florida, Inc. v. United States, CBP reported that as of late May 2026 CAPE had accepted submissions covering approximately $85 billion in potential and certified refunds for processing, and that approximately $20.6 billion in duties plus interest had been completed, certified, and sent to Treasury for disbursement, with both totals rising as CAPE processed more claims, while finally liquidated entries were treated as a separate and later problem. The 80-day line is a CBP implementation cutoff inside CAPE rather than the statutory ninety-day reliquidation period of Section 1501, and CAPE eligibility is a question of CBP processing design rather than ultimate refund entitlement.
CBP has also published a phased schedule. Phase 1, live since April 20, 2026, covers unliquidated entries and entries liquidated within 80 days. Phase 2, scheduled for June 29, 2026, extends the process to entries flagged for reconciliation (entry types 01, 02, and 06) for which a Type 09 reconciliation entry has not been filed, limited to unliquidated entries or entries within 80 days of liquidation, which CBP has estimated at roughly 2.8 million additional entries and about $28.7 billion, bringing combined Phase 1 and Phase 2 coverage to roughly $130 billion of the $166 billion in IEEPA duties at stake. A Phase 3 expected in late July 2026 would reach finally liquidated entries that CBP has said it will refund only for importers that have filed suit at the CIT. That last condition builds the government's litigation position directly into the refund plumbing, because an importer with finally liquidated entries that never sued has no administrative lane even after Phase 3 opens.
The government has not accepted the finally-liquidated half of the order. In a motion filed in V.O.S. Selections on May 29, 2026, the Department of Justice took the position that once an entry is finally liquidated CBP has no authority to reliquidate or refund without a court order, that the universal injunction exceeds the court's jurisdiction and equitable authority under Trump v. CASA, and that refunds for finally liquidated entries must come through importer-specific orders. The government has since appealed the universal refund order, which is now pending, and has framed the fight around whether finally liquidated entries of importers that never sued can be reliquidated absent importer-specific relief or a valid class mechanism. The appeal deadline in the oldest refund order, in Euro-Notions Florida, Inc. v. United States, was reported to fall on June 6, 2026. That posture is why final liquidation is now a litigation variable rather than a back-office customs status.
CASA gives the government a party-scope objection but does not decide the customs question
CASA did not arise from customs refunds. It arose from universal injunctions, and in a stay posture the Supreme Court held that universal injunctions likely exceed the equitable authority Congress has given federal courts, granting partial stays to the extent the injunctions reached beyond complete relief for plaintiffs with standing. The opinion also contrasted universal injunctions with class actions and faulted the former for producing de facto class relief without Rule 23 protections.
CASA gives the government a vocabulary for attacking nonparty refund relief. It does not answer the customs question by itself. The government can argue that an order requiring refunds for absent importers resembles the universal relief CASA rejected. Importers can answer that Section 1581 gives the CIT specialized customs jurisdiction, and that the CIT in V.O.S. Selections treated tariff uniformity as a reason the challenged orders could not be unlawful only as to named plaintiffs. The point is not that CASA decides the refund issue. The point is that CASA makes party scope a live appellate issue, which is the ground the government has chosen to fight on.
Rule 23 is the bridge for nonparties and Dukes narrows it
The class question is no longer hypothetical. On June 4, 2026, an importer plaintiff moved in V.O.S. Selections to certify a class of importers whose IEEPA refund claims fall outside current CAPE eligibility, using the recognized exception that lets a certified class carry relief across its members where a universal injunction cannot. On June 25, 2026, the government filed its opposition to certification. At a hearing on June 9, 2026, the court had signaled reluctance to proceed on a class basis and pressed the government to abandon its appeal of the refund order instead. In its opposition the government also stated for the first time that CBP is prepared to begin refunds by August 2026 for the roughly 4,000 importers that have filed suit, should the court order reliquidation in those cases, which underscores that filing at the CIT remains the most assured route. That motion is what ties the title of this analysis to a live filing rather than a thought experiment.
Dukes is the reason certification is contested. It held that Rule 23(b)(2) fits where one indivisible remedy would provide relief to each class member, and that individualized monetary claims belong instead in Rule 23(b)(3), with predominance, superiority, notice, and opt-out protections. The refund class can be framed two ways. Plaintiffs can describe the relief as one uniform injunction directing CBP to process unlawful IEEPA duties through a single mechanism. The government can describe the same relief as money for each importer, calculated entry by entry and shaped by liquidation status, protest posture, interest, and offsets. On the second framing the government will argue that Dukes makes Rule 23(b)(2) unavailable and pushes the case toward Rule 23(b)(3) or individual suits. Class actions are rare at the CIT, the certification question can take years, and it is unsettled whether the pendency of a putative class tolls the limitation clock, so an importer that relies on the class and loses could find itself time-barred.
Customs finality makes entry posture decisive
Three statutes sort importers into different refund postures. Section 1501 permits CBP to reliquidate an entry within ninety days from the date of original liquidation. Section 1505 requires CBP to refund excess money with interest when the excess is determined on liquidation or reliquidation. Section 1514 makes listed customs decisions final and conclusive on all persons unless a protest is filed or a protest-denial action is brought in the CIT, sets the general protest window at 180 days after liquidation, and feeds the jurisdictional choice between Section 1581(a) protest-denial actions and Section 1581(i) residual jurisdiction. Section 2636 bars protest-denial actions after 180 days and bars covered Section 1581(i) actions two years after accrual.
These provisions decide the clock, and the merits holding does not. An unliquidated entry, an entry inside the reliquidation window, an entry inside the protest window, and a finally-liquidated entry do not share the same route to money. The merits holding is common. The refund clock is not.
The final-entry problem is not simply that liquidation is final. It is that finality is now the line the government is defending against nonparty refunds. The CIT orders point toward broad relief, while the government's appeal seeks to preserve the argument that finally liquidated entries cannot be reopened for non-suing importers absent importer-specific litigation or a valid class mechanism. For finally liquidated entries, Section 1581(i) should be treated as a contested preservation theory rather than a clean substitute for protest jurisdiction. The government can argue that Section 1514 finality and the availability or adequacy of protest remedies foreclose residual jurisdiction. The importer's answer depends on timing, entry posture, whether a protest route was available, and whether the CIT refund orders survive appeal.
What importers should do now
Importers with material IEEPA exposure should sort entries by liquidation status before deciding anything about litigation, because the posture controls both the available channel and the deadline.
| Entry posture | First action | Key risk |
|---|
| Unliquidated | File a CAPE claim for liquidation without IEEPA duties | CAPE eligibility conditions |
| Liquidated within 80 days | File a CAPE claim inside the window | CAPE 80-day cutoff |
| Inside protest window | File a protest that identifies the IEEPA exaction and refund theory with enough specificity to preserve jurisdiction | A vague protest risks a later Section 1514 or Section 1581(a) defect |
| Protest denied | Calendar the 180-day CIT clock from denial or deemed denial | Missing it converts a live refund claim into a finality problem |
| Finally liquidated, no suit | Weigh class coverage, appeal outcome, and a contested Section 1581(i) action with counsel | Class tolling and universal relief are both contested, so neither is a safe default |
| Finally liquidated, pending suit | Confirm case-specific order scope and any later CAPE phase | Stay, appeal, and order scope |
Beyond the schedule, importers should keep one master record for the affected universe showing entry number, importer of record, IEEPA payment, liquidation date, protest deadline, reliquidation window, CAPE status, surety involvement, and responsible entity, and should refresh it as CBP liquidates, reliquidates, denies protests, or issues refund instructions. A static schedule can misstate the operative deadline after a new customs event. Importers should also separate the customs refund right from downstream pass-through disputes. The customs claim follows the importer of record and the entry record. Whether customers, suppliers, carriers, platforms, or affiliates ultimately bore the tariff cost belongs to a different contract and consumer-law layer.
What would change the calculus
The benchmarks have narrowed. The first benchmark is no longer whether CBP will build a refund process, because CAPE exists and is paying claims. The live benchmarks are whether CAPE's planned Phase 3 for finally liquidated entries, which CBP has limited to importers that filed suit, survives the appeal unchanged, whether the Federal Circuit stays or narrows the CIT order on finally liquidated entries, whether the CIT certifies a class as a valid nonparty vehicle or declines and presses the government to drop its appeal instead, and whether final liquidation remains a statutory barrier for importers that did not sue. A stay that pauses refund processing during the appeal would not change the merits holding, but it would change the expected value of waiting for aggregate relief, most of all for importers whose entries are already final.
Caveats
This analysis should not be read to describe refunds as automatic. CAPE is an implementation channel rather than a blanket self-executing refund right for every importer and every entry posture.
The analysis distinguishes ordinary customs finality from the disputed IEEPA refund orders. Section 1514 finality remains a government defense, but the CIT orders and the pending appeal make final-entry treatment an active litigation issue rather than a settled exclusion.
Statements about class relief are tied to the actual certification motion and do not imply that Rule 23(b)(2) certification is either assured or impossible. Dukes gives the government a serious objection. It does not decide the motion in advance.
The analysis does not tell every importer to sue. The safer formulation is that importers with material final-entry exposure should evaluate importer-specific CIT preservation while CAPE scope, class certification, and appellate review remain unsettled. Later docket items should be confirmed against official records before they are relied on.