Customs and Border Protection collected roughly $166 billion under the IEEPA tariffs, across what government declarations describe as more than 330,000 importers and over 53 million entries. CBP has built an administrative refund channel that is already moving tens of billions of dollars. As currently configured, though, it reaches only part of the affected population. A large and growing share of entries sits outside it, and the government now says those entries cannot be refunded at all without a court order covering the importer that filed them. So an importer's refund path today depends less on the merits, which are decided, than on the liquidation posture of its individual entries. It is an entry posture problem before it is a legal one.
Three postures, three different paths
The most useful way for a trade team to read its own exposure is to sort its IEEPA entries into three groups, because each group faces a structurally different route to recovery.
The first group is entries CBP can still process on its own authority. These are entries that have not yet liquidated, plus entries that liquidated recently enough that CBP can still reliquidate them on its own initiative. They are what CBP's refund tool, known as CAPE, the Consolidated Administration and Processing of Entries module inside the ACE portal, is built to handle in its current phase. For this group recovery is mostly administrative. File, validate, get paid.
The second group is entries that have liquidated but remain inside the 180-day protest window under 19 U.S.C. § 1514. These may fall outside CAPE's current intake, but the importer still holds a statutory tool that keeps the liquidation from becoming final and conclusive. A protest here is a preservation device, not a refund mechanism in itself.
The third group is the contested one, entries that are finally liquidated or about to be. CBP's current litigation position is that it has no authority to reliquidate or refund these without a court order covering that importer, which means recovery may require the importer to have sued. This is where the money is most at risk, and where the V.O.S. Selections class fight is being waged.
The reason the third group is so large, and so urgent, lies in how liquidation timing works.
Why final liquidation is the fault line
Liquidation is CBP's final computation of the duties owed on an entry. For a typical formal entry CBP must liquidate within a year, and in practice the cycle begins around day 314. Two clocks then matter, and the practitioner vocabulary depends on keeping them apart.
The first is the 90-day voluntary-reliquidation window under 19 U.S.C. § 1501, during which CBP may reliquidate an entry on its own initiative. The second is the 180-day protest window under 19 U.S.C. § 1514, during which the importer may protest the liquidation. Once the protest window closes without a protest, the liquidation becomes "final and conclusive upon all persons," including the United States.
CAPE's current intake is keyed to the first clock, not the second. In its present phase the tool reaches mainly unliquidated entries and entries that liquidated within roughly the last 80 days, along with certain entries whose liquidation is suspended, extended, or under review. The 80-day cushion is an operational buffer that leaves CBP time to finish reliquidation before its 90-day § 1501 authority lapses. The boundary comes from CBP's own reliquidation authority, not from the protest deadline, and the difference matters. The government and several practitioner commentaries use "finally liquidated" loosely to mean entries past the 90-day reliquidation horizon, while finality for protest purposes under § 1514 turns on the 180-day mark. An entry can be outside CAPE's reach and outside CBP's voluntary-reliquidation authority while the importer's own protest right is still alive. That gap is exactly why the protest window is the practitioner's backstop for the middle group.
A point Judge Eaton has pressed repeatedly sharpens the stakes. Informal entries, the small-value shipments that dominate by volume, are generally deemed liquidated at or near the time of entry. They move into liquidation posture far faster than formal entries, so the refund and preservation windows close much earlier. Eaton has flagged the "millions of informal entries where liquidation was simultaneous, or nearly simultaneous, with the time of entry, and for which the liquidation is now final," and has noted that the government has offered no proposal for them. Because the IEEPA tariffs ran from April 2025 to late February 2026, a very large share of the earliest and smallest entries is already final. The importers most likely to be shut out of the administrative channel are, disproportionately, the smallest ones.
How the CAPE channel actually works
For entries CAPE can reach, the mechanics reward getting the filing right the first time and punish small administrative gaps. CAPE is the CBP portal for administrative IEEPA refund processing. It does not displace protests or CIT actions used to preserve rights on entries it cannot reach, and importers should treat it as one of several tools rather than the only one.
A claim takes the form of a CAPE Declaration, a CSV file uploaded through the CAPE tab in ACE that lists nothing but entry numbers, up to 9,999 per declaration, with multiple declarations permitted. Only the importer of record, or the licensed broker who filed the original entry summaries on that importer's behalf, can file. No supporting documentation is required at submission, but a declaration cannot be amended once submitted, so the grouping and accuracy of each CSV deserve care before upload. ACE validates the file and then each entry, confirming that the entry exists, carries a dutiable IEEPA Chapter 99 line, and is not duplicated. Invalid entries are rejected one by one while valid ones proceed. The system then strips the IEEPA lines, recalculates the duty as if those tariffs had never applied, and reliquidates.
Two operational details cause most of the failures. The refund is paid as a consolidated electronic deposit per importer, so an importer without current ACH banking information on file in ACE will see its refund rejected, a problem that has already stalled thousands of consolidated refunds. And the recalculation removes only the IEEPA layer. Stacked Section 232 and Section 301 duties, and the Section 122 surcharge that replaced IEEPA in late February 2026, all survive and remain payable. An importer expecting a clean refund of everything it paid on a given entry will be disappointed, and in some cases the recalculation can even surface a net bill. CBP indicates refunds generally issue within roughly 60 to 90 days of acceptance, and acceptance itself adds a lag, so a realistic timeline from filing to deposit runs to several months.
Recovery includes statutory interest under 19 U.S.C. § 1505(c), computed under 19 C.F.R. § 24.36 at the rate the IRS sets each quarter under 26 U.S.C. § 6621. The rate is not fixed, and it resets every quarter, so writing a single current figure into a durable guide is a trap. For the quarter running April 1 through June 30, 2026, the overpayment rate is 6 percent for filers other than corporations and 5 percent for corporations, and the slice of a corporate overpayment above $10,000 earns 3.5 percent. On duties held for well over a year, even at these reduced rates the interest is a meaningful addition to the principal. The principal and the interest are also generally treated differently for tax purposes, a detail worth flagging to finance early.
The fight over the finally-liquidated entries
For finally-liquidated entries the administrative channel currently offers nothing, and whether it ever will is now the center of the litigation. CBP told the court in late May 2026, for the first time, that once an entry is finally liquidated it lacks authority to reliquidate or refund without a court order. That position sits awkwardly against the agency's own earlier public guidance, which suggested such entries would be handled in later phases. The Justice Department has built on it in the appeal, arguing both that CBP cannot reach these entries on its own and that Judge Eaton's order directing refunds for all importers, not just those who sued, is an impermissible universal remedy under Trump v. CASA, the Supreme Court's 2025 decision limiting nationwide injunctions.
Both sides lean on the same precedent for the reliquidation question, Shinyei Corp. of America v. United States, a 2004 Federal Circuit decision, and they read it in opposite directions. The plaintiffs and Judge Eaton read it to mean that liquidation finality does not bar a court from ordering reliquidation when the underlying exaction was unlawful. The government reads it as a narrow exception that each importer must invoke through its own suit. Whether Shinyei speaks only to the finality of administrative protest determinations or reaches a presidential tariff order struck down on its face is the genuine open question underneath the procedural noise. It has not been resolved, and a practitioner should treat the outcome as uncertain rather than assume the broad reading prevails.
The drama around the June 9 hearing says more about strategy than about substance. Eaton ordered the CBP Commissioner to testify in person on the compliance timeline. The government resisted, won a temporary stay at the Federal Circuit, and the fight dissolved once Eaton agreed to accept senior CBP trade officials instead. The real signal is institutional. DOJ is trying to move the dispute from implementation speed to remedial power, from how fast refunds happen to whether the court can compel them at all for importers who never sued. That shift is what should shape planning.
The class-action workaround
The plaintiffs' answer to the finality problem is to convert the contested universal relief into a form CASA left intact, a class action. CASA curtailed universal injunctions but expressly preserved class actions as a route to classwide relief, so a certified class would let the court reach importers who never sued without running into the universal-injunction objection. Based on public summaries of the June 4, 2026 motion, lead plaintiff Terry Precision Cycling asks the court to certify a class of importers whose IEEPA refund claims are not currently eligible for processing through the administrative channel, which is functionally the finally-liquidated population. Public trade-press summaries describe the motion as seeking certification under Rule 23(b)(2), but the exact class definition and operative language should be checked against the CIT docket before reliance.
Whether the trade court will certify such a class is a real question, not a formality. Class actions are rare at the CIT and have generally not been certified there. In past refund litigation the court preferred a test-case mechanism over certification. The strongest argument for certification this time is that the usual obstacle, identifying who paid what, has evaporated. CBP's own data identifies every importer and every payment, and the merits question is identical and already decided for everyone. The realistic read is that certification is plausible but not assured, that even a grant would likely draw an appeal, and that the more conventional path remains the lead-case procedure the court has been running. An importer should not treat eventual class coverage as a substitute for protecting its own position.
What a trade team should be doing now
The sequence is straightforward in principle. The first task is visibility. Pull a complete ACE dataset of IEEPA entries across the April 2025 to February 2026 window and sort them into the three postures above, because every later decision depends on knowing which entries sit where. In parallel, confirm that current ACH banking information is on file in ACE, since the most common cause of a stalled refund is a missing or stale bank record rather than anything substantive.
For the first group, the entries the administrative channel can still reach, the priority is to file CAPE declarations promptly. The government is appealing the broader refund order and may seek a stay. Getting eligible declarations accepted before any stay takes hold is the most valuable administrative step available, and the bar on amending a declaration after submission is a reason to get the grouping right, not a reason to wait.
For the second group, entries that have liquidated but remain within the protest window, a protective protest preserves the right against finality and buys time while the litigation sorts out whether later administrative phases will cover them. It is a backstop, to be unwound if a future phase makes it unnecessary.
For the third group, the finally-liquidated entries, this is a question for trade counsel rather than a step to run from a guide. Where an importer has material finally-liquidated exposure, the right move is to ask counsel whether an individual action at the CIT is needed to preserve a refund path, rather than relying on class certification succeeding or on a later administrative phase materializing. For planning purposes the point is simple. Under the government's stated position, a court order covering the importer may be the only reliable route for these entries, and the longer one waits the fewer options remain.
What would change the calculus
Three developments would shift the strategy, and each is worth watching directly rather than through secondhand summaries. If the Federal Circuit stays the broader refund order, administrative access for importers who never sued could freeze, which would push even routine liquidated entries toward the litigation track and make early filing more urgent. If the class is certified and survives appeal, the finally-liquidated population might be reached without individual suits, and individual actions would become insurance rather than necessity. And if CBP announces a later administrative phase covering final entries with a firm timeline, protests and suits on those entries may turn into unwinding exercises, at which point the posture should be reassessed.
Until one of those happens, the safe assumption is the conservative one. The merits are decided, the remedy is contested, the administrative channel reaches only part of the population, and an importer's recovery depends on the posture of its entries and on the steps it takes, or fails to take, before finality closes the door. The figures cited here for collections, refund processing, and entry counts come from government declarations and contemporaneous reporting current to early June 2026 and will move. They, and the docket developments above, should be confirmed against the live CIT record before any decision turns on them.