The most expensive sentence I hear in a compliance review is "we noticed that a while ago." Someone in the trade team spotted a classification that had drifted, or a royalty that never made it into declared value, and the file went into a drawer because raising it felt riskier than sitting on it. In one 2025 review I priced a drawer like that at $2.3m of unpaid duty across four years of entries. For years that instinct was merely costly. From September 2026 it becomes a good deal worse, because the arithmetic that made sitting still tolerable is being rewritten by CBP.
On 3 June 2026 the White House signed the executive order on strengthening customs enforcement. Latham & Watkins Most of the coverage went to the importer-of-record and bonding provisions, which is fair, since those reshape who may even file an entry. The provision I would put in front of a CFO sits further down the document: within 90 days, so by roughly 1 September 2026, CBP must revise its penalty mitigation standards to set a minimum penalty floor of not less than 50% of the assessed penalty, absent exceptional circumstances materially affecting national security, and to eliminate mitigation entirely for repeat offenders. Snell & Wilmer
Read that twice, and note where the exception sits. The carve-out is not a general "exceptional circumstances" escape hatch of the kind that fills most enforcement policies. It is tied specifically to circumstances that materially impact national security. Snell & Wilmer For the overwhelming majority of commercial violations, which is to say a misclassified component or an unreported royalty, there is no national-security dimension and therefore no route around the floor. Read as written, 50% is a hard minimum for ordinary trade cases.
One thing to be clear about before going further, because I have already heard it misread. The 90-day instruction is a deadline for CBP to revise its standards, not a grace period for importers. Snell & Wilmer Nothing in the order suspends enforcement until September, and nothing protects a violation merely because it existed before the new guidelines issue. What the calendar gives you is a window in which the old mitigation arithmetic still governs your options, which is a different and much narrower thing than immunity.
This guide is about that arithmetic. If you want the vetting, data, and bonding side of the same order, we covered it separately in our walkthrough of what the 2026 customs enforcement executive order means for importers of record. What follows is strictly the penalty machinery.
How a customs penalty is actually built
Almost nobody outside the trade bar can explain where a penalty number comes from, which is why a first notice from CBP lands as a shock. The governing statute is section 1592 of title 19 of the US Code, and it sorts every violation into one of three culpability levels. 19 U.S.C. 1592 The level, not the size of the error, drives the exposure, and the spread between the top and bottom tier on the same facts is routinely a factor of 10 or more.
Negligence sits at the bottom. A negligent violation carries a civil penalty of up to two times the lawful duties, taxes, and fees the United States was deprived of. Where the violation did not affect the assessment of duties at all, the ceiling is 20% of the dutiable value of the merchandise instead. 19 U.S.C. 1592
Gross negligence doubles the multiplier. The ceiling becomes four times the deprived duties, taxes, and fees, or 40% of dutiable value where revenue was not affected. 19 U.S.C. 1592
Fraud abandons multipliers altogether. The statutory maximum is the domestic value of the merchandise. 19 U.S.C. 1592 Not the duty, not a multiple of the duty, the value of the goods. I have seen a fraud allegation on a duty shortfall of under $90,000 generate an exposure figure larger than the annual profit of the importing entity, purely because the merchandise behind those entries was worth about $6m.
| Culpability | Statutory ceiling (revenue affected) | Ceiling (no revenue effect) |
|---|---|---|
| Negligence | 2× duties, taxes and fees deprived | 20% of dutiable value |
| Gross negligence | 4× duties, taxes and fees deprived | 40% of dutiable value |
| Fraud | Domestic value of the merchandise | Domestic value of the merchandise |
Those are ceilings, not invoices. The gap between the ceiling and what an importer actually pays has always been filled by mitigation, and mitigation is exactly what the June 2026 order targets.
Mitigation is the part that is disappearing
CBP publishes its mitigation framework openly, in the informed compliance publication on fines, penalties, forfeitures and liquidated damages. U.S. Customs and Border Protection In practice a penalty notice arrives at or near the statutory maximum, counsel files a petition under part 171 of title 19 of the customs regulations, and the agency reduces the number in light of mitigating factors: contributory error by CBP itself, cooperation during the investigation, immediate remedial action, a clean prior record, inexperience in importing, and so on.
The reductions have been substantial. In the negligence cases I worked between 2019 and 2025, a well-documented part 171 petition citing prompt remedial action and a clean history routinely brought a penalty down to a small fraction of the 2x ceiling, sometimes to within 10% or 20% of the underlying duty loss itself. That is the practical reason importers have tolerated known exposure: the worst case was rarely the actual case.
A floor of not less than 50% of the assessed penalty removes most of that cushion, and because the only stated exception turns on national security, a commercial importer should assume the floor applies to it. Snell & Wilmer If a gross-negligence assessment starts at four times a $400,000 duty loss, the assessment is $1.6m, and a floor at half of it means the best realistic outcome after a full petition is $800,000. Under the framework that has applied until now, a strong petition on the same facts could plausibly have landed far below that. Nothing about the underlying error changed. The negotiating range did.
The second half of the provision is the one that should worry any importer with a history. The order directs CBP to eliminate mitigation for repeat offenders, which on its face means a company with a prior penalty on file loses access to the reduction machinery altogether. Snell & Wilmer A first violation would stop being a one-off cost and become a change in status.
Here I have to be honest about what is not yet decided. The order mandates the outcome and leaves the definition to CBP, and as of late July 2026 the agency had not said whether a repeat offender means anyone with any prior violation, a specific count, or a violation inside a defined lookback period. Trade counsel writing through July flagged that this single definitional choice will drive most of the provision's practical effect. Treat what follows as the conservative reading, which is the one I plan around: assume any prior 1592 penalty counts until CBP says otherwise. If you have ever settled a 1592 case, and many of the 200-plus importers I have advised since 2018 have, plan on the possibility that your exposure on the next one differs materially from that of a competitor with a clean file.
Prior disclosure, and why the calendar matters
The statute contains a mechanism that most importers know exists and few use well. If the person concerned discloses the circumstances of a violation before, or without knowledge of, the commencement of a formal investigation into it, the merchandise is not seized and the monetary penalty is capped at sharply reduced amounts. 19 U.S.C. 1592
The reduction is not cosmetic. As the statute is written, a valid prior disclosure on a negligent or grossly negligent violation limits the penalty to the interest on the duties that were not paid, and on a fraud violation it limits the penalty to one time the loss of duties rather than the domestic value of the goods. Torres Trade Law You still tender the underlying duty. What you buy is the collapse of the multiplier, and in a fraud case the difference between one times the duty and the value of the merchandise can be two orders of magnitude.
The timing condition is the whole game. The disclosure must precede CBP's formal investigation, or at least your knowledge of it. Once the agency opens a file, or once a CF-28 request for information or CF-29 notice of action lands on the subject, the door is closing. Practitioners across the trade bar reached the same conclusion in June 2026: importers should be deciding now whether a disclosure is appropriate, before the revised standards take effect and before CBP finds the issue on its own. Snell & Wilmer
I want to be precise about what that means, because "disclose before September" is being repeated in two wrong directions at once. It is not a cliff for the disclosure mechanism: prior disclosure under 1592(c)(4) is statutory and remains available after the standards are rewritten. It is also not an amnesty window: a violation you are sitting on today is fully exposed today, and CBP can open a file in August as easily as in October.
What the date changes is the value of your alternative. Today an importer can weigh disclosure against the odds of a favourable part 171 petition later. Once a 50% floor is in force, with only a national-security exception to it, and once a prior violation may forfeit mitigation entirely, the petition path is worth considerably less. Disclosure becomes the dominant option across a much wider set of cases. The decision does not expire. The comfortable version of the decision does.
The self-review I run, in the order I run it
A disclosure decision is only as good as the facts behind it, and most importers do not have those facts to hand. When a client asks me to scope exposure inside two weeks, I work through the 4 areas where the money actually hides, in this order.
- Declared value. Assists, royalties, licence fees, commissions to related parties, post-importation price adjustments. This is where the largest under-declarations I have found originated, and it is rarely anyone's deliberate choice. It is usually a contract signed by a team that never spoke to the customs broker.
- Classification and origin. Tariff lines that were set once and never revisited while the product changed, and origin claims resting on supplier certificates nobody has re-validated. With current duty rates, a single wrong line on a high-volume SKU compounds fast.
- Trade-programme claims. Preference claims, exclusions, first-sale valuation. Each is legitimate and each requires documentation you must actually possess. Our guide to the first sale rule and legal ways to cut US duty sets out the evidence standard, and the same discipline applies to every other claim on your entries.
- Restricted-goods and forced-labour exposure. Violations here do not always affect duty assessment, which pushes them into the 20% and 40%-of-dutiable-value branch of the statute rather than the duty-multiple branch. Our 2026 forced-labour import compliance guide covers the supply-chain evidence CBP now expects.
Two practical points about scope. Run the review under privilege, with counsel, because you may be documenting a violation. And bound it by the limitations period: the general customs statute of limitations runs five years, with the clock on fraud starting from discovery of the violation rather than the entry date. 19 U.S.C. 1621 Five years of entry data is a large pull, and it is the pull that matters.
Two files, one decision, and why they went different ways
Earlier in 2026 I reviewed two importers within a month of each other. Both had found a valuation problem. The similarity ended there.
The first was a mid-sized consumer-goods importer paying a design fee to a supplier's affiliate and booking it as a marketing cost. It was an assist under the valuation rules, it belonged in declared value, and it had been running for 3 years across roughly 400 entries. Clean record, no prior 1592 penalty, cooperative management, documentation intact. We quantified the duty loss at just under $600,000, filed a prior disclosure, tendered the duty, and closed it. The cost was the duty plus interest plus professional fees. Under a 50% floor the same facts, discovered by CBP instead of by us, would have been an entirely different conversation, and the company would have carried a first violation into every future dispute.
The second had a classification question on a product line where the technical answer was genuinely arguable between two HTS headings. Reasonable specialists disagreed, and there was a defensible CBP ruling history pointing their way. That is not a disclosure case, it is a ruling case, and the right move was to seek a binding ruling from CBP prospectively rather than to concede a violation that may not exist. The distinction I hold to is simple: disclose what you know is wrong, and get a ruling on what you are unsure about. Conflating the two either buries real exposure or manufactures a violation out of a legitimate difference of view.
What the floor does to the way you argue a case
If you end up in a penalty proceeding after September 2026, the strategy shifts upstream. Under the old arithmetic, a great deal of effort went into the part 171 petition, because the petition was where the number moved. With the floor in place, the number moves most where culpability is decided, and the step from negligence at 2x to gross negligence at 4x is now worth twice as much as it was, because the 50% floor applies to whatever assessment you failed to argue down.
That puts a premium on evidence you can only create in advance. A documented compliance programme, written procedures, records of internal audits, proof that you acted on what you found, and a demonstrable reasonable-care posture under the statutory standard. 19 U.S.C. 1484 Those artefacts are what separate a 2x negligence finding from a 4x gross-negligence one, and they cannot be assembled after a CBP notice arrives. Build the file while nothing is happening, because its only job is to exist on the day something does.
The bonding side deserves a glance too, since penalty exposure and bond sufficiency travel together. If your duty spend has climbed with current tariff rates, your bond may already be undersized, and a penalty file is a bad moment to discover it. Our comparison of continuous versus single-entry customs bonds covers the sizing arithmetic.
The other side of the ledger
One thing I insist on when a board hears the word penalty for the first time: the enforcement story is only half of the picture. The same body of law that assesses penalties also contains substantial refund and reduction mechanisms that plenty of importers never claim. Drawback recovers up to 99% of duties paid on goods that are subsequently exported or destroyed, and our 2026 duty drawback guide walks the mechanics. Duty deferral through a bonded warehouse or a foreign trade zone changes when, and sometimes whether, duty is owed at all.
I raise these in the same meeting deliberately. A trade function that only ever reports risk gets funded like an insurance premium. One that recovers 99% drawback and reduces penalty exposure gets funded like an operation, and it is the funded one that finds the valuation problem before CBP does.
What I would do in August
If I were sitting in an importer's compliance seat in August 2026, my sequence would be short. Pull 5 years of ACE entry data and run the 4-area review above, under privilege, starting with declared value because that is where the largest numbers live.
Where you find a clear violation, get a disclosure decision made on a calendar, not in a drawer, and remember that the alternative to disclosure is worth less than it was in June. Where you find genuine ambiguity between two HTS headings, file for a binding ruling instead of conceding. Check whether you have a prior 1592 penalty on file, because under the conservative reading of the repeat-offender provision your risk profile may not be the same as a competitor's with a clean record. And write down the compliance programme you already run informally, because that document is what argues the culpability tier for you later.
Then watch for two specific pieces of text. The first is the revised CBP mitigation guidelines themselves, due within 90 days of 3 June and therefore around 1 September, which will show how narrowly the agency reads the national-security exception. The second is the definition of repeat offender, which the order does not supply and which will determine whether one old settlement follows a company indefinitely or drops out of consideration after a defined period. Neither was public as of late July 2026, and both matter more to your exposure than anything in the order's headline.
Until they issue, the safe assumption is the one Executive Order 14411 states plainly: the discount you were counting on is being cut to 50% for everything that does not touch national security. Do not read the September date as protection for the months before it. Read it as the last stretch in which the old arithmetic still shapes your choices, and act while it does.
This is general information for trade and logistics professionals, not legal advice. Penalty and disclosure decisions turn on specific facts and should be made with qualified customs counsel.