CBP Proposes Bonds and a $1,000 Floor on Every US Entry Under $2,500

The Analysis

CBP put out a 103-page proposal on 8 October that rewrites how goods valued at 2,500 dollars or less come into the United States, and nearly every headline I have read on it is about duty. This one is not about duty. It is about paper, and about who is allowed to sign it.

Here is what is actually on the table. Informal entry type 11 has to be filed electronically on or before the day of importation, and the 15-day post-arrival window goes away. A new type 13 shows up for mail, and it wants the shipper's identity and address plus a UPU S-10 tracking number. Every filer, type 11 and type 13 alike, has to post a basic importation and entry bond, single transaction or continuous, with liquidated damages set at cargo value, tripled for restricted goods, and a floor of 1,000 dollars. Mail that is not entered within 15 days is deemed abandoned. Comments close on 7 December.

Let me get the quiet part out of the way before anything else. This is a proposed rule. Nothing in it binds anybody today. But if you move low-value parcels into the US, the thing that decides whether you can keep filing your own entries is not the duty line at all. It is the bond.

Take the three hardest facts out of those 103 pages and leave the rest for later. Fact one: the filing clock moves. Type 11 has to be filed electronically on or before importation, full stop — the old rhythm where you filed within 15 days of arrival is gone. That sounds like an operations detail and it is not, because a same-day filing requirement means the data has to be complete before the aircraft lands or the truck reaches the border, not after. Fact two: a bond becomes mandatory on both type 11 and type 13, single transaction or continuous, your choice in principle. Fact three: liquidated damages carry a floor of 1,000 dollars, and restricted goods go to three times cargo value. Everything else in the document — the mail provisions, the S-10 identifier, the abandonment rule — hangs off those three.

Now let me walk a shipment through it, because that is the only way I know how to read a rule like this. A parcel worth 2,500 dollars or less comes in today and goes through five steps, and the proposal touches every one of them.

Classification comes first. On low-value informal entries, classification has historically been loose — nobody argues hard about a tariff line on a 40-dollar item, and the whole point of informal entry was that you did not have to. Put a bond and a liquidated-damages exposure behind that same line and the calculation changes completely: a wrong tariff number is now a wrong declared value, and liquidated damages are assessed on cargo value. Look at the tariff number first, then look at whether any exemption or preference applies. That order is not negotiable and it is exactly the order people skip on small parcels.

Declaration elements come next, and this is where the two new teeth are. One is timing — on or before importation, which means the entry has to be built before the goods arrive, not reconstructed afterwards. The other is content: type 13 wants the shipper's identity and address and a UPU S-10 tracking number. On the postal channel that identifier is not something a broker can conjure up at the border. It has to be generated upstream, at the point where the label is created, which in most operations means the order-taking system. That is a two-to-four-week change to a data pipeline, by my estimate, and it is the kind of change that gets discovered late because everyone assumes somebody else has the field.

Examination is the third step, and the proposal barely talks about it, which is exactly why I am nervous. The treble damages provision turns on what counts as restricted, and I do not have that definition in front of me. Food contact articles, cosmetics, anything with a battery, anything with a children's-product certificate behind it — the list matters enormously here, because tripling the exposure moves a marginal SKU into unviable territory overnight. Which items land in that bucket is something you have to check against the final text and then check again against how your port actually treats them. Single cases are not general rules, and I have watched people get burned generalising from one port's practice.

Taxes and duties are the fourth step, and here is the part the headlines get wrong: this proposal does not change what you owe in duty. That is why I keep saying this is not a duty story. What it adds sits outside the duty line entirely — liquidated damages are a penalty exposure, not a tariff, which means they do not net into your duty drawback, they do not get covered by any duty deferral arrangement you have, and they do not respond to the usual levers you pull when landed cost goes up. If your finance team is modelling this as a duty increase, the model is wrong before it starts.

Release is the fifth step, and the mail provision there is the hardest sentence in the whole document. Mail that has not been entered within 15 days is deemed abandoned. Not delayed, not returned, not held pending correction — abandoned. If a type 13 filing stalls because the shipper identity is missing or the S-10 number never made it into your system, the clock runs and the goods are gone. There is no cure period written into that sentence as I read it. For a postal shipper running thousands of low-value pieces a day, a 15-day hard stop with no cure is the single most dangerous clause in the proposal, and it is getting about a tenth of the attention that the duty angle is getting.

Why now? I am reading between the lines here, so treat this as my view rather than CBP's stated reasoning. Low-value volume into the US has grown to a scale where the informal entry system is processing enormous quantities of goods with very thin data attached, and nowhere is the data thinner than on the postal channel, where the shipper is often a name on a label and nothing more. Requiring the S-10 identifier and the shipper's identity is the tell: CBP is asking for the one thing the postal channel has never reliably given it. And when a regulator wants behaviour to change but cannot efficiently police millions of small filings, the cheapest lever is to make somebody post money. A bond does that without adding a single inspector.

Who feels it, and how much — this is where the answer stops being uniform. A small seller filing its own entries, the independent store shipping 2,000 pieces a month at 30 dollars a pop, is the exposed party, because a bond is not something you decide to have. It is something a surety decides to extend to you, based on your financials and your history, and a lot of small importers simply will not qualify for a continuous bond. A seller working through a platform or an established agent absorbs the change as a fee line rather than as an existential one.

The postal channel carries the data burden I described above plus the abandonment risk. The express and courier channel mostly has to compress its filing window to the day of importation, which is an operational squeeze rather than a structural one. Restricted-category shippers carry the treble exposure. And a traditional importer that already runs a continuous bond for its formal entries is largely unaffected — for them this is one more entry type with one more filing deadline, and the marginal cost is close to zero.

Let me price the part that nobody is pricing. Take a piece with a declared value of 30 dollars. Liquidated damages at cargo value would be 30 dollars. The 1,000-dollar floor makes the starting point 33 times the value of the goods. Assume a seller doing 2,000 pieces a month with a historical rejection or correction rate of 0.5 percent — my assumption, yours will differ, and yours is the one that matters — that is 10 events a month, 10,000 dollars a month, which spread across 2,000 pieces is 5 dollars a piece. Assume a gross margin of 25 percent, so 7.5 dollars a piece. The exposure eats 5 of your 7.5 dollars. At that structure, filing your own entries does not survive contact with arithmetic. And note the direction of that floor: it is a floor, not a cap. Higher-value goods get assessed on value, restricted goods get tripled, and only the bottom of the range is pinned at 1,000.

Now price the bond itself, because the answer is not where people expect. Assume a continuous bond at an annual premium of 500 dollars — my assumption, since real pricing depends on bond amount and on the surety's view of your credit, and this is only an order-of-magnitude figure. At 24,000 entries a year that is about 0.02 dollars per entry, which is nothing. So the bond premium is not the barrier. The barrier is qualification. If you cannot get a continuous bond, you are left with single transaction bonds, and at an assumed 50 to 150 dollars per transaction — again my assumption, check your own quotes — a single-transaction bond on a 30-dollar piece is arithmetically absurd. Which is why the real conclusion here is not about cost. It is about who is allowed to file.

That is the point I have not seen in the coverage, and it is the reason I am writing this at all. Every headline frames this as the end of duty-free treatment for low-value goods. Maybe, but the mechanism that actually reorders the market is the bond. Duty travels with the goods and anybody can calculate it. A bond travels with the person — it is the surety's judgement about you, your balance sheet and your track record, and it is not something you can buy your way into on a bad week. The 1,000-dollar floor then decouples the penalty from the value of the shipment, so the exposure stops scaling with what you shipped and starts being a fixed entry price for doing business at all. That is not a tax. It is a licence, and it decides who still gets to file.

And that is what rewrites the cost structure for small sellers. Assume the marginal cost of self-filing today is 0.50 dollars a piece, which is basically staff time spread over volume. Moving to a qualified agent, assume a quote of 3 dollars a piece for consolidated volume and 15 dollars a piece for one-off, non-consolidated filings — my assumptions, and the spread between them is the whole story. At 3 dollars, 2,000 pieces a month is 6,000 dollars a month, and your 7.5-dollar margin becomes 4.5. Thin, but alive.

At 15 dollars, you lose 8 dollars a piece and the business does not exist. So the action is not go find an agent. The action is to work out, before the rule is final, whether you can consolidate to the volume that earns the 3-dollar rate — I would use 5,000 pieces a month as a working threshold — or whether you raise prices, or whether you drop the low-value SKUs entirely. That calculation should be finished in November, not after the final rule lands.

Timing, then, because the sequencing here is genuinely awkward. Comments close on 7 December. After that CBP issues a final rule on a timeline nobody can promise; I would plan on six to twelve months from the close of the comment period, and I want to be clear that is my estimate rather than a schedule. Effective dates commonly run 30 to 60 days after a final rule publishes, and that is general practice rather than a guarantee. Two things, though, have lead times longer than the rule itself. A continuous bond takes time to underwrite — call it four to eight weeks by my estimate — and the label and order-system changes for shipper identity and S-10 need two to four weeks.

Working backwards from today, that puts the bond question in front of your finance people by mid-November at the latest. Miss that and you have a four-to-eight-week gap between the rule taking effect and your bond being in place, and nobody has answered what you do with the parcels that arrive in the middle of that gap. That is the part that keeps me awake, not the duty.

Dr. Ingrid Voss would slow me down right here, and she would be right to: a proposed rule is a proposal, and the applicability, the exceptions and the definitions — including exactly what counts as restricted for the treble provision — live in the final text, not in the notice. The 7 December window is not a formality either. A flat-dollar floor is precisely the kind of provision that gets revised after industry comments land, because somebody always produces the arithmetic showing what it does to a 30-dollar parcel. So nothing above is a legal conclusion, and you should not file it as one.

Where could this go the other way? Plenty of directions, and one variable flips most of what I have written. If the final rule keeps the bond requirement but replaces the flat 1,000-dollar floor with a percentage — say 200 percent of cargo value — then a 30-dollar piece carries a 60-dollar exposure instead of a 1,000-dollar one, and self-filing stays viable for a lot of sellers who would otherwise be pushed out. The variable to watch is whether the floor is absolute or proportional. The other real possibility is a small-filer exemption: a threshold below which you do not need a continuous bond, perhaps tied to annual entry count or annual value. If that appears, the small seller is back in business and the whole consolidation argument softens. I do not know which way it goes, and nobody who tells you they do is being straight with you.

And on the postal side there is a second-order effect worth naming. If S-10 identifiers have to be captured at label creation, then the party that controls the label controls whether your entry can be filed at all. For most postal shippers that is not the broker, it is the platform or the consolidator or the label vendor. So the practical question you have to answer is not whether your broker can file a type 13. It is whether the system that prints your label can hand you a shipper identity, a shipper address and an S-10 number in a structured field, before the piece moves. If it cannot, the entry cannot be built, the 15 days run, and the goods are abandoned. Documents that do not match cause trouble all the way down the line, and this is that problem in its purest form.

So what do you do, and by when, and who does it. Today, before close of business, have your customs desk pull twelve months of your own type 11 filings: entry count, value distribution — specifically how many pieces sit below 30 dollars and how many between 30 and 100 — and your rejection and correction rate. Owner is customs, not finance, and it should take an afternoon. Every number I have used above is an assumption until you replace it with that table.

Before the end of October, go to your surety channel and ask three questions: what continuous bond amount can you actually get, how long does underwriting take, and what financials do they want. Allow four to eight weeks. Do this even if you conclude you will end up using an agent, because the answer determines which side of the line you are on and therefore what you can negotiate.

By mid-November, build the two-path cost model — self-filing versus a qualified agent — and find your break-even volume. That model needs the real bond quote, the real agent quote and the real value distribution from step one. Then decide whether you consolidate, reprice, or cut the low-value tail. Waiting for the final rule to do this is the mistake, because by then the four-to-eight-week bond lead time sits inside your peak season instead of beside it.

Before 7 December, file comments. That is the only date in this whole sequence where you can still change the outcome, and it will not come back. Comments carry weight in proportion to how specific they are, so bring the numbers: entries per month, value distribution, margin per piece, and what a flat floor does to that margin. A comment that says the rule is burdensome gets read and discarded. A comment that says a 1,000-dollar floor costs 5 dollars a piece against a 7.5-dollar margin on 2,000 pieces a month gets read twice.

On the systems side, get shipper identity, shipper address and the S-10 field into your order and label pipeline now, with a two-to-four-week window, and confirm in writing that your postal or express channel can actually deliver an S-10 number in a structured field. On the classification side, re-walk the tariff lines on every SKU that sits under 2,500 dollars and moves volume, because a wrong line now carries a damages exposure assessed on value. On the contract side, sit down with legal and sales and settle who bears liquidated damages — you, the platform, or the customer — because right now almost nobody's terms and conditions address it, and an unallocated penalty is a penalty you end up paying. This shipment can still be saved, but the work has to start this afternoon, and how the 15-day abandonment clock is applied in practice will come down to the port.

↑ Back to the key points

— By Derek Xu