Mexico, Peru and Korea push back at USTR forced-labor tariff hearing

The Analysis

On 7 July the USTR held a hearing on a proposed duty of 10% to 12.5% on imports from 60 economies, on the ground that their exports involve forced labour. Ministers and representatives from Mexico, Peru, Guatemala and Ecuador rejected the claim, and Mexico's deputy economy minister called the proposed tariff unreasonable. South Korea said the measure was excessive and unjustified. The duty is meant to take effect later in July.

Strictly speaking, nothing was decided on 7 July. A hearing builds a record, a record supports a determination, a determination becomes an instrument, and an instrument has a scope clause and an effective date. The ministers were not arguing with each other; they were putting evidence into a file that someone else will read when deciding whether the measure survives review. The loudest sentence in the room is rarely the one that matters.

The question I would put first is not how high the duty is. It is what the duty attaches to, an economy, a sector, or a named facility, because that single choice decides who has to prove what, and there is a category of importer who cannot prove it at any price.

Forced-labour measures in trade practice have usually attached to entities and facilities: a named producer, a named region, a named product line, each with its own finding and its own evidence. A duty aimed at 60 economies as economies is a different instrument, because it makes origin the proxy for conduct. Once origin is the proxy, the burden of proof inverts. Under a facility-based measure the authority must identify the facility, and your problem is traceability, meaning can you show which factory made this carton. Under an economy-based measure you must show the negative, that this consignment, from this factory, in this province, is not in the category. Proving a negative costs more than proving a positive, and it costs most for the importer with the least documentation.

That inversion, and not the two and a half points between 10% and 12.5%, is what I would carry into a board meeting. A rate is a number you can put in a model. A burden of proof is a capability you either have or do not have on the effective date, and no amount of hedging builds it in three weeks.

The exceptions live in the scope clause. Goods whose specific supply chain contains no labour input of the kind described sit outside the rationale even if they sit inside the country list. Goods already covered by an entity-specific measure raise a double-counting question the instrument must answer. Goods whose origin is outside the 60 economies are untouched, which makes origin rather than ownership the operative fact. Goods entered before the effective date are governed by the rate in force at entry. Each of those is a question to put to the text, not an assumption to build a quotation on. The band itself carries information: a range of 10% to 12.5% means somebody chooses inside it, and the criteria for that choice are the part you most need to read.

The chain runs from the record to the entry. The hearing closes a record, a determination follows, the instrument is published with a scope and an effective date later in July, entries after that date carry the duty, and liquidation with the cash arrives months later. Notice where the discretion sits: in the determination and in the scope, both still open, which is why the diplomatic pushback matters procedurally even though it will not move a single rate.

Layered by who feels it, the split follows documentation depth rather than size. A US importer of Mexican, Peruvian, Guatemalan, Ecuadorian or Korean goods with a compliance function and a supplier audit trail absorbs this as administrative cost and can defend its own classification. A mid-size importer with the same goods and no audit trail absorbs it as an unquantifiable exposure, because it cannot say which purchase orders are in scope and must assume all of them are. An exporter in one of the 60 economies feels something else again, an order deferral rather than a duty, because the buyer pushes the risk back to origin by asking for paperwork before it asks for a price. Importers with small exposure relative to landed cost, and importers with short auditable chains, are close to indifferent.

The effective date is published and the entry date governs. Cargo that arrives before the effective date and is entered after it does not get the old treatment; cargo entered before does, whenever it clears. So the deadline for an import team is not arrival and not sailing, it is the date the entry goes in, which in the last week of any tariff window is a broker-capacity question. The asymmetry runs both ways: the range is published before the criteria are, so every July quotation carries a hedge that cannot be priced, and every buyer asking for a firm price in July is asking you to underwrite a determination you have not seen.

The part the coverage has not touched follows from the number of economies rather than from anything said at the hearing. A 60-economy scope is an admission that the authority cannot name facilities at scale. Sixty countries cannot be audited factory by factory, so the measure legislates against the flag instead. A measure that legislates against the flag does not create a compliance market, it creates a documentation market. The effect is distributional and sharp: inside one country, exporters with clean records take share from exporters without them, because the buyer's cost of believing them is lower. Economies outside the list become candidates for minimal transformation, which is the pattern the next instrument will be written to close. That last sentence is a forecast about the shape of the next text, not a finding about this one, and I mark it as such.

A small account, with my assumptions rather than the article's numbers. Take a US importer with two million dollars a year of FOB purchases from one of the 60 economies, duty currently at zero on that line, no ability to change origin inside the year. At 10% the annual duty is 200,000 dollars; at 12.5% it is 250,000. The band alone is 50,000 dollars a year, about 4,000 a month, of cost that cannot be quoted forward. If net margin on that book is 6%, which is 120,000 dollars, the band is roughly 42% of the margin. That is the arithmetic that turns a range into a decision, and the reason the criteria matter more than the rate.

Run the counterfactual both ways. Treat the hearing as the decision and you reprice in July on a measure that may be narrowed, delayed or stayed, then hand your buyer an increase you cannot defend when the text lands. Treat the range as noise and you enter cargo in the last week of July at whatever rate applies, then discover at liquidation that your product fell on the wrong side of a subheading. The second error is worse, because the first is a commercial loss that can be renegotiated and the second is a legal position that arrives with interest attached.

The trap is assuming the measure needs your goods to actually involve forced labour. It does not. It needs the authority's finding to hold, and your inability to rebut it. A measure of this shape is answered with documents, not with facts, which is uncomfortable and also true.

All of this holds while the scope remains economy-wide and the effective date remains prospective. If the final text narrows to named facilities, traceability cost falls and compliance cost rises, and everyone who spent July building country-level files spent it on the wrong file. If a court stays the measure before the effective date, the exposure disappears and the paperwork cost stays. Read the scope clause before the rate column, and read both before you sign a price for the fourth quarter.

↑ Back to the key points

— By Dr. Ingrid Voss