De Minimis as a Live Control
What a 2006 ice-cream ruling still teaches trade compliance about the tolerance rule, its carve-outs, and where origin exposure actually sits.
By Julio d’Arbelles, Global Trade Compliance, linkedin.com/in/juliodarbelles
The first origin file I read cover to cover was a dispute over ice cream. It was 2008, the ruling was already two years old, and I was young enough to read the entire record without knowing which cases would matter. That one did. It is where I learned de minimis before the term had become common vocabulary in the region, and I have carried its logic into every origin determination I have run since. The rule reads like a footnote in the origin chapter. In practice it is a value-based control that has to be managed month to month, and its edges are where most of the exposure lives.
The baseline, and the tolerance
De minimis only makes sense against the baseline, so start there. Origin is not conferred by where a good is assembled or packed. It is conferred by transformation, usually expressed as a required change in tariff classification. Joining finished components, or moving a finished product into a new box, does not create origin. The rule is deliberately strict, and precisely because it is strict, the drafters left a tolerance: a small quantity of non-originating material that fails the required classification change can be disregarded, up to a set percentage of value.
The ice-cream case is the cleanest illustration I know. A milk-based ice cream carried non-originating powdered milk at between 2.78 and 3.8 percent of the good’s value. The milk was the only input that failed its required tariff shift. Article 11 of the Central American origin regulation set the de minimis tolerance at seven percent, so on its face the product qualified. The importing administration resisted on two grounds worth remembering, because both recur in practice. First, that once the milk was transformed into ice cream it had already changed classification, making de minimis irrelevant. Second, and more aggressively, that the calculation should aggregate every non-originating material in the good, not only the milk, which pushed the figure well past the threshold.
The Council of Ministers settled the method in Resolution 137-2005: de minimis is calculated only on the materials that fail the required classification change, not on the total foreign content of the good. On 13 June 2006 the tribunal held that the tolerance applied, that the goods were originating, and that the measures denying them were incompatible with the treaty, with sixty days to withdraw. The detail I keep returning to is that the tribunal relied on an almost identical earlier case, a milk-based ice cream from Honduras already held originating on the same reasoning. The principle was settled before the dispute began.
Thresholds, and the exclusions that matter
That seven percent was not a regional peculiarity. It was NAFTA’s figure. The USMCA raised the tolerance to ten percent of the value of the good, or of total cost, in Article 4.12. CAFTA-DR runs its own de minimis at ten percent of the adjusted value, in Article 4.6. The European Union’s origin protocols apply a tolerance in the same range. A practitioner who understands the number from one instrument understands it across most of them, which is exactly why the mechanism deserves more attention than it usually gets.
The exclusions deserve more attention still. De minimis has never been available for everything. The lists, from NAFTA forward, carve out dairy for goods of Chapter 4, certain sugars in Chapter 17, and, in a detail that matters to anyone working Central American agri-food, the coffee bean going into instant coffee, along with specified fats and oils. The pattern is not random. The tolerance runs generous toward an industrial component and closes hard on the commodities a producing country is expected to actually produce. If your portfolio sits in dairy, sugar, or coffee, the safe assumption is that the window is shut and the qualifying rule must be met on its own terms.
Automotive goods are the case practitioners cite most and rely on least. A passenger vehicle has to clear a regional value content of seventy-five percent, a labor-value-content test, and steel and aluminum sourcing requirements, and that regime is under negotiation again in the 2026 USMCA joint review. A ten percent tolerance does not rescue a good from an architecture like that. The lesson is not that de minimis is trivial for autos, it is that the tolerance is only ever one instrument in a stack, and the stack is what decides.
Managing the tolerance
Because de minimis is a calculation of value, it moves. A supplier price rises, a currency shifts, and a material that sat comfortably under the threshold in one quarter brushes against it in the next, with no signal on the plant floor. The teams that stay compliant treat the tolerance as a monthly control, re-run against the current bill of materials, rather than a determination made once and filed. A procurement decision to substitute a local input for a cheaper foreign one is an origin decision, and it belongs in front of compliance before it is executed. A useful discipline is to read your own import statistics as an early-warning system: when a material begins arriving in volume for a plant and it is not one you were tracking, that is a question to walk over and ask.
The consequence of getting it wrong is not abstract. Under CAFTA-DR and the USMCA, the importing authority can verify origin after entry, reach the producer and the underlying records, and, if the arithmetic does not hold, deny the preference retroactively and assess duties and penalties against the importer. The exposure lands on the customer, rarely on whoever changed the formula.
The valuation trap
A second-order trap catches sophisticated operations. De minimis is a fraction: non-originating value over the value of the good. When a product drifts over the line and the non-originating input cannot be reduced, the arithmetic invites raising the denominator, the price of the good, so the same foreign value becomes a smaller share of it. Between related parties that move is usually unavailable. The transfer price is disciplined by the arm’s-length standard and often fixed in an advance pricing arrangement, and it cannot be adjusted because an origin ratio needs help. Customs valuation and transfer pricing reach for the same figure from opposite directions, and origin sits on top of both. Few functions in an organization can see all three at once, which is the practical argument for placing trade compliance where it can.
Claims, marking, and the sanitary registry
The confusion that does the most quiet damage lives on the label. “Originating,” “wholly obtained,” country-of-origin marking, and a marketing claim of “one hundred percent” are four different tests, and they do not always agree. A good can qualify as originating under an agreement while carrying a de minimis foreign component, which means it is legally originating and, at the same time, not wholly anything. A “100%” claim printed on that package asserts something the record cannot support. The sanitary registry compounds the error, because people treat it as a passport: a health registration authorizes sale in a market, it does not confer customs origin. Regulatory domicile and origin are separate questions, and when the label and the declaration tell different stories, a verifying officer eventually reads both.
The multilateral backstop that never arrived
None of this is a regional curiosity, and the clearest proof is the altitude at which the same argument has been fought. Whether a country’s origin rules measure origin or quietly protect an industry went to the WTO in DS243, where India challenged the United States over the non-preferential rules it applied to textiles under Article 2 of the Agreement on Rules of Origin. The panel confirmed that origin rules answer to that discipline, though it did not find a breach, and the report was adopted in 2003. The deeper point is what did not happen. The same agreement launched a program to harmonize non-preferential rules of origin, intended to take three years. More than thirty years on, it is unfinished. The most basic question in trade, where a thing is from, is the one on which the trading system has never reached a single answer, and absent a treaty rule you fall back to each country’s own non-preferential regime.
What the file taught
De minimis rewards the teams that treat it as a control and catches the ones that treat it as a year-end discovery. The tolerance is an honest admission built into the rules, that no product worth manufacturing is sourced entirely at home, and that a small measure of the foreign does not erase the local work around it. A file I read in 2008, over a tub of ice cream and a line of imported milk, taught me that much. The number was 2.78 percent. It was enough to start a dispute, and enough to teach a method.
Primary sources
- Guatemala v. Costa Rica, Central American Trade Dispute Settlement Mechanism, MSC-04-04, Final Resolution, 13 June 2006 (SIECA).
- COMIECO-EX Resolution 137-2005, interpretive guidance on de minimis calculation methodology.
- USMCA, Chapter 4 (Rules of Origin), Article 4.12, De Minimis.
- CAFTA-DR, Chapter 4 (Rules of Origin and Origin Procedures), Article 4.6, De Minimis.
- WTO Agreement on Rules of Origin, Article 2; WTO, United States, Rules of Origin for Textiles and Apparel Products, DS243 (Panel Report adopted 2003).
Download attachments: