Where a thing is from

Not where it shipped from

A shipment arrives claiming a preferential tariff under a free-trade agreement, the reduced or zero duty that two countries grant each other's goods. The whole claim rests on one question: is this good actually from the country it claims to be from? And the surprising part, to anyone who has not met rules of origin before, is that the answer has almost nothing to do with where the good was shipped from. A product can be exported from a country and not originate there in any sense that customs recognizes. Origin is not a location in a logistics sense. It is a status the good earns through how and where it was made.

This is what rules of origin determine, and they exist because without them every preferential tariff would leak. A good made in a country outside the agreement could be routed through a member country and relabeled, capturing a benefit it was never meant to have. Rules of origin are the test that stops that, and the test is intricate.

Origin is a qualification, not a threshold

Here is where origin is a genuinely different shape of rule from the ones in most domains. A deadline is an event and a window. A disclosure is a threshold and a form. Origin is a qualification: a derivation that asks whether the good, given how it was produced, meets the criteria to count as originating.

The simplest case is wholly obtained: a good entirely grown, mined, or produced in one country plainly originates there. The hard cases involve goods made from inputs that came from several countries, and there the agreements use substantial transformation tests. One common form is a change in tariff classification: if the inputs enter under one tariff heading and the finished good leaves under another, a substantial transformation has occurred and the good qualifies. Another is regional value content: if a sufficient percentage of the good's value was added within the agreement's region, it qualifies. Many goods are tested against both.

{
  "rule_id": "ORIGIN-RVC-001",
  "title": "Regional value content below the agreement threshold",
  "source": "FTA rules of origin; WTO Agreement on Rules of Origin",
  "severity": "block",
  "expected_outcome": {
    "action": "review",
    "message": "Regional value content for this good is below the threshold the agreement requires for preferential origin. The good does not qualify for the preferential tariff on this basis. Confirm against any alternative tariff-shift criterion before claiming preference."
  },
  "conditions": [
    { "type": "numeric_comparison", "path": "good.regional_value_content_pct", "operator": "<", "target": "agreement.rvc_threshold_pct" }
  ],
  "deterministic": true,
  "validation_status": "expert_reviewed"
}

The comparison itself is simple once the regional value content is computed, and computing it, tracing the origins and values of every input, is the hard upstream work, the normalization of this domain. But notice the rule is one branch of a qualification. A good can fail the value-content test and still qualify under a tariff-shift criterion, so the real determination is a structure of alternative tests, any one of which can confer origin. That is the qualification shape: not a single line, but a set of paths to the same status.

Fail closed, because preference is a claim

A preferential tariff is a claim a importer makes and a customs authority can audit, with duties and penalties owed if the claim was wrong. So the determination cannot be a guess. When the inputs to the origin test cannot be resolved, when the value content or the tariff classification of components is unknown, the engine does not conclude that the good qualifies. It declines to certify preferential origin and flags the claim for review. An origin you cannot establish is not an origin you can claim, and treating an unproven origin as a qualifying one is exactly the leak the rules exist to close.

The point

Rules of origin matter here because they show the engine handling a shape beyond thresholds and deadlines. Origin is a qualification: a derivation over how a good was made, tested against alternative criteria, any of which can confer the status. The arithmetic at the end is simple. The structure of paths to qualification, and the work of resolving the inputs each path needs, is where the domain lives. And the discipline holds as everywhere else: a preferential claim that cannot be established is not waved through on the assumption it would have qualified. Where a thing is from is a status it has to earn, and the rule is the test of whether it did.