Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- By 8 June 2026, complete a line-level register of every open aluminum, steel and copper entry, carrying tariff line, country of origin, entry date and the instrument cited on the entry summary, at 100 percent coverage, before any rate assumption enters a forecast.
- By 10 June 2026, one day after the hearing, obtain counsel's written answer to three questions: which entries fall inside the scope of any order, which are already liquidated, and what deadline attaches to each. No refund may be booked as a receivable until all three answers are in writing.
- By 12 June 2026, split the IEEPA payment ledger into paid and processed, paid and not processed, and not yet determined, and apply no percentage factor to any bucket; if a fifty-one percent factor already sits in the plan, remove it and mark the gap instead.
- By 15 June 2026, obtain the operative text of the tariff adjustment, proclamation, annex and publication date, and mark every covered line with its effective entry date; no rate movement may enter third-quarter planning before that text is on file.
- By 30 June 2026, re-run the third-quarter scenario under both branches of the 9 June outcome and require the plan to hold in each; a plan that only works if the refund arrives is not a plan, it is a forecast with a signature on it.
The White House issued a presidential action in early June further adjusting tariff regimes for imports of aluminum, steel, and copper into the United States under Section 232 authority, continuing the administration's active trade policy posture. The action comes ahead of a critical June 9 federal court hearing that will determine whether the U.S. government must issue full retroactive IEEPA tariff refunds — with only $85 billion of $166 billion in potential refunds having entered processing so far according to U.S. Customs data. The twin developments underscore the persistent legal and regulatory volatility surrounding U.S. trade policy that shippers must navigate in their Q3 planning.
The Analysis
Two documents are being reported as one story. In law they are not one story.
One is a presidential action issued in early June, described as further adjusting the tariff regimes for aluminum, steel and copper imports into the United States under Section 232 authority. The other is a federal court hearing on 9 June that will decide whether the government must issue full retroactive refunds of IEEPA tariffs. Reported together they produce a single word, volatility, which is being asked to carry more weight than it can bear. One instrument looks forward, names products and sets rates. The other looks back, and concerns money already collected. They attach to different entries, different goods and different periods.
The threshold question with the presidential action is not what it signals. It is what it does. Further adjusting presupposes regimes that already exist and continue, with modifications. It is not a reset, and anyone treating it as a clean slate has the wrong starting point. What determines exposure is the operative text: which tariff lines are named, how countries of origin are treated, what the rate is, and from which entry date it applies. The summary in front of us supplies none of those four. Strictly speaking, a headline saying the regimes were adjusted tells you the direction of travel and nothing about the price. I will not convert it into a number, because the number would be mine and not the document's.
The boundary of an action of this kind is drawn by its annex and its scope language, not by the three metal names in the title. A product that contains aluminum is not thereby a product covered by the aluminum regime. Whether a given derivative sits inside or outside is a drafting decision, and drafting decisions are where most of the real exposure lives. The same caution applies to exclusions. I am not asserting the terms of any exclusion process here; the point is that the question has to be put to the text, and that an exclusion's own expiry date is often what moves a company's cost, rather than the adjustment that made the news. If you have been treating an exclusion as permanent, that assumption deserves a date written on it.
The refund figure is where I would slow down further. Eighty-five billion of a potential one hundred and sixty-six billion has entered processing, according to the Customs data cited in the release. As a fraction that is about fifty-one percent, and that fraction is going to be used badly this quarter. Entered processing is an administrative state. It is not approved, it is not paid, and it is not a legal determination of anything. The eighty-one billion that has not entered processing is not a queue of approved claims waiting for a cheque. Nothing in front of us says what it consists of, how much of it has been claimed, or how much of it is claimable. A ratio whose numerator and denominator measure different things is not a recovery rate. It is a throughput rate, and it describes how fast an office is working rather than how much you are owed.
Who is exposed, precisely. Importers whose entries fall on covered lines after the operative date carry the tariff question. Importers whose lines sit outside the annex do not, whatever the title of the action suggests. Separately, anyone who paid IEEPA duties carries the refund question, and that population is defined by entries and payments rather than by product. These are different sets of companies, and frequently different departments inside the same company, which is why the two questions get merged in planning even though they are not merged in law. The timing is asymmetric too, and this is the part that gets flattened into a single calendar. The 9 June hearing date is fixed and readable. The operative date of the tariff adjustment runs off publication, and until the publication is in hand you do not have a date at all. The clock you can read is not the clock that moves your money.
Reporting the two developments as twins invites a netting instinct, and the netting is not supported. The instinct runs like this: the refund will offset the tariff, so the exposure is broadly flat. Put the two instruments side by side and the flaw is plain. A refund attaches to an entry already made and a duty already paid. A tariff adjustment attaches to an entry not yet made and a duty not yet owed. One is cash returning from a past period. The other is cost going out in a future one. Both may be large. They do not net, and a plan that assumes they net has replaced two questions with one answer that nobody gave.
The same error appears at smaller scale inside the refund number itself. Assume your company paid two point four million dollars in IEEPA duties, purely for arithmetic. Fifty-one percent of that is one point two two million. If that figure is already sitting in your third-quarter plan as a probable recovery, you have assumed a pro-rata distribution rule applied entry by entry across a population you have not seen. No document in front of us states such a rule. The arithmetic is right and the premise is invented, which makes the answer worse than no answer.
Suppose the court orders full retroactive refunds on 9 June. The correct conclusion is narrow: the obligation question is answered. The amount, the population of entries falling inside the order, and the timing of payment are separate questions, and each can fail independently of the others. Suppose the court does not. The eighty-one billion does not become zero. It becomes contingent on appeal, and the appeal runs on its own calendar, which is longer than a quarter.
What would make this reading wrong. If the operative text of the tariff action turns out to name the same tariff lines, the same countries of origin and the same entry dates as the population the court is examining, then the two instruments stop being separable in practice, the netting question becomes real, and the argument I have just made weakens considerably. That is a factual question about documents I have not read, and it can be settled in an afternoon by reading them. Until it is settled, the boundary I would hold is this: two legal bases, two directions of cash, one calendar that is fixed and one that is not. Change any one of those three and the conclusion changes with it.
↑ Back to the key points