Trade policy built for containers is missing a trillion dollars of exports

American trade policy is argued almost entirely in the language of goods. Schedules, lines, rates, origin, ports, containers. The vocabulary is precise and it is old, and it describes something that can be stopped at a border and counted. Meanwhile the Bureau of Economic Analysis records services exports at $1.2 trillion, and virtually none of the apparatus applies to them.

That is the argument I want to make, and it is narrower than it sounds. I am not claiming services matter more than goods. I am claiming that the instruments and the scorekeeping have drifted out of alignment with the exposure, and that a negotiation conducted entirely in tariff terms will trade away things nobody priced.

A line the schedule cannot reach

Set the figures next to each other carefully, because they come from different places. Total exports of goods and services stood at $3.32 trillion in 2025 on the annual series, up from $2.57 trillion in 2021. Services exports are recorded separately at $1.2 trillion. The total series is titled goods and services, so services sit inside it rather than beside it, and the two are compiled on different bases — I will not difference them or take one as a share of the other. What the pairing establishes is only an order of magnitude, and that is enough: a trillion-dollar line is not a rounding error on anything.

Now ask what a tariff does to it. Nothing. There is no customs entry for an engineering review, a licensing payment, a reinsurance contract, a cloud instance or a semester of tuition. The instrument set that trade policy spends its political capital on does not touch this trade at all, in either direction. It cannot protect it and it cannot be used to retaliate over it.

What does reach it is a different list, and it is one that rarely makes a headline: professional accreditation and mutual recognition, data localisation rules, licensing and market-access conditions, standards bodies, procurement eligibility, and visa policy. Those are the tariffs of the services economy. They are administered quietly, by regulators rather than at borders, and they can close a market without a single announcement that looks like a trade action. They also tend to sit with ministries that are not in the trade negotiation, which is part of why they survive it intact.

The objection this has to answer

The strongest reply is blunt and I think it is largely right. Services are not strategic in the way that chips, minerals and munitions are. You cannot build a hull out of a consulting engagement. A services surplus does not qualify a supplier, does not shorten a lead time, does not employ a machinist, and does not give a defence programme a second source. When an adversary restricts a mineral, the shortage is physical and immediate; when a partner makes life harder for an American professional services firm, the firm is inconvenienced and the country is not disarmed.

All of that is true, and it is the reason economic-security policy is properly weighted toward physical supply. I would not reweight it. Scarcity of a physical input is a category of risk that no amount of services income substitutes for, and a brief that pretended otherwise would deserve to be ignored.

But the objection answers a claim I am not making. The question is not whether services can stand in for industrial capacity. It is whether the country is defending and negotiating over its actual exposure. A trillion-dollar export line that is vulnerable to instruments nobody has costed is exposure, whether or not it is the most strategic exposure. And the asymmetry runs the wrong way: a partner can impose real cost on it through a licensing rule that attracts no attention, while the retaliation available in return is drawn from a schedule of goods.

What this data cannot show

I have one number. It is a single latest-year value, with no breakdown into travel, transport, finance, professional services, royalties or education, no partner detail, and no history. That is why this page carries no chart. One figure drawn as a picture would suggest a trend or a composition I cannot show you, and I would rather print the number in a sentence and tell you what stands behind it.

The absence is itself part of the argument. The goods trade is measured to a granularity that supports whole careers of analysis. The services trade, on this evidence, resolves to one scalar. It is difficult to defend what you can barely see, and the measurement gap and the instrument gap are the same gap viewed from two sides.

So the practical ask is modest and specific. In any negotiation conducted over tariff lines, require an accounting of what is being conceded on licensing, recognition, data and visas, and require it in the same document. Those concessions are usually cheap to give in the room because nothing on the table measures them. On this evidence they are not cheap at all, and the bill arrives somewhere the trade press does not look.