A trade agreement is worth the terms it fixes and not the exports it is credited with

The case for a trade agreement is usually made in the currency of an economic development strategy. A market opens, exports rise, and the jobs that make the exported thing are counted at home. Ministries budget against that chain. Development offices put the agreement in the same column as a site, a grant and a training programme, and score it the same way.

That chain is mis-specified at its first link, and it is worth saying so plainly. An agreement changes the terms on which trade happens. It does not change the capacity to produce for export and it does not create the demand that would use it. What an agreement is worth to a development strategy is the durability of the terms it fixes, which is a real good and worth paying for. It is not export growth, and writing it in as export growth puts the wrong asset on the wrong line.

The line was going up anyway

Exports of goods and services ran 2.57 trillion dollars in 2021, 3.04 in 2022, 3.07 in 2023, 3.22 in 2024 and 3.32 in 2025. Five readings, four rises, and a total gain of 0.75 trillion dollars, which is 29.2 percent of where the series started.

Inside that total the four steps are 0.47, 0.03, 0.15 and 0.10 trillion dollars. The first is larger than the other three combined, which come to 0.28. Put another way, 62.7 percent of five years of export growth arrived between 2021 and 2022. Whatever produced the shape of that line had mostly finished by the second reading.

Nothing in this file holds an agreement. There is no schedule here, no rule of origin, no date of entry into force, and no export series for the country that signed nothing. What the file holds is a total that went up at every step available. A total of that shape is always available to be claimed by whoever was in office while it rose, and that availability is the defect in the usual case rather than evidence against the instrument.

How much of the economy is in play

Set the exports against the output they come out of. Gross domestic product in current dollars was 30.8 trillion in 2025. Exports of goods and services in the same year were 3.32 trillion. The quotient is 10.8 percent. The Bureau of Economic Analysis publishes its own share series, carried on this site at zero decimals, and that reads 11. Two different computations, landing in the same place.

A tenth of output is the whole exposure, before any agreement narrows it further. An agreement covers named partners and named lines, so the share it can reach is some fraction of that tenth, and this file does not hold the fraction. A strategy that expects an agreement to move its headline numbers is asking a lever attached to a tenth of the economy to move the rest of it.

The price nobody negotiated

Tariff schedules are negotiated line by line over years. The exchange rate re-prices every line at once and asks no one. The broad dollar index, set to 100 in January 2006, read 117.8 in 2020, 113.1 in 2021, 120.7 in 2022, 120.3 in 2023, 122.9 in 2024 and 122.7 in 2025. Between the low of those six and the high there are 9.8 index points, 8.7 percent of the 2021 reading.

Each of those points is an average of that year's daily observations, so 9.8 points is the distance between two annual averages rather than the widest move the index made; the wider figure is not in this file. An index of this kind records the dollar against a trade-weighted basket of currencies. It is not an export price and it is not converted into one here. For the argument it is enough that a number of that size moved while no one was at a table moving it.

The constituency the case is made to

Agriculture is where the case for an agreement is put most directly, to growers who can name the market they are being promised. U.S. agricultural exports were 196 billion dollars in 2022, 175 billion in 2023, 177 billion in 2024 and 175 billion in 2025. The fall from the first reading to the second is 21 billion dollars, 10.7 percent. The three readings since sit within 2 billion of one another, and the most recent is identical to the one two years before it.

Four annual observations settle nothing about cause. Prices, harvests, exchange rates and the purchasing decisions of governments abroad all sit inside that number and none of them is separable here. The narrower point is harder to avoid. In the sector where agreements are sold hardest, the published series is flat and below where its window opened, and anyone pointing at it to score an agreement is pointing at a figure that cannot carry the weight.

What an agreement is actually for

The defensible case runs through duration rather than volume. A plant takes years to build and longer to pay back. What kills that investment decision is rarely the tariff itself; it is the prospect that the tariff will be something else by the time the plant is running. An agreement is a written commitment that a rule of origin, a tariff line and a dispute procedure will still be there when the asset is commissioned. That good is real, it is scarce, and the six years above argue for its scarcity rather than against it.

Priced that way, an agreement belongs in a development budget as insurance on a long-lived asset. Insurance is worth its premium and it does not appear as revenue. An office that writes an agreement into its forecast as additional exports has booked a contingent benefit as a cash flow, and when the exports arrive or fail to, it will have no way of saying whether the instrument did anything at all.

The objection worth answering

The strongest reply is that these aggregates are the wrong altitude. A single tariff line can decide whether a firm exports at all, and that firm's experience is real whether or not it registers at 10.8 percent of output. I accept this without reservation. A line that matters absolutely to one exporter is invisible in a national total, and nothing above says otherwise.

Accepting it also fixes the scope of the disagreement. The claim under attack here is the aggregated one: the agreement presented as a source of national export growth and scored on a total that would have moved regardless. At the level of the firm the agreement does roughly what its advocates say, and the evidence for it is a contract rather than a series. Development strategies should make the case they can support and stop making the one they cannot.

This page carries no chart. The argument runs across an export total, a share of output, a currency index and an agricultural series of four readings, and no single picture holds them without implying a relationship none of them establishes. What is left is short enough to sign. An agreement fixes terms. Terms are worth money to anyone financing an asset that outlives a political cycle. Export growth is not the receipt, and a strategy that treats it as one has given up the ability to tell whether the instrument worked.