Technical standards decide admission where a tariff only sets a price, and nothing in the record counts admission

Market access is usually argued as a question of price. A duty is set, a quota is fixed, a preference is granted, and the argument runs over what a good costs once it has crossed. Underneath that argument sits a prior question that is rarely negotiated in public and never counted: whether the good is admissible at all. A voltage, a frequency band, a test method, a tolerance, a data format, a safety threshold. Meet the specification and the market is open at whatever the duty happens to be. Miss it and the duty is irrelevant, because there is no transaction for a duty to apply to.

Where this is written from

The vantage decides what follows, so it belongs before the argument. I am an analyst writing inside The Cefa Brief, working from published statistics and public documents and from nothing else: annual series redistributed by FRED and retrieved for this site on 19 September 2026, together with the agencies' own accounts of what each series counts. I sit on no technical committee, hold no brief from any government or firm, and have read no test report, no conformity certificate and no file on a consignment that was turned back. That position sees the flows that cleared and the aggregates the agencies chose to publish. It cannot see a single decision that kept a good out, and on this subject the decision is the whole instrument.

What the position does well, it does plainly. It can say what a series counts, on what base, at what frequency and with what left out, and it will not assert anything the series does not carry. What it cannot do is reach the layer where standards actually operate. That layer is made of committee drafts, national deviations, accredited laboratories, recognition arrangements between regulators and decisions taken at a port, and none of it is aggregated into a public time series by any agency whose work appears in this brief. A reader who wants the count of goods refused on conformity grounds will not get it here, and should be careful with anyone who offers one without saying where it was counted.

A price and a condition are not the same instrument

A tariff is continuous. It moves a margin, and a firm with a cost advantage can absorb it, pass it on or route around it, and the decision stays commercial. A standard is a condition. It sorts the world into admissible and inadmissible, and no cost advantage moves a product from the second set into the first. The difference in shape decides who holds the instrument. A duty is imposed by a ministry and can be lifted by the same ministry in an afternoon. A specification is settled in a technical body over years, is referenced by the regulations of several jurisdictions at once, and outlives the government that helped write it. The slower instrument binds harder.

The surface where the instrument would show

Standards sit thickest where products have to interoperate, and the record does hold output series for exactly those industries. Computer and electronic product output, indexed to 2017 equal to 100, reads 105.7 in 2020, 110.3 in 2021, 117 in 2022, 116.6 in 2023, 118.3 in 2024 and 126.8 in 2025, a rise of 21.1 points across the window with 8.5 of it arriving in the final year. U.S. semiconductor production, on the same 2017 base, runs 121 in 2021, 131 in 2022, 135 in 2023, 144 in 2024 and 165 in 2025, a rise of 44 index points. Each point in both series is an annual average and needs no further qualifier. They record volume produced. Neither records whether a unit of that volume met a foreign specification, was tested against one, or was kept out by one.

Both legs fell and nothing attributes the fall

The bilateral series with mainland China show the difficulty in its sharpest form. U.S. goods exports to mainland China run 10 billion dollars a month in 2020, 13 in 2021, 13 in 2022, 12 in 2023, 12 in 2024 and 9 in 2025. Goods imports from China run 36, 42, 45, 36, 37 and 26 across the same six years. Both are stated in billions of dollars at a monthly average, so each annual point is the average of that year's twelve monthly readings and is not a total for the year. Exports are 4 billion dollars a month below their 2022 figure and 1 below their 2020 figure. Imports are 19 below their 2022 figure and 10 below 2020. Two legs of one trading relationship contracted at once, and the series carry no field that says why. A duty, an export control, a specification a supplier could no longer meet, a procurement rule, a change in demand, an assembly step moved to a third country: each would appear in these numbers in the same way, as a smaller number.

Advantage is taken where the text is written

Industrial advantage in standards does not come from enforcing them. It comes from having written them. Where a specification is drawn close to the practice a country's firms already follow, those firms are compliant on the day it takes effect and their installed base carries over into the new regime. Everyone else redesigns, retools, retests and recertifies, and pays for it in the same years they are supposed to be competing on price. That adaptation cost is real capital and real engineering time. It lands on a firm's accounts as ordinary product development, indistinguishable there from any other engineering spend, and it appears nowhere in the national accounts as a cost of foreign rule-making. The advantage is durable for the same reason it is invisible. It is embedded in a text that a later government cannot repeal, only renegotiate.

What follows for a policy that cannot count its own instrument

Exports were 11 percent of GDP in the latest reading this site holds of that measure, and exports of goods and services rose from 2.57 trillion dollars in 2021 to 3.32 trillion in 2025. Services exports stand at 1.2 trillion dollars, governed less by product specification than by whether a qualification, a licence or an audit is recognised on the other side. A material share of national output therefore has to satisfy a rulebook written in another jurisdiction, and the size of that share is the one part of this that carries a published number. What it costs to satisfy is not published by anyone. Policy follows what it can count. A tariff schedule yields a rate, a subsidy yields an award, an export control yields a licence decision, and each of the three can be reported to a legislature at the end of a year. Standards participation yields a committee seat held for a decade and a text that may not take effect until the seat has changed hands twice. It loses the argument for money and attention not because it works less well, but because it cannot be scored over the interval in which budgets are set.

So the conclusion has a hole in it, and the hole is the point. This piece can show that output in the two industries it cites grew, that both legs of the bilateral goods flow with mainland China fell, and that eleven percent of GDP depends on being admitted somewhere. It cannot show that any one of those movements was caused by a standard, because no series in this record separates admission from price, and asserting a mechanism the data cannot carry would be the worse error. What that leaves is a rule for reading rather than a finding. Treat participation in technical standards as a capital commitment with a decade-long payback, judge it on presence and continuity rather than on the trade figures of the following year, and distrust any claim about standards that arrives attached to an export number, including the export numbers above.