The Tariff Made a Purchasing Decision Into a North American Content Decision

Mexico put a tariff on tool steel in December 2025. Most of it landed on the wire rod instead.

The reform that took effect on 1 January 2026 is the largest single tariff action Mexico has taken in decades. The Senate approved duties of up to 50 per cent on 10 December 2025 by 76 votes to 5 with 35 abstentions; the decree ran in the Diario Oficial three weeks later; the schedule covers 1,463 tariff lines, most of them at 25 or 35 per cent and a minority at the full 50, aimed at China, India, South Korea, Thailand and Indonesia.

Read line by line off the 2026 Chapter 72 schedule, and the shape of it changes. Every tool-grade long product in the chapter sits at 35 per cent. The 50 sits on commodity flat and wire rod: the products Mexican mills already make in volume. Mexico has no tool steel mill. It imports more than 90 per cent of the moulds and dies it uses, and it imports the steel those moulds are cut from.

Which means the reform, applied to this corner of the trade, does the opposite of what the headline says. It raises the cost of the input above the cost of the finished article. A mould shipped into Mexico from Michigan or Guangdong clears at zero. The H13 bar that a Mexican tool shop would need to cut that same mould from pays 35 per cent.

What the schedule actually says

There is no ambiguity to argue about here. The rates are published:

The 50 per cent is not a fiction. It is simply aimed at a different product. The line that carries the tool rooms, 7228.30.01, carries 35.

The inversion, in one quote

Jorge Ayala of the AMMMT, the Mexican mould and die association, put the problem plainly in September 2025, three months before the decree: finished moulds enter Mexico without paying tariffs while the steel to make them faces 20 per cent.

The reform widened that gap rather than closing it. It lifted tool-grade bar from roughly 20 to 35 points and left the finished tool where it was, at zero. A tariff that taxes the input above the output is not protection in a country with no mill. It is a tax on the version of the industry that employs people inside Mexico.

The scale of what is being taxed differently is easy to understate. Mexico is the second-largest importer of moulds and dies in the world, with consumption calculated at USD 5.0 to 6.0 billion. In 2024 it imported USD 1.72 billion of injection moulds, up 20.6 per cent, and USD 1.052 billion of dies, up 55.1 per cent. That is USD 2.77 billion of finished tooling, against a tool-grade steel import market of USD 142.8 million: the finished goods trade is roughly nineteen times the size of the raw material trade beneath it.

Around 1,800 foreign suppliers serve those buyers. The AMMMT counts more than 300 tool shops in Mexico and thinks the country could carry 1,000. INA, the national industry association, has put an import-substitution programme at USD 7.0 to 9.0 billion over five to ten years, with a new tool shop costing about USD 12 million. Every one of those numbers is an argument for making tooling in Mexico. None of them is an argument for putting 35 per cent on the bar.

Who actually ships it, and how little the duty reaches

The other reason the tariff lands differently than reported is who supplies Mexico. Of the 2024 flow through the alloy bar basket:

72.5 per cent of it is North American. US and Canadian steel moves under the USMCA and is untouched by the new schedule. Japan and Germany enter preferentially under their own agreements. China, the country the reform is named after, is 8.2 per cent of the basket.

Put the two filters together and the arithmetic is small. Take a 16 per cent dutiable slice of a USD 142.8 million market and apply 35 per cent: about USD 8.0 million a year of duty. Apply 50 per cent instead and it is USD 11.4 million. Not a wall, not an exclusion, not a reshaping of supply. A number in the low single-digit millions, small enough to be absorbed at the border, passed through to the customer, or designed out by buying the finished mould instead.

All three are happening, which is why the interesting quantity is not the duty but the origin file. Under the new regime a Mexican buyer has to be able to show where the steel was melted and poured, not merely where it was last cut. That paperwork, not the rate, is what will decide which suppliers keep the account.

The number under the number

Anyone sizing this market has to make one honest choice, and it is worth stating because the wrong version of it looks better than the right one.

The customs line that carries tool-grade bar, HS 7228.30, is not a tool steel line. It is an alloy bar basket that carries 4140, 4340 and 8620 alongside H13, D2 and the high-speed grades. In 2024 it was USD 593.0 million; in 2025, USD 571.0 million. Calling that figure the tool steel market would overstate it by a factor of four.

At a 25 per cent tool-grade share the market is USD 142.8 million in 2025, rising to USD 203.0 million in 2033 at 4.5 per cent. The band is USD 180.8 million at 3.0 per cent and USD 227.5 million at 6.0 per cent, a spread of 1.26 times. The share assumption is stated with its sensitivity: 20 per cent gives USD 114.2 million, 30 per cent gives USD 171.3 million.

Cross-check it against the world. The global tool steel market is USD 4,906.0 million in 2025. A Mexican market of USD 142.8 million is 2.9 per cent of it, falling to 2.7 per cent by 2033 as the global market compounds faster. That is a plausible weight for the twelfth-largest economy. A market of USD 571 million would have made Mexico 11.6 per cent of global tool steel consumption, which no reading of the industry supports. The basket is a throughput figure. The market is what stays.

Why 4.5 per cent and not 8

The forecast is a blend of two forces pulling in opposite directions, and the rate is the residue of them.

Pulling up: Mexican mould and die consumption is growing at double digits in places, imports of injection moulds rose 20.6 per cent in 2024, and more than 90 per cent of what the country uses is imported. Every new tool shop is a new buyer of bar.

Pulling down: the tariff is designed to substitute. If the INA programme works even partially, some of that demand is met by domestic capacity or by finished tooling that enters at zero instead. A 35 per cent duty on the input, against a zero duty on the output, actively rewards the substitution that removes tool-grade bar from the import series.

Four and a half per cent is roughly what you get when a market growing in the high single digits is partly met by something other than imports. It is a view about tool shop formation and tariff design, not about steel demand in the abstract.

The falsifier

Two things would break this argument, and both are readable inside one year of trade data.

If the DOF adds 7228.30.01 to the 50 per cent list, the premise that tool grades were deliberately left at 35 collapses. And if reported 7228.30 volumes fall by more than 15 per cent in 2027 while mould and die imports keep growing, then substitution is working and the 4.5 per cent is too low rather than too high. Watch the ratio, not the rate.

The tariff is real. It is 35 per cent on the bar, zero on the mould, and about a sixth of the trade. For a Mexican tool room the decision it creates is not whether to pay it. It is whether to keep buying steel, or to stop being a tool room and start being an importer of finished tools. That is a strange thing for a protection measure to ask.