While neither Canada nor Mexico is willing to withdraw its steel and aluminum tariffs in its FTAs, they have been willing to offer various figleaf concessions towards their Fortress pitch.
This song and dance that began in the run-up to USMCA continued throughout the Biden Administration, and, per Secretary Ebrard’s Aug. 30 remarks quoted above, remain Mexico’s game plan.
Disastrous Failure No. 1: the 2019 Steel and Aluminum Agreements That Facilitated Import Surges, Circumvention, and US Plant Closures
In May 2019, USTR announced it had reached agreements on steel and aluminum with both Canada and Mexico. Under the agreements, American manufacturers were assured that there would be “aggressive monitoring” and that if covered metal imports “surged meaningfully beyond historic volumes of trade”, then the Section 232 tariffs would be restored.
The agreements also promised to “[p]revent the transshipment of aluminum and steel made outside of [North America].”
The agreements weren’t worth the paper they were written on.
Mexico violated promise to avoid surging exports comprehensively, as CPA has documented: by 2023, steel mill shipments ran 36% over baseline, Chapter 72 iron and steel imports were 134% over, and electrical steel conduit – the humble pipe carrying wiring through every American building – 472% over baseline, heading for nearly 700%, until Mexico controlled 87% of U.S. conduit imports.
It wasn’t just the surge, either, but also the transshipment, despite the “aggressive monitoring” promised in the agreement. The transshipment of Chinese steel and aluminum via Mexico into the United States continued, and without any tariffs or quota to control import penetration, the predictable result happened. American mills were shuttered as Mexico took the market.
USTR took note of the deal’s failure. On February 16, 2024, USTR issued a readout of a meeting with Mexico stating that “Ambassador Tai stressed the urgent need for Mexico to take immediate and meaningful steps to address the ongoing surge of Mexican steel and aluminum exports to the United States and the lack of transparency regarding Mexico’s steel and aluminum imports from third countries.”
Mexico once again managed to dodge tariffs, however, despite scoffing off every aspect of the 2019 Agreement.
Instead of tariffs, Mexico successfully sold the Biden Administration on the Fortress North America plan. Under this new plan, covered steel and aluminum imports could only be freely imported if the importer certified that the steel was melted and poured, or the aluminum was smelt or cast, in North America.
But North American rules of origin keeps American mills and fabricators in a price competition they cannot win. The rules are also fruitless because they remain critically undermined by scoping issues (discussed below), varying metal costs across the three countries, and fail to learn from the disastrous North American Automotive Rule of Origin.
Throughout this time period, U.S. production of steel and aluminum shrank.
For steel:
- In 2018 and 2019, the USGS recorded 86.6 and 87.8 million tons of raw steel production respectively.
- By 2024 and 2025, those numbers were down to 79.5 and 82 million tons respectively.
For aluminum:
- In 2018 and 2019, the USGS recorded 891,000 and 1,093,000 metric tons of American primary aluminum production respectively.
- By 2024 and 2025, those numbers were down to 676,000 and 660,000 metric tons of American primary aluminum production respectively.
Disastrous Failure No. 2: The USMCA Automotive Rule of Origin
On March 26, 2025, President Trump imposed the global 25% tariff that should have happened in 2019. The Proclamation noted diplomatically that the Automotive Rule of Origin had been a bust: “the revisions to the United States-Korea Free Trade Agreement and the United States-Mexico-Canada Agreement (USMCA), have not yielded sufficient positive outcomes. The threat to national security posed by imports of automobiles and certain automobile parts remains and has increased.”
The numbers don’t lie: In 2018, Mexico sent 2,663,805 cars to America while we sent only 140,560 to Mexico. By 2025, Mexico had increased its exports to 2,700,297 while our exports to Mexico had fallen to 126,221.
Put simply: Mexico sends us 20 cars for every 1 car we send to Mexico. And that happens because they have lower wages, and USMCA continued the NAFTA promise to automakers of unlimited U.S. market penetration for Mexican vehicles.
On July 1, 2026, USTR issued a report on the failure of the USMCA Automotive Rule. In the report’s Foreword, authored by Greer, acknowledged that:
“the United States has run a persistent, substantial, and growing trade deficit in autos and parts with Mexico for over 30 years. In 2024, that deficit peaked at $138.2 billion. In recent years, analysis indicates that the share of U.S. content within Mexico’s automotive exports to the United States has been declining, while the share of Chinese or other Asian content in those goods has been increasing.”
Regrettably, despite this acknowledgment, the Foreword indicated a doubling down on tougher ‘rules’: “Through strategic improvements, USTR seeks to raise regional and U.S. content, reduce reliance on inputs from third-countries, and build a more resilient and secure supply chain”.
Conjured-up academic rules which contemplate one nation’s entire industry being offshored so long as the competition was ‘fair’ should be a non-starter. Instead, USTR should be negotiating firm, outcome-oriented measures, like quota caps on the amount of vehicles imported. For example, in USTR’s May 2025 UK trade deal, rather than rules, the UK was put on a quota: 100,000 cars at a 10% tariff, 25% thereafter.
Disastrous Failure No. 3: Another USTR “Monitoring Agreement” Hollows Out US Capacity to Build Transformers, Highlighting Scoping Issues Around Steel and Aluminum
Grain-oriented electrical steel — GOES — is the specialty steel at the heart of every grid transformer. Exactly one company melts it in the United States, Cleveland-Cliffs at Butler, Pennsylvania; neither Canada nor Mexico melts a single ton.
But fabricators in Mexico can buy it duty-free from suppliers in Asia, and from 2018 until 2025 were able to export GOES products duty-free into the United States with only the mildest treatment.
This is due to scoping issues inherent in tariff actions involving primary metals. When we talk about the “steel” and “aluminum” tariffs, that does not automatically include every product containing those metals. So when Canadian and Mexican fabricators can source cheaper metal than American fabricators, and export their products to the United States duty-free, American fabricators get wiped out.
Transformers, which are critical equipment for our civilization’s existence, became a perfect case study.
While the 2018 steel tariff did cover GOES, it did not extend to the laminations and cores one step downstream of the metal.
The arbitrage assembled itself immediately: Japanese GOES was slit, stacked or wound — a transformation so minor that GOES remains 70–90% of the article’s value — and the components entered the United States tariff-free. GOES imports fell 56% in 2019 while import penetration of laminations reached 88%; since 2020, roughly 95–98% of both countries’ exports of transformer parts have gone to a single market, the United States. Neither country built a components industry. They built a tariff bypass.
In October 2020, as part of a Section 232 investigation into transformers and transformer components, Commerce formally found that “transformer laminations, stacked cores and wound cores are being imported into the United States in such quantities and under such circumstances as to threaten to impair the national security.” The remedy? No tariffs, but instead another monitoring agreement: Mexico would operate “a strict monitoring regime,” and in exchange Mexican transformer laminations and cores would “not be subject to any action” under Section 232. The monitoring regime then monitored the following: imports rising every single year, up 321% by 2024 over 2018, from $126 million to $527 million. It fell to the sole American producer to petition Commerce, in May 2025, to pull laminations and cores inside the tariff as steel derivative products — granted August 18, 2025, seven years into the steel 232 and five years after the government found the threat. And the coda: Census data now runs ten months past the inclusion, and imports have not fallen a dollar — roughly $40 million a month before, $40 million a month since, all of it duty-paid. The offshored capacity simply pays the toll, because Asian imports don’t flinch at a 25% tariff.
What played out with GOES applies to countless other downstream metal products still outside the scope of the steel tariffs: from boats to bicycles, from gun safes to golf carts.
Costly Fig Leaf Number 3: Canada and Mexico’s WTO-compliant steel tariff maneuvers
Mexico’s January 2026 reform raised MFN duties to 25–35% on household appliances, aimed squarely at China — the very instrument Washington is being asked to accept as its northern and southern wall. Six months of data later: Chinese shipments were down just 2.5%, and China’s share of Mexican appliance imports actually rose, to 68.1%, because the across-the-board hike crushed the second-tier suppliers (Thailand down 71%) while China absorbed the tariff through its cost advantage and CPTPP-preferenced Vietnam sailed through untouched. Meanwhile the same CPTPP door has been swinging in Canada: the Canadian appliance market grew 49% from 2016 to 2024 while the U.S. share collapsed from 42.9% to 29.8%, with Vietnam going from zero to $113 million — concentrated in large washers, Louisville’s flagship product. Haier builds washers at its plant in Biên Hòa, Vietnam; LG and Samsung supply Canada from Vietnam and Thailand rather than their U.S. plants. And the White House has celebrated Haier’s GE Appliances — a Chinese-owned manufacturer — as a showcase of American manufacturing. If a Chinese-owned OEM is the celebrated face of American appliance making, Ottawa and Mexico City hold a green light to court Chinese OEM plants of their own and call the output North American. A fortress where Haier is simultaneously the American champion in Louisville and the CPTPP importer from Biên Hòa is not a fortress. It is a franchise.
The tally: a restraint agreement, violated. A monitoring agreement that monitored a 321% increase. An MFN tariff wall that raised China’s market share. The 2026 “harmonization” pledge would be promise number four — the largest and least verifiable yet.
When the Sheinbaum government raised MFN tariffs to 25–35% on numerous steel and aluminum products, the decrees expressly preserved preferential treatment for treaty partners. A 35% “steel tariff” sounds formidable until one asks who pays it: in late August, Mexican domestic hot-rolled coil was assessed around $982–$989 per metric ton while Japanese HRC — expressly FTA-exempt — was offered into Mexico at $685–$700 CFR. One of the world’s largest steel producers ships qualifying material at zero through the wall’s dozens of doors. That is not a North American external tariff.
Harmonize to what, exactly? The U.S. steel action’s own scope has been redrawn six times in eighteen months: the February 2025 derivative annex; household appliances added that June; 407 more tariff lines — charcoal barbecues to transformer components — that August; an April 2026 proclamation that rebuilt the entire architecture into three duty tiers, removed products and terminated the inclusions process; and an August 2026 proposal to start adding again. A refrigerator is a national-security derivative in Washington and ordinary commerce in Ottawa; a bicycle that Washington expressly refused to cover carries a 35% tariff in Mexico City; gun safes are protected by no one. Three governments cannot harmonize schedules when they cannot agree on — and Washington cannot itself hold still — the list of products the fortress protects.
Nor does a promise to “consult” solve anything. Greer himself denied that the proposed text gave Washington a veto over Canadian trade deals. Without such authority, harmonization is inherently temporary: Canada or Mexico could negotiate a new preferential agreement tomorrow and open another hole in the perimeter. A true common external tariff is not a one-time concession; it is a continuing surrender of national trade-policy autonomy — which is precisely why no one will agree to it.
The “Fortress North America” Mirage: Why Tariff Harmonization Will Never Work
The collapse of the U.S.-Canada trade deal is providing a rare glimpse into the secret world of international trade negotiations.
Beyond just the insider view, we are being given hope that the mirage of “Fortress North America” can be put to rest.
From the day President Trump returned to office, Mexico has aggressively pitched the necessity of “Fortress North America”. One month into the Administration, in February 2025, Treasury Secretary Scott Bessent confirmed that Mexican leaders had proposed matching U.S. tariffs on China, repeating Mexico’s “Fortress North America” framing.
On May 9th, Canadian Prime Minister Mark Carney said “Like Mexico, Canada remains open to deeper integration, including options for ‘Fortress North America’ in selected sectors.”
By May 28th, in a fireside chat following his remarks at the Economic Club of New York, Carney said ““It’s our strong view, and we’ve been clear about this, a ‘Fortress North America’ in autos, in steel, in aluminum – that’s in everyone’s interest.”
Those three sectors were flagged by Carney because they were the three subject to high profile Section 232 national security investigations in President Trump’s first term.
The steel and aluminum actions resulted in nominal 25% and 10% global tariffs respectively in early 2018, but both Canada and Mexico were promptly awarded total exemptions. More exemptions would follow.
The Automotive Section 232 report, finished May 2019, offered the President three alternative recommendations: 25% tariff on all cars; a 35% tariff on SUVs/CUVs; or, forego tariffs and instead jawbone some USMCA ‘rules’ with Canada and Mexico on requiring a higher level of North American content before a car could become eligible for USMCA duty-free status. Assurances from automakers convinced the President to accept the latter.
The result was a heightened ‘rule of origin’ for Automotive imports in USMCA, considered the flagship improvement over NAFTA. It went so far as to include a headline-grabbing $16/hour wage requirement, although this would only ever apply to a minority of workers under an absurdly complicated formula. (More on this failure below).
The USMCA Automotive result was instructive to free traders in Canada and Mexico: avoid U.S. tariffs by pitching North American integration and heightened ‘rules of origin’.
Mexico isn’t giving up. On August 30th, even in the fallout of the Canada deal, Mexico’s Secretary of Economy Marcelo Ebrad assured his colleagues “El año que entra seguramente el debate internacional ya no van a ser los aranceles, van a ser las reglas de origen.” (Translation: “Next year, the international debate will likely no longer be about tariffs, but about rules of origin.”)
We’ll see about that.
Despite Carney’s earlier supportive remarks on tariff harmonization for steel, aluminum, and automotive, Carney either had a massive change of heart, or never understood the request to involve raising tariffs on those sectors for countries outside the ‘fortress’.
On August 22, announcing the collapse of the talks, the Prime Minister framed the U.S. tariff harmonization request as follows: “The U.S. introduced, in the last hours, efforts to restrict our ability to have other trade deals. Because we believe in free trade, we’re the partner of choice, in many respects, for countries around the world, and Americans wanted to restrict that, they had language to restrict that. Unacceptable.”
Greer was clearly exasperated following the Canadian pivot. “For a year and a half,” Greer told CBC, Canadian officials have been coming to Washington advocating a “Fortress North America” — “their term, not mine.” If the United States is going to lower its national-security tariffs for Canada, it expects its partners to “also protect their borders.” What USTR wants is “mutual protection”: “we can’t have Canada be a place where you can get flooded with steel and aluminum from other countries and be used as a back door into the U.S.”
Greer double-downed on calling out Canada in his CBC interview: “Fortress North America, you can say it, but when it gets down to brass tacks and you want to actually collectively protect this market and these critical sectors, they’re not gonna do it.”
Thank heavens. Greer was correct in his expectation, but it’s been clear for a decade that neither Canada nor Mexico would pull back tariff concessions in their other FTAs.
The “Fortress North America” pitch is understandable to anyone who grew up playing Risk, the classic boardgame. Consolidate the continent to unlock tremendous gains! But in the real world there is no prospect of a genuine North American tariff perimeter, and foregoing important sectoral tariffs for more ‘rules of origin’ was always going to be a terrible trade.
This article focuses on why substituting national home market U.S. tariffs for a common North American external tariff is a fruitless, counter-productive goal, and how tariff harmonization efforts to-date have proved immensely harmful to American manufacturing.
A single, European-style customs union for North America that encompasses various national currencies and legislatures is also an unworkable and bad idea, but those arguments are beyond the scope of this article.
First, Acknowledge What the Fortress Concedes
Before the political and technical flaws, we should acknowledge what the very idea concedes. The “Fortress North America” proposals from Canada and Mexico require the U.S. to concede that unlimited cross-border price competition with Canada and Mexico is worthwhile, even as workers in both countries earn less.
International free trade — cross-border price competition — is a race to the bottom, full stop. The jurisdiction with the lowest labor and environmental standards soaks up the investment.
Within a nation, federal preemption across policy spaces backstops the race to the bottom. Not just on labor and environment, but on policy areas from intellectual property to consumer protection, antitrust to anti-corruption.
Wherever production moves inside the United States, a common federal policy floor – girded by a real judiciary – moves with it.
No such floor follows a factory across the Rio Grande or the St. Lawrence.
For that reason alone, “Fortress North America” should be dropped.
But assume you still believe in unlimited cross-border price competition with Canada and Mexico. Even on its own terms, the fortress is a mirage that can never be built, and no U.S. jobs should be sacrificed in its pursuit.
The Emptiness of Fortress North America Claims Was Apparent Long Before Carney’s Explicit Rejection
Carney’s explicit refusal to harmonize steel and aluminum tariffs – calling the request “Unacceptable” – was a welcome and rare instance of candor to those who have watched Canada and Mexico’s simultaneous embrace of global free trade while touting a walled-off North America.
Greer’s “they’re not gonna do it” acknowledgment could have been confidently said at any point in the last ten years.
Many Americans remember the 2016 election fight over the Trans Pacific Partnership, a free trade agreement with eleven other Pacific-rim nations, including Canada and Mexico. During the campaign, then candidate Donald Trump said TPP was a “another disaster done and pushed by special interests who want to rape our country, just a continuing rape of our country.” He withdrew the U.S. from TPP as soon as he took office in January, 2017.
What is not widely known in the United States is that after U.S. withdrawal, the other eleven nations continued on with the deal, renaming it the Comprehensive and Progressive Agreement for Trans-Pacific Partnership, or “CP-TPP”. They did this while they were re-negotiating NAFTA into USMCA.
Canada and Mexico haven’t stopped driving CP-TPP expansion. Costa Rica and Uruguay are wrapping up negotiations to join, and the UAE, Philippines and Indonesia are in formal negotiations. In June 2026, the Sheinbaum government completed Mexico’s ratification of the United Kingdom’s accession to CP-TPP, allowing the country to become the 12th nation to join the bloc.
Vietnam and Malaysia are CP-TPP founding members of this pact.
A recent CP-TPP-wide analysis from New Zealand’s trade ministry reported that 93.6% of tariff lines were duty-free within the CP-TPP bloc, and that percentage will grow to 98.8% as implementation continues to phase-in.
How is such a far-flung, massive free trade regime compatible with a North American “fortress”? Someone needs to put that question explicitly to Carney and Sheinbaum.
Canada and Mexico Won’t Even Say No To China Joining CP-TPP
China aims to join the CPTPP as well. Earlier this year, China’s Commerce Minister, Wang Wentao, while discussing China’s trade strategy for the 15th Five-Year Plan, said China would “[c]ontinue advancing accession to the [CP-TPP]”.
On July 15th, MOFCOM’s China Free Trade Area Service Network reported that China was “steadily advancing” accession to the CPTPP.
Accession to CP-TPP requires consensus. Any single member holds a free, quiet, procedurally trivial veto. However despite their Fortress North America rhetoric, neither Canada nor Mexico has ever been willing to say no to China. Not once. Governments supposedly eager to wall China out of North America cannot bring themselves to pronounce a one-word veto over China’s application to their global free trade agreement.
Thus, for anyone who wants a simple test of the Fortress North America concept, ask: are Canada and Mexico willing to reopen their steel, aluminum and automotive concessions in the CP-TPP? If not, the conversation can end right there.
Refusing To Touch Other FTAs Further Negates Harmonization
While China’s accession to the CP-TPP is a simple test of political seriousness, an honest assessment would ask whether Canada, Mexico, or, for that matter, the United States, were willing to revisit tariff concessions to any other country.
When the United States re-affirmed its steel and aluminum tariffs in February 2025, it ended all country-exemptions. The whole world faced the same tariff rate.
While not formally withdrawing its FTA, or WTO (more on this in a moment) steel, aluminum and automotive tariff commitments, the United States had effectively nullified them.
Mexico and Canada have refused to do the same.
Mexico boasts 14 free-trade agreements covering 52 countries, plus nine limited-scope agreements through the Latin American Integration Association. Canada has 15 free trade agreements covering 52 countries, and has six active comprehensive free trade negotiations aside from CP-TPP accession. This includes one with India.
Both countries have steadfastly refused to revisit the schedule of concessions under any of their FTAs.
So again, what Fortress?
The ‘Fortress North America’ Pandering We’ve Gotten Since 2018
While neither Canada nor Mexico is willing to withdraw its steel and aluminum tariffs in its FTAs, they have been willing to offer various figleaf concessions towards their Fortress pitch.
This song and dance that began in the run-up to USMCA continued throughout the Biden Administration, and, per Secretary Ebrard’s Aug. 30 remarks quoted above, remain Mexico’s game plan.
Disastrous Failure No. 1: the 2019 Steel and Aluminum Agreements That Facilitated Import Surges, Circumvention, and US Plant Closures
In May 2019, USTR announced it had reached agreements on steel and aluminum with both Canada and Mexico. Under the agreements, American manufacturers were assured that there would be “aggressive monitoring” and that if covered metal imports “surged meaningfully beyond historic volumes of trade”, then the Section 232 tariffs would be restored.
The agreements also promised to “[p]revent the transshipment of aluminum and steel made outside of [North America].”
The agreements weren’t worth the paper they were written on.
Mexico violated promise to avoid surging exports comprehensively, as CPA has documented: by 2023, steel mill shipments ran 36% over baseline, Chapter 72 iron and steel imports were 134% over, and electrical steel conduit – the humble pipe carrying wiring through every American building – 472% over baseline, heading for nearly 700%, until Mexico controlled 87% of U.S. conduit imports.
It wasn’t just the surge, either, but also the transshipment, despite the “aggressive monitoring” promised in the agreement. The transshipment of Chinese steel and aluminum via Mexico into the United States continued, and without any tariffs or quota to control import penetration, the predictable result happened. American mills were shuttered as Mexico took the market.
USTR took note of the deal’s failure. On February 16, 2024, USTR issued a readout of a meeting with Mexico stating that “Ambassador Tai stressed the urgent need for Mexico to take immediate and meaningful steps to address the ongoing surge of Mexican steel and aluminum exports to the United States and the lack of transparency regarding Mexico’s steel and aluminum imports from third countries.”
Mexico once again managed to dodge tariffs, however, despite scoffing off every aspect of the 2019 Agreement.
Instead of tariffs, Mexico successfully sold the Biden Administration on the Fortress North America plan. Under this new plan, covered steel and aluminum imports could only be freely imported if the importer certified that the steel was melted and poured, or the aluminum was smelt or cast, in North America.
But North American rules of origin keeps American mills and fabricators in a price competition they cannot win. The rules are also fruitless because they remain critically undermined by scoping issues (discussed below), varying metal costs across the three countries, and fail to learn from the disastrous North American Automotive Rule of Origin.
Throughout this time period, U.S. production of steel and aluminum shrank.
For steel:
For aluminum:
Disastrous Failure No. 2: The USMCA Automotive Rule of Origin
On March 26, 2025, President Trump imposed the global 25% tariff that should have happened in 2019. The Proclamation noted diplomatically that the Automotive Rule of Origin had been a bust: “the revisions to the United States-Korea Free Trade Agreement and the United States-Mexico-Canada Agreement (USMCA), have not yielded sufficient positive outcomes. The threat to national security posed by imports of automobiles and certain automobile parts remains and has increased.”
The numbers don’t lie: In 2018, Mexico sent 2,663,805 cars to America while we sent only 140,560 to Mexico. By 2025, Mexico had increased its exports to 2,700,297 while our exports to Mexico had fallen to 126,221.
Put simply: Mexico sends us 20 cars for every 1 car we send to Mexico. And that happens because they have lower wages, and USMCA continued the NAFTA promise to automakers of unlimited U.S. market penetration for Mexican vehicles.
On July 1, 2026, USTR issued a report on the failure of the USMCA Automotive Rule. In the report’s Foreword, authored by Greer, acknowledged that:
“the United States has run a persistent, substantial, and growing trade deficit in autos and parts with Mexico for over 30 years. In 2024, that deficit peaked at $138.2 billion. In recent years, analysis indicates that the share of U.S. content within Mexico’s automotive exports to the United States has been declining, while the share of Chinese or other Asian content in those goods has been increasing.”
Regrettably, despite this acknowledgment, the Foreword indicated a doubling down on tougher ‘rules’: “Through strategic improvements, USTR seeks to raise regional and U.S. content, reduce reliance on inputs from third-countries, and build a more resilient and secure supply chain”.
Conjured-up academic rules which contemplate one nation’s entire industry being offshored so long as the competition was ‘fair’ should be a non-starter. Instead, USTR should be negotiating firm, outcome-oriented measures, like quota caps on the amount of vehicles imported. For example, in USTR’s May 2025 UK trade deal, rather than rules, the UK was put on a quota: 100,000 cars at a 10% tariff, 25% thereafter.
Disastrous Failure No. 3: Another USTR “Monitoring Agreement” Hollows Out US Capacity to Build Transformers, Highlighting Scoping Issues Around Steel and Aluminum
Grain-oriented electrical steel — GOES — is the specialty steel at the heart of every grid transformer. Exactly one company melts it in the United States, Cleveland-Cliffs at Butler, Pennsylvania; neither Canada nor Mexico melts a single ton.
But fabricators in Mexico can buy it duty-free from suppliers in Asia, and from 2018 until 2025 were able to export GOES products duty-free into the United States with only the mildest treatment.
This is due to scoping issues inherent in tariff actions involving primary metals. When we talk about the “steel” and “aluminum” tariffs, that does not automatically include every product containing those metals. So when Canadian and Mexican fabricators can source cheaper metal than American fabricators, and export their products to the United States duty-free, American fabricators get wiped out.
Transformers, which are critical equipment for our civilization’s existence, became a perfect case study.
While the 2018 steel tariff did cover GOES, it did not extend to the laminations and cores one step downstream of the metal.
The arbitrage assembled itself immediately: Japanese GOES was slit, stacked or wound — a transformation so minor that GOES remains 70–90% of the article’s value — and the components entered the United States tariff-free. GOES imports fell 56% in 2019 while import penetration of laminations reached 88%; since 2020, roughly 95–98% of both countries’ exports of transformer parts have gone to a single market, the United States. Neither country built a components industry. They built a tariff bypass.
In October 2020, as part of a Section 232 investigation into transformers and transformer components, Commerce formally found that “transformer laminations, stacked cores and wound cores are being imported into the United States in such quantities and under such circumstances as to threaten to impair the national security.” The remedy? No tariffs, but instead another monitoring agreement: Mexico would operate “a strict monitoring regime,” and in exchange Mexican transformer laminations and cores would “not be subject to any action” under Section 232. The monitoring regime then monitored the following: imports rising every single year, up 321% by 2024 over 2018, from $126 million to $527 million. It fell to the sole American producer to petition Commerce, in May 2025, to pull laminations and cores inside the tariff as steel derivative products — granted August 18, 2025, seven years into the steel 232 and five years after the government found the threat. And the coda: Census data now runs ten months past the inclusion, and imports have not fallen a dollar — roughly $40 million a month before, $40 million a month since, all of it duty-paid. The offshored capacity simply pays the toll, because Asian imports don’t flinch at a 25% tariff.
What played out with GOES applies to countless other downstream metal products still outside the scope of the steel tariffs: from boats to bicycles, from gun safes to golf carts.
Costly Fig Leaf Number 3: Canada and Mexico’s WTO-compliant steel tariff maneuvers
Mexico’s January 2026 reform raised MFN duties to 25–35% on household appliances, aimed squarely at China — the very instrument Washington is being asked to accept as its northern and southern wall. Six months of data later: Chinese shipments were down just 2.5%, and China’s share of Mexican appliance imports actually rose, to 68.1%, because the across-the-board hike crushed the second-tier suppliers (Thailand down 71%) while China absorbed the tariff through its cost advantage and CPTPP-preferenced Vietnam sailed through untouched. Meanwhile the same CPTPP door has been swinging in Canada: the Canadian appliance market grew 49% from 2016 to 2024 while the U.S. share collapsed from 42.9% to 29.8%, with Vietnam going from zero to $113 million — concentrated in large washers, Louisville’s flagship product. Haier builds washers at its plant in Biên Hòa, Vietnam; LG and Samsung supply Canada from Vietnam and Thailand rather than their U.S. plants. And the White House has celebrated Haier’s GE Appliances — a Chinese-owned manufacturer — as a showcase of American manufacturing. If a Chinese-owned OEM is the celebrated face of American appliance making, Ottawa and Mexico City hold a green light to court Chinese OEM plants of their own and call the output North American. A fortress where Haier is simultaneously the American champion in Louisville and the CPTPP importer from Biên Hòa is not a fortress. It is a franchise.
The tally: a restraint agreement, violated. A monitoring agreement that monitored a 321% increase. An MFN tariff wall that raised China’s market share. The 2026 “harmonization” pledge would be promise number four — the largest and least verifiable yet.
When the Sheinbaum government raised MFN tariffs to 25–35% on numerous steel and aluminum products, the decrees expressly preserved preferential treatment for treaty partners. A 35% “steel tariff” sounds formidable until one asks who pays it: in late August, Mexican domestic hot-rolled coil was assessed around $982–$989 per metric ton while Japanese HRC — expressly FTA-exempt — was offered into Mexico at $685–$700 CFR. One of the world’s largest steel producers ships qualifying material at zero through the wall’s dozens of doors. That is not a North American external tariff.
Harmonize to what, exactly? The U.S. steel action’s own scope has been redrawn six times in eighteen months: the February 2025 derivative annex; household appliances added that June; 407 more tariff lines — charcoal barbecues to transformer components — that August; an April 2026 proclamation that rebuilt the entire architecture into three duty tiers, removed products and terminated the inclusions process; and an August 2026 proposal to start adding again. A refrigerator is a national-security derivative in Washington and ordinary commerce in Ottawa; a bicycle that Washington expressly refused to cover carries a 35% tariff in Mexico City; gun safes are protected by no one. Three governments cannot harmonize schedules when they cannot agree on — and Washington cannot itself hold still — the list of products the fortress protects.
Nor does a promise to “consult” solve anything. Greer himself denied that the proposed text gave Washington a veto over Canadian trade deals. Without such authority, harmonization is inherently temporary: Canada or Mexico could negotiate a new preferential agreement tomorrow and open another hole in the perimeter. A true common external tariff is not a one-time concession; it is a continuing surrender of national trade-policy autonomy — which is precisely why no one will agree to it.
Nobody Can Enforce It Anyway
Even identical headline rates would not build a perimeter, because North America runs on special customs regimes beneath them. Mexico’s IMMEX program admits inputs duty-free for export production — a central feature of its manufacturing model, and one Mexico itself tightened in 2024 after finding what its own government called “technical smuggling.” Mexico’s antidumping order on Chinese and Vietnamese hot-rolled steel had to enumerate definitive imports, temporary imports, three bonded regimes and Rule Eight entries — an admission that a tariff imposed only at the ordinary border is not enough. A genuine perimeter would need common, enforceable, continuously audited rules for where metal is melted, smelted and cast, how origin attaches through processing, and how IMMEX, FTZ and drawback interact — across three countries.
Washington has already built a preview of the problem into its own aluminum action: since April 2026, imported articles made entirely from U.S.-smelted-and-cast metal enter at 10% instead of 50% — a two-tier rate keyed to an origin claim no one at the border can verify, because CBP has no operational system for confirming where the metal inside an extrusion was smelted. The history counsels zero optimism. The Department of Justice’s $549.5 million settlement with Perfectus Aluminum this May — the largest customs-fraud recovery in U.S. history, involving 2.2 million Chinese extrusions disguised as “pallets” — took fifteen years to resolve. That scheme required elaborate physical disguise; the U.S.-smelted carve-out requires only paperwork. Every incentive points toward a repetition of the Perfectus fraud, at larger scale, with less effort.
And that is the American side — a professional customs service working a single fraud. The Mexican side is worse. Earlier this decade Mexico purged its civilian customs corps amid corruption scandals and handed all fifty-odd customs offices to the military — army at the land crossings, navy at the ports. In 2025, Mexican prosecutors charged a network of naval officers, including admirals related to the former navy secretary, with running a massive fuel-smuggling operation through the very ports the navy was assigned to guard. The institution being asked to co-administer melt-and-pour certificates, CPTPP origin declarations and IMMEX inventories is a military customs service whose flagship achievement to date is a smuggling indictment of its own admirals.
The Price of Getting It Wrong
This would be academic if North American metal prices were converging. They are not — and the market has already told policymakers what a concession costs. When reports emerged that the Canadian steel tariff might be halved, U.S. steel futures sold off $50 a ton in a single day. When the same reports touched aluminum, the September Midwest Premium dropped 8.2% in one session. A Canadian Section 232 concession is not a diplomatic courtesy; it materially changes the U.S. metal price.
The worst outcome is asymmetric integration: high U.S. prices created by comprehensive Section 232 protection, combined with lower-cost Canadian and Mexican production that retains access to third-country inputs under different tariff rules, followed by preferential access for the resulting articles into the United States. That is not Fortress North America. It is a tariff arbitrage system.
Do Not Pay 25 Points for a Slogan
Greer is right about the underlying problem. The United States cannot maintain a serious industrial policy if Canadian and Mexican markets become staging grounds for third-country steel and aluminum, and Washington is entirely justified in demanding that its partners’ policies not undermine Section 232. But recognizing the problem does not make the proposed solution possible. Canada is not going to abandon CPTPP; its prime minister just called the suggestion unacceptable. Mexico is not going to dismantle a 52-country treaty network; its president will not even concede the transshipment problem exists.
And Americans should be careful what they wish for, because sovereignty cuts both ways. A genuinely common trade policy requires common decision-making: pressure on Washington to consult Ottawa and Mexico City before changing American tariffs. What happens when a future Canadian government wants a lower steel tariff, or Mexico demands an exclusion as the price of maintaining the wall?
Their governments may raise MFN rates, adopt safeguards, improve traceability, crack down on Chinese circumvention. All useful. None is tariff harmonization. USTR should therefore adopt a simple rule: no Section 232 concession for promises of coordination. A reduction from 50% to 25% should occur only for reciprocal protection that is real, legally enforceable and economically equivalent — including treatment of third-country metal entering through CPTPP preferences, special customs regimes, transformation rules and downstream derivatives. Anything less leaves American producers paying the price for a fortress whose northern and southern gates stand open.
Three promises have already been tested. Three have failed. There is no reason to give away half the wall in exchange for a fourth — least of all for a slogan.
MADE IN AMERICA.
CPA is the leading national, bipartisan organization exclusively representing domestic producers and workers across many industries and sectors of the U.S. economy.
TRENDING
The “Fortress North America” Mirage: Why Tariff Harmonization Will Never Work
AI Data Centers and Pharma Imports Send July Deficit to $119.59 Billion, Highest All Year
Why Beef Prices Are Still So High: More Imports Won’t Fix America’s Cattle Crisis
After Canada Impasse, Mexico Quietly Left Washington Empty Handed, But Reiterates Section 232 Tariff Goals.
Five Reasons the Collapse of the Canada Trade Deal Is a Win for America
The latest CPA news and updates, delivered every Friday.
WATCH: WORTH FIGHTING FOR
Get the latest in CPA news, industry analysis, opinion, and updates from Team CPA.
CHECK OUT THE NEWSROOM ➔