REVIEW: The MEI is right on Section 338, wrong to tell us to shelve our counter-tariffs

Crates at the border

September 23, a paper

The Montreal Economic Institute published a paper on September 23, 2026, that gives Canada two opposite pieces of advice. Trump will probably lose in court. And Ottawa should perhaps rethink its retaliation.

That is the paradox. The harshest case against Trump’s tariffs asks Canada to lower its guard.

The author is Ilya Somin, a law professor at George Mason University and an associate researcher at the MEI. He co-signs with Renaud Brossard, the institute’s vice-president of communications.

A serious jurist. A serious paper. A risky piece of advice.

September 29, one minute past midnight

Six days after the paper, on Tuesday at 12:01 a.m., Washington starts banning certain Canadian spirits, dairy products, molasses and motorcycles, CBC reports.

In Ontario warehouses, crates wait under fluorescent lights. The packing tape squeaks one last time. Nobody knows if the truck will leave.

The paper is counting on the courts. It admits, though, that when it was written, no lawsuit had been filed against these duties.

No lawsuit filed, and we are asked to wait.

The law is on our side. The calendar is not.

Nine minutes against Trump

Harmful, unjust, illegal

The paper takes nine minutes to read. It pulls no punches.

Trump’s tariff policy, Somin writes, is harmful, unjust and illegal. It damages the American, Canadian and global economies. It needlessly poisons relations with crucial allies.

He is not writing from a distance. On behalf of five small businesses, he helped argue the case that struck down the duties imposed under the International Emergency Economic Powers Act, the IEEPA.

He knows what he is talking about. That is what makes his final advice weigh so much.

He even grants the main point. The urge to retaliate is entirely understandable, he writes. If Trump were the only one hurt by it, there would be no argument against it.

Law, evidence, usefulness, blind spot

I judge the paper on four points. The accuracy of the law. The strength of the evidence. Its usefulness to Canada, now. And what it leaves out.

On the law, it is strong. On the evidence, solid. On usefulness, fragile. On what it leaves out, costly.

I will say it early. This column sides with the paper against Trump, and against its advice to rethink Ottawa’s counter-tariffs. Because as of September 23, no lawsuit had been filed against Section 338. And the last battle, over the IEEPA, took ten months.

You put a counterpunch away once the judge has spoken. Not before.

A good analysis can give bad advice.

Ten months for the IEEPA

The Supreme Court, already

The precedent is real. In February 2025, Trump invoked the IEEPA to impose 25% duties on Canada. Officially, to fight fentanyl trafficking, the paper recalls.

The Court ruled that the law does not authorize tariffs at all. The Court recalled that tariffs are a power of Congress. Not of the executive.

Three justices and two lower courts also invoked the major questions doctrine. It requires Congress to speak clearly on decisions of vast economic significance.

The Court of International Trade then struck down duties based on Section 122 of the 1974 trade law. That case is on appeal.

That leaves Section 301. Trump is using it to impose 10% to 12.5% on sixty partners, Canada among them, on forced-labor grounds. Somin sees a pretext. Three lawsuits have already been filed, one of them by 25 states. A ruling is expected in the coming weeks, he says.

US$166 billion collected

And yet the victory cost time. The IEEPA case took ten months from start to finish, Somin writes.

Meanwhile, Trump collected about US$166 billion in payments the paper considers illegal. Not all of it has been refunded yet.

Ten months. Billions collected. Partial refunds.

The paper itself supplies the strongest evidence against its own advice.

A court can rule for Canada. It cannot give back the months.

Winning in court means winning late.

Section 338, dug up

July 20, three proclamations

Section 338 comes from the 1930 tariff act known as Smoot-Hawley. That law deepened the Great Depression, the paper recalls.

On July 20, 2026, Trump became the first president to impose duties under that section. So write Georgetown legal scholars Peter Harrell and Jennifer Hillman.

Three proclamations. Spirits. Dairy. Vehicles.

The 50% duties took effect on August 22, after a three-day delay, according to the law firm WilmerHale. On September 8, Trump signed exclusion proclamations. They ban certain Canadian imports as of September 29.

US$5.6 billion, US$19.3 billion

On the law, the paper is right. Others make the case better than it does.

Section 338 authorizes duties to “offset” discrimination. Harrell and Hillman run the numbers. The White House puts lost auto sales at US$5.6 billion.

The matching proclamation imposes 50% on about US$19.3 billion in imports. That comes to some US$10 billion in duties a year. Nearly twice the alleged harm.

The duties hit hockey sticks and cement. The vehicle measure does not even cover a single vehicle tariff line, they write.

Read that way, the law would allow a tax on almost anything the United States imports. There are more than 380 preferential agreements among its partners, they point out. That many pretexts.

The dairy quotas under fire were negotiated by Washington itself in CUSMA. The targeted spirits come down to purchases by provincial liquor boards. Not customs.

A 1930 law wakes up, badly applied.

No lawsuit as of September 23

A door still shut

Here is the hole in the paper. The author wrote it himself.

At the time of writing, no challenge had been filed against the Section 338 duties, Somin acknowledges. He expects some soon. From states, importers or others.

Soon is not a date. And if soon means winter…

There is no lawsuit. There is no hearing scheduled. There is no judge on the case.

The U.S. International Trade Commission, for its part, is considering reviving a role it abandoned decades ago. It is taking comments on Section 338 until November 9, 2026, WilmerHale reports.

The Commission acknowledges it has no established practice for identifying foreign discrimination. None. In 2026.

A ban cannot be bargained down

While the legal case takes shape, the bans are arriving. They are designed to discourage any further retaliation, according to a senior White House official and experts cited by CBC.

A tariff is a number. It can be negotiated down. A ban is really hard to lift, lawyer Barry Appleton warned on CBC.

The paper talks about tariffs. It does not talk about bans.

They had been proclaimed fifteen days before it came out.

You argue a tariff. You endure a ban.

57% of Americans against

Ipsos, YouGov, the lake

The paper’s other sound half is political. It is valuable.

According to Ipsos, 57% of Americans oppose the tariffs on Canada. Only 20% support them. The poll dates from September 1, the paper notes.

According to Ipsos, 63% of Americans also reject renaming Lake Ontario.

An Economist/YouGov poll, taken August 28 to 31, counts 58% opposed to new tariffs on Canada.

In March, 55% of Americans viewed Canada favorably. Against 10%.

More popular than the president

Canada is far more popular in the United States than Donald Trump, Somin writes. Hence his advice: work with the Americans who oppose the tariffs.

That is right. It is even the best page in the paper.

The lever exists. It is American. Who in Washington is listening to that 57%?

The paper also proposes a clause for a future CUSMA. Any court ruling that strikes down an illegal tariff would take effect immediately. No stay until the appeals run out.

Ottawa should put that idea on the table tomorrow morning.

Our best ally votes in the United States.

The September 8 counterpunch

C$27.6 billion, dollar for dollar

Now let us look at what the paper criticizes. As Ottawa wrote it.

On August 25, Finance Minister François-Philippe Champagne announced counter-tariffs of 15%, 25% and 50% as of September 8. They target C$27.6 billion in American imports. Dollar for dollar. Rate for rate.

The sectors: steel, dairy, appliances, farm equipment, pulp and paper, electronics.

The same day, Ottawa added C$7.5 billion in support for affected workers and businesses.

And yet the counterpunch has a release valve. The remission framework stays open to requests for exceptional relief, the department says. A struggling importer can ask for mercy.

The goal Ottawa wrote down

Here is the point the paper does not discuss.

The Department of Finance writes that the primary goal of the counter-tariffs is to protect Canadian workers and producers. By putting them in a better position against American products, in the Canadian market.

Not to make Washington fold. To hold the market here.

The paper judges a tool against a goal the government did not give it first.

That is a flaw of method. It weighs heavily in a paper built on method.

A shield is judged by what it protects.

The consumer pays

Furniture, clothing, 50%

Now I have to write what proves the MEI right. It costs my side.

The 50% counter-tariffs hit furniture, clothing and apparel, the Department of Finance specifies. These are not factory inputs. They are household purchases.

The 25% rate hits appliances and some cheeses.

The stove you replace. The couch you push to next year. The price tag you flip over in the store, then put back.

September 8, two signatures

Economist Vincent Geloso, in an MEI viewpoint published in early September, adds the sharpest argument. Once in place, counter-tariffs attract the companies that profit from them. Those companies lobby to keep them. They become hard to repeal.

The timing is awkward too. On September 8, the day our counter-tariffs took effect, Trump signed the ban proclamations, according to WilmerHale.

I do not claim one caused the other. I note the date.

Households pay. The lobby waits.

It is true that the consumer pays. It is true that the counterpunch can dig in. It is true that both signatures land on the same day.

And that does not change the rest of the equation.

Our counter-tariffs have a price, and we pay it.

1930, as Geloso reads it

Mackenzie King, Bennett

The MEI’s historical evidence comes from 1930. The same year as Trump’s law.

Mackenzie King had let a likely retaliation hang in the air. Richard B. Bennett’s Conservatives, elected in July 1930, imposed steep tariff hikes that September, Geloso recalls. They targeted about 30% of American exports to Canada.

Those counter-tariffs had no visible effect on American tariffs, he writes. Bennett eventually backed down.

The real change came from Washington. The Democratic victory of 1932. Then the Reciprocal Trade Agreements Act of 1934.

Geloso admits a limit. It is hard to separate the effect of tariffs from the general collapse of trade. Canada’s share of trade fell by about a third during the Depression, according to the work he cites.

The footnote

The history is accurate. It proves less than it claims.

In 1930, Canada was retaliating against a law passed by Congress. In 1930, the president who signed it was not being challenged in court on that point. In 2026, Ottawa is retaliating against proclamations a court could overturn, against a policy that is unpopular at home.

And Geloso, to his credit, slips in a footnote. Researchers found that China’s 2018 retaliation weighed on the U.S. midterm elections. He answers that China is a large economy. Canada is not.

That is an argument. It is not proof that Canadian pressure counts for nothing.

China carries weight. So does Canada, in a different way.

1930 teaches caution. Not surrender.

The argument already on file

Somin’s worry

That leaves the legal argument against our counter-tariffs. It is the most technical. It is also the weakest.

Section 338 requires discrimination against American products. Somin worries that our new counter-tariffs, which by definition target American products, weaken that argument before the judge.

He admits the plaintiffs could still win. They could argue that there were hardly any discriminatory Canadian tariffs when the Section 338 duties were imposed, he writes.

The 2025 auto tariffs

But Harrell and Hillman write something else. One of the three justifications Trump invoked on July 20 targets Canadian tariffs on American cars. Tariffs adopted in retaliation for Trump’s first tariffs, last year.

And the Department of Finance confirms that Canada’s existing counter-tariffs, including those on autos, remain in place.

If so, retaliation was already in the file on July 20. Shelving the September 8 round would not erase that exhibit.

The worry comes too late. What does one more counterpunch change in a file that already holds the first?

Pulling the counterpunch does not erase the exhibit.

Six days, ten months, two years

Factory time

Here is what the paper puts in no table. The time of the people who wait.

Six days between the paper and the bans. Ten months for the last case won. About two years until the end of Trump’s term, by Somin’s own count.

His best reforms, a new CUSMA and an end to stays, will almost certainly not arrive while Trump is in office, he writes.

Two years, for a brewery, means two summers. For a protein plant, two budgets. For an employee, two lease renewals.

What a brewery says

Six days. Ten months. Two years.

Once you add 50% duties, selling into that market becomes extremely difficult, Matt Johnston of Hamilton’s Collective Arts brewery told CBC.

BioSteel’s boss in Windsor, Dan Crosby, talks about resources, money and time sunk into the market. Then taken back at the eleventh hour.

The paper asks these people to trust a lawsuit that does not exist yet.

I cannot ask them that.

Not when the other camp is putting nothing away.

In court, months. At the plant, weeks.

The half that holds

What the MEI brings

The paper deserves credit for what it brings. It is a lot.

It is right about Section 338: never used, badly applied. It is right about the IEEPA: ten months, then Trump’s defeat. It is right about American opinion: 57% against. It is right about stays: US$166 billion collected before the ruling. It is right about the price: our furniture at 50%.

Right five times.

What it costs

And yet it lands badly. Its advice rests on a bet.

The bet that a court will act fast. That a lawsuit will be filed soon. That Washington will put its weapons away if we put ours away.

None of these three bets has a date.

The sound half is legal. The half that costs is strategic.

And the second one is what we pay for at the checkout.

Right on the law. Costly on time.

Right on the law, wrong on waiting

November 9

There is one date ahead of us. November 9, the end of comments at the International Trade Commission.

It lifts no ban. It can help document what Washington calls discrimination.

Ottawa should take from the paper what helps. The American allies. The clause against stays. Support for the lawsuits. And keep what protects, until the judge speaks.

The cost of waiting

Who should carry the cost of waiting, the consumer who pays our counter-tariffs or the worker who pays theirs?

Keeping the counterpunch is not loving tariffs. It is refusing to show up unarmed in court.

The 1930 law may fall. The crates are already waiting.

Right on the 1930 law. Wrong to await a lawsuit that does not exist.

Sources:

Primary Sources:

• MEI — Ilya Somin and Renaud Brossard on the legal battle over Trump’s tariffs, September 23, 2026
• MEI — Vincent Geloso on why counter-tariffs fall short, September 3, 2026
• Department of Finance Canada — September 8 counter-tariffs and C$7.5 billion in support, August 25, 2026

Secondary Sources:

• Reason (Volokh Conspiracy) — Ilya Somin presents his MEI paper, September 23, 2026
• Mondaq (WilmerHale) — the International Trade Commission and Section 338, September 23, 2026
• Reason (Volokh Conspiracy) — Harrell and Hillman on challenges to Section 338, August 3, 2026