Not all parts of free trade are equal. Some countries have lower wages. Some have more abundant resources of one thing or another. With Canada, the assumption is usually that things are basically the same — same wages, same costs, same rules. When it comes to softwood lumber, Canada can sell at a lower price. In some cases, as with dairy, countries set quotas to protect their local industry. For softwood, that was never the US ask. What the US wanted was head-to-head competition — supposedly fair competition. In the case of wood, the assumption from the US point of view has always been that Canada was competing unfairly by subsidizing its industry.
How? The raw trees are owned by the government in Canada. In the US, the land is mostly private. Even if the land in Canada were private, the two countries might not have equal costs. Either way, the way the Canadian government charges for logging is called a stumpage fee — not an open-market price, but a rate set administratively. The claim has been that those fees are too low, and that this makes it impossible for US firms to compete.
The fight was never winnable on its own terms
Ninety percent of Canadian forest land supplying the lumber industry is Crown land. Provinces set stumpage fees by schedule. Most US forest land supplying the industry is private — sometimes owned outright by the mills that cut it, sometimes bought from independent landowners through a negotiated sale or a competitive auction. Either way, when a price does get set, it comes from a buyer and seller striking a deal or bidding against other buyers, not from a government schedule. The US Commerce Department revisits the rate every year in what’s called an administrative review, or AR — a fresh recalculation, with a new number attached, on a clock that never stops. Four of the most recent reviews alone have tried to benchmark the two systems against each other, as if an administered fee and a market price were the same instrument measuring the same thing.
They aren’t, and USMCA and NAFTA Chapter 19 panels have said so, repeatedly — ruling the US benchmark methodology unsound. The US kept using it anyway. AR5 became AR6 became AR7 became AR8, each one re-litigating the same comparison, because the two countries’ forestry systems were never going to converge into something a single price benchmark could fairly measure.
British Columbia already ran that experiment, partially. Since the 2006 Softwood Lumber Agreement, BC has sold roughly 20 percent of its Crown timber at open, competitive auction through BC Timber Sales, then used those auction results to extrapolate a stumpage rate for the rest of the province’s harvest. It’s not a full market conversion — most of BC’s timber still moves at the extrapolated rate, not an actual auction price — but it’s the closest thing to the private-market pricing the US says it wants, built specifically to satisfy Washington. The duties came anyway. Commerce has leaned on other provinces — Nova Scotia’s private stumpage rates, at points — as the benchmark instead, over Canada’s objection that the comparison doesn’t hold. Even a partial, one-province market experiment, built to the US’s own specification, didn’t end the dispute.
The way out Canada already uses — for dairy
There’s a way to sidestep the stumpage comparison entirely: stop trying to align two structurally different pricing systems, and just fix a volume. Canada already does this for its own protected industry. The dairy tariff-rate quota doesn’t ask whether Canadian milk pricing is comparable to US milk pricing. It sets a number and stops arguing.
Softwood never went that route. The US lumber industry chose litigation instead, filing the first countervailing duty petition in 1982, claiming Canadian stumpage fees were an unfair subsidy. That’s a legal remedy under US trade law, triggered by a private petitioner — not a negotiated instrument between states, the way dairy’s quota was. Softwood’s fight was picked by an industry, in court, and has stayed there ever since.
Earlier this year, that looked like softwood was headed to a quota system also. BC Premier David Eby raised quotas as a path to resolving the dispute. Prime Minister Mark Carney said a future deal could include “some element of managed trade” on softwood. For a few months, quotas were being discussed as a real option — the first genuinely new instrument on the table in this dispute in years.
What was actually offered
That’s not what showed up in the room at the end. When the US and Canada sat down for the final push in August 2026, the offer on softwood wasn’t a quota. It was narrower: drop the 10 percent Section 232 tariff — the national-security tariff added in October 2025 — by September 1. The underlying “anti-dumping”/”countervailing duty” (AD/CVD) tariffs, the ones running through AR5 to AR8 since the last agreement lapsed in 2015, would stay exactly where they are.
So the quota talk that ran through the first half of 2026 didn’t survive contact with the actual negotiation. What survived was an offer to peel back the newest layer of the dispute while leaving the oldest layer — the one built on the stumpage argument — untouched.
Why it fell apart anyway
Talks collapsed on August 21–22, and softwood wasn’t the reason. Carney said the deal broke down over US terms he said would “destroy” Canada’s auto, steel, and aluminum sectors — sectors with nothing to do with stumpage fees or Crown land. He also named bilingual labeling as a red line, USTR Jamieson Greer called that a “funny, fake story,” and Carney fired back that it was real. I’ve covered that dispute on its own terms in “Washington’s Complaint About Canadian Bilingual Laws” — the short version is that Carney folded a fifty-six-year-old, uncontested federal labeling law together with a one-year-old, already-softened Quebec trademark rule into a single undifferentiated grievance, because the merged version plays better as a soundbite than either piece does on its own. That’s not a one-off. The same flattening shows up in how he’s handled dairy, covered in “Canadian Dairy Quotas: The Taxi Medallions of Their Dairy Industry.” Trump makes these fights harder than they need to be with the “51st state” routine — see “Does Taunting Canada Have Any Rational Purpose?” — but Carney isn’t a plain-spoken counterweight to that. He’s playing the identical game to his own audience: real grievances and manufactured ones collapsed into one outrage, engineered for the domestic news cycle rather than for accuracy. Softwood was part of the package that fell — Trump had also reportedly offered to cut the auto tariff — but it wasn’t what killed the talks, on either side’s account.
Which means softwood lumber didn’t get resolved and didn’t get rejected. It got shelved, dragged down by sectors that were never part of the stumpage argument in the first place.
Where that leaves it
Forty years in, the pattern holds. A new instrument gets floated — this time a quota, dairy-style, treaty-durable in a way the old side deals never were — and it evaporates before it’s tested, not because either side rejected it on the merits, but because it was never negotiated on its own this time. Softwood has been settled on its own terms before — a 1986 MOU, the 1996 Softwood Lumber Agreement, the 2006 SLA — and none of the three lasted. Canada unilaterally withdrew from the 1986 deal in 1991; the 1996 and 2006 agreements ran their fixed terms and lapsed. Either way, the underlying AD/CVD machinery was always still there, waiting. That’s arguably worse than never settling at all: even a negotiated fix doesn’t retire the statutory process that produced the dispute in the first place. It just pauses it.
The underlying AD/CVD duties are still there, unaffected by any of this. AR8 is still running its own review cycle, on its own schedule, indifferent to whether the two governments are speaking to each other. Whatever the next opening looks like, it’ll have to survive being attached to something else again — because in this dispute, it never gets to stand on its own.