Negotiating CUSMA on a Dog Day Afternoon
Dan Ciuriak is director and principal of Ciuriak Consulting and senior fellow at the Centre for International Governance Innovation where he writes on the digital economy.
ICYMI, things aren’t going so good for The Donald. In the “you broke it, you bought it” department, he’s trying to offload a property (the Iran War), bought on spec and not living up to promise, onto a new set of investors (specifically naming China, France, Japan, and South Korea). Turns out they aren’t all that interested in the property, and might not be returning his calls due to insults and threats (which, by the way, is not a recommended “Marketing 101” approach to these things). Gasoline prices are up; energy rationing is being imposed around the world; critical disruptions in material supplies are emerging for semiconductors (helium), electric vehicles (aluminum) and other sectors; markets are down, airports are in chaos, the Boss is de-ICEing Minneapolis, and The Economist cover just dropped with a very unflattering photo and an even less flattering lede: “Never interrupt your enemy when he’s making a mistake”.
We are watching a loose 2026 adaptation of Dog Day Afternoon (if you don’t get the reference, Google it, it repays the read).
So, is this a good time to be negotiating the renewal of the USMCA (CUSMA to Canadians)?
It might be the best of times, as one comment pointed out: “Iran Oil Shock Is Canada’s Golden Opportunity in USMCA Talks”. Others point to the looming mid-terms as working in Canada’s favour because tariffs don’t improve Americans’ affordability. So Canada should strike while the iron is hot to use a metaphor from the early days of industrialization, when everyone was familiar with how blacksmiths worked (and what blacksmiths actually did).
But it also might be the worst of times. We asked Chat: “On balance, this is a poor moment for Canada to negotiate USMCA renewal. The U.S. is operating in a highly volatile, litigation-constrained, and politically reactive environment, which weakens the credibility and durability of any commitments it might make. Absent clear counterpart stability, Canada’s leverage is limited and the risk of negotiating into a moving target is unusually high.” (emphasis in the original)
As for me, I really don’t know (beyond my pay grade). But I do have some ideas I’d like to share about how Canada should approach these negotiations based on the economics, which is something (checks credentials) that I hope I do know a little about.
Dog Day Afternoon 2026 – The Economics
First, Canada has already absorbed the main shock of U.S. protectionism – the uncertainty bomb dropped long ago. It has undermined foreign investment in Canada to serve the U.S. market for an indefinite period (ask yourself, what Board of Directors in its right mind would build in Canada to sell to the United States?).
For Canada, investments in accessing the US market for manufactures is a sunk cost. What investment is coming to Canada is coming for the weather and the rare earths and is not hostage to tariff threats. So Canada should build on the latter rather than trying vainly to recapture the former.
Second, there is no tariff Sword of Damocles hanging over Canada’s economy. The Bank of Canada projects a 2.4 percent hit to GDP under an across-the-board 25 percent tariff. Oxford Economics puts its worst case at 1.8 percent. To put this in perspective, this is one to two years worth of forgone growth. In other words, the worst-case scenario for Canada has it remaining as prosperous as we are now for a couple of years max, before we start growing again at our potential rate. Put another way, we get to where we would be income-wise sometime in 2030 instead of 2029. And that is if Canada does not take countermeasures – and every firm and level of government in Canada is taking countermeasures (the last phrase should be read in all caps).
The reason why the discount on Canada’s economic prospects from a worst-case outcome is this modest is that Canada only exports about 15% of its value-added production to the United States, and of that about half is comprised of products that the United States really needs and will buy from Canada, tariffs or (when reason prevails) no tariffs.
And as for the stuff that Canada ships to the United States that is vulnerable to tariffs, the hurt in Canada of seeking new markets is at least partly matched by the hurt in the United States in seeking new suppliers – and new customers, because as Canada turns inward to serve domestic markets, that market share is captured from American suppliers. So there’s a lot of hurt to go around. And hurt is not good at the ballot box. Dog Day Afternoon stuff.
For these basic reasons, Canada should not run scared in approaching the negotiations. But wait, there’s more.
And now for something completely different
Canada should not be bracing itself to concede freebies to the United States, because the cost of providing the United States the “Special and Deferential Treatment” that the Trump administration is demanding is higher than cost of saying “No thanks”.
To follow the advice from Stephen Harper and others, one needs to “know what you’re getting before making any concessions.” And to build on that, one needs to know the value of the concessions.
Let’s focus on the latter: what is the most valuable trade in play in these negotiations? Hint: it’s not dairy. The U.S. economy has only one visible means of support at the moment and that is the tech sector boom based on investment in the infrastructure for artificial intelligence (AI) and data.
Here’s the interesting part: while the United States is a net exporter of energy (the “old oil”) and so can walk away from the train wreck in the Gulf, it is a net importer of the data (the “new oil”) that drives the AI economy. That data is very, very valuable.
Now, here’s where it gets kind of crazy. The value of data flows across borders is recorded as having zero value.
So when I say it is very, very valuable, there will be skeptics: “Dan, is this very, very valuable data in the room with us now?” they might ask. And my answer would make it sound even crazier as I would say something like, “It’s invisible”. And, yeah.
But let’s suspend judgement on my sanity and suppose that Nvidia did not become a US$ 5 trillion company (a mark it reached on 29 October 2025) by providing equipment to process low-value assets and posit that data is indeed very, very valuable.
Which brings us to the blank canvas: how are these international flows of valuable data structured – who are the exporters and who are the importers?
In the barter trade of free internet services for data, the United States exports the internet services through companies such as Alphabet, the company formerly known as Google, which sells internet search and other services, and Meta, the company formerly known as Facebook, which sells cat-picture sharing and other stuff.
The consideration these companies receive in return is data. This data does not flow into a money bank because there is no price attached – instead, they flow into a data bank, which then makes the company – Alphabet or Meta – very valuable to its shareholders and especially to its founders.
So here is the nub of the issue: if the consideration received is making one side of the trade phenomenally rich, is it crazy to think that the value of the data received exceeds considerably the value of the internet services provided?
So that’s me on the couch asserting this basic point – that this is a “beads for Manhattan” deal.
Now let’s take this to the international trade negotiations table.
Under the USCMA/CUSMA, the United States gets the flow of data from Canada tax-free and royalty-free. This is really important to the United States, not on grounds of principle, but on grounds of money: Canada’s attempt to impose a digital services tax to capture some of that value drew an immediate threat of nuclear tariffs; and Canada’s musings about a sovereign cloud have been immediately listed by the United States Trade Representative (USTR) as a “trade irritant”. Someone is paying attention to the value proposition – in Washington if not in Ottawa.
So if our negotiators go into this negotiation thinking they are giving away beads while US negotiators are tasked with ensuring they secure Manhattan, Canada signing onto the renewal of CUSMA locks in this massive freebie for the United States.
That’s bad enough, but the United States will also ask that Canada throw in sweeteners like expanded concessions for Canada to absorb U.S. excess supply of U.S.-subsidized dairy products.
And worse, it means signing onto absorbing the deadweight costs of excessive rules-of-origin certification, now required for every shipment using the agreement — pure economic waste that benefits nobody, not even the U.S. Treasury. A tariff at least generates revenue; compliance paperwork just burns money and time. Don’t get me started.
And it means committing to a market – and thus deepening exposure to that market –even though that market now regularly churns out unpleasantries such as the unilateral ending of de minimis exceptions and the sudden removal of regulatory harmonization for acceptance of Canadian standards for food and drugs, for one simple reason:
Because. It. Is. Isolationist.
Now, to keep this very simple, the proposition facing Canada – a trade-oriented nation – is whether to orient its economic policy around access to an isolationist market? If you don’t believe this is a good idea, then I have at least one reader to keep me company.
U.S. isolationism is not a superficial posture that can be reversed with a stroke of the pen by a future U.S. president issuing an Executive Order that reverses a Trump Executive Order the way Trump reversed Biden-era decrees. It emerges from a more fundamental shift in the political demography of the United States conditioned by the technological and economic conditions that now shape its political alignments: the university-educated demographic with progressive values, vested in a rules-based order of which they were the well-paid administrators, started to shrink around 2010 when the economy came to be ruled by college drop-outs, not post-grads, and that shrinkage is continuing and accelerating in the world of machine knowledge capital, in which young people are counselled to apprentice in the trades rather than to seek a university education.
Perhaps the greatest challenge will be in redefining the role of human capital and supporting investment in it. This will be particularly important for younger generations facing the questions of why and in which areas to invest in human capital when (to paraphrase the iPhone’s marketing slogan) “there is an AI for that” (Ciuriak 2024).
So, should we stay or should we go? In my submission on Canada’s reservation price in the CUSMA review, I argue that the trade flow genuinely at risk from tariffs is a fraction of what headline statistics suggest and that the costs of a deal that entrenches compliance burdens without addressing the full range of threats to trade from the United States, or offering any guarantee that what is agreed will be honoured, cannot be considered a good deal.
So let’s be prepared to let Mr. Trump play out his Dog Day Afternoon and put up his tariffs – and Canada can prepare to keep its data. Fair is fair.
 | A guest post by
|
Very interesting new perspective, and well written.