The Receipt

In January 2026, Canada and China announced a "Preliminary Joint Arrangement" to reset trade relations. Canada codified its side in the Canada Gazette: a 49,000-vehicle EV import quota at 6.1% tariff, growing 6.5% annually, with no sunset clause. Canada also extended steel and aluminum tariff remissions and cancelled previously announced tariffs on solar panels and semiconductors from China. These are enacted regulatory instruments.

China's side is structured differently. The canola seed tariff reduction, from roughly 85% to 14.9%, was formally implemented through MOFCOM's anti-dumping process. That ruling provides a five-year legal form distinct from the temporary suspensions covering other products, though public evidence does not establish how much the final rate was shaped by the investigation alone versus the broader political arrangement. The tariff suspensions on canola meal, peas, lobster, and crab, collectively worth roughly $2.4 billion annually, expire December 31, 2026. No extension had been announced as of publication. Canola oil remains at 100%. Pork remains at 25%. Neither was addressed in the deal.

In the Global Affairs Canada backgrounder, the only publicly available Canadian government summary, China's agricultural commitments are described three times using the phrase "Canada expects." Canada's EV commitments use language that was subsequently codified in law. The full text of the arrangement has not been published. No published document specifies what happens if a three-year review finds that Canada's expected benefits have not materialized.

The deal was presented as a balanced exchange: canola for EVs. Previous reporting in this series examined what actually arrived on the EV side. This piece asks a different question: are the two halves of the deal built the same way? One side is Canadian law, published in the Canada Gazette, enforced by federal agencies, operative regardless of whether the other side follows through. The other side is a set of tariff suspensions, most of which expire at year-end, described in a backgrounder using the word "expects."

That structural mismatch has consequences beyond any tariff schedule. If the concessions expire, Canada's laws remain. If the relationship deteriorates, Canada's regulations are still on the books. The deal's only accountability mechanism, a three-year review, does not specify what happens if expected benefits fail to materialize. The question isn't whether the arrangement delivers short-term value. It's whether the two sides are bearing the same kind of commitment.

What Canada Committed

Canada's side of the arrangement is built on enacted regulatory instruments. The centrepiece is the EV import quota: 49,000 Chinese-manufactured vehicles per year at the 6.1% Most Favoured Nation tariff rate, growing by 6.5% annually. [1] This quota was codified in the Canada Gazette through SOR/2026-32, is administered by Global Affairs Canada through import permits, and is enforced by the Canada Border Services Agency. [1] It contains no sunset clause. It is Canadian law.

Beyond the EV quota, Canada extended and expanded tariff remissions on Chinese steel and aluminum. Sixty-six product categories that had been receiving remissions were extended through 2026, eleven of them fully duty-free and fifty-five at partial relief. Forty-nine company-specific remissions were maintained. Thirteen new product categories were added to the remission list, effective retroactively from January 1, 2026. [5]

Canada also cancelled previously announced tariffs on solar panels and semiconductors from China. The 2024 Fall Economic Statement had declared an intention to impose these tariffs starting in 2026. [12] Under the arrangement, they were formally dropped before they were ever collected. [2] This distinction matters: Canada did not lower an existing tariff. It withdrew a planned measure before it took effect. The cancellation may still have affected investment expectations and bargaining value, but it surrendered a future policy instrument rather than removing duties already being collected.

Taken together, Canada's commitments share three structural characteristics. They are codified in regulation or published in official instruments. They have no stated expiry dates tied to the arrangement (the steel and aluminum remissions run through 2026 but were pre-existing policy extended, not created by the deal). And they are enforceable through domestic institutions without requiring action from the other side. A future Canadian government could amend or repeal any of these instruments through the same domestic machinery used to create them, but unless it does, they remain operative.

What China Committed

China's side of the arrangement covers multiple agricultural products, but they differ significantly in origin, duration, and legal form.

The largest single line is canola seed, worth approximately $4.0 billion annually in Canadian exports to China before the tariff escalation. [6] [7] [19] China's combined tariff on Canadian canola seed was reduced from roughly 85% to 14.9%, effective February 28, 2026. [3] China implemented this reduction through MOFCOM's anti-dumping process, which provides a five-year legal form distinct from the temporary suspension instrument used for the other agricultural products. [3] [4] Public evidence does not establish how much the final 5.9% anti-dumping rate was determined by the investigation alone versus the broader political arrangement. The rate is set for five years, subject to MOFCOM's administrative review provisions, which allow for interim reviews at anniversary dates but are discretionary and not automatic. [10] [16]

The remaining agricultural concessions are different. They are tariff suspensions, created directly by the arrangement, and they expire. Canola meal, which had faced a 100% tariff, was suspended to 0%. Tariffs on Canadian peas, lobster, and crab were likewise suspended. All suspensions took effect March 1, 2026 and expire December 31, 2026. [2] No extension has been announced as of publication.

These expiring concessions are collectively worth roughly $2.4 billion annually: canola meal at approximately $918 million, peas at roughly $700 million, lobster at $472 to $604 million depending on year, and crab at approximately $173 million. [6] [8] [9] [11] [13] [14] [18]

The proportionality deserves direct statement. The canola seed line, at $4.0 billion, is larger by value than all expiring concessions combined. More than a third of the cited pre-disruption agricultural trade value is covered by tariff suspensions that expire in December; approximately two-thirds is carried by the canola seed tariff, which was implemented through a five-year anti-dumping determination rather than a temporary suspension. (These values are based on 2023-2024 Canadian export data to China and may not reflect trade volumes during the 2025 tariff-escalation period.)

Two significant products were excluded entirely. Canola oil still faces a 100% tariff as of publication. [2] Canada exports canola as seed, meal, and oil; the deal covers two of three. Because canola oil is produced by crushing seed in Canada, exporting oil rather than raw seed retains more processing activity domestically. Pork, subject to a 25% retaliatory tariff imposed in March 2025, was not addressed. [5] [17] Beef market access was described as "resumed," restoring a pre-existing suspension rather than creating new access, with the government characterizing it as "accelerate the resumption" with no published timetable. [2]

The Language of the Deal

Because the full text of the arrangement has not been published, public understanding of its terms depends on government summaries. The most detailed Canadian source is the Global Affairs Canada backgrounder issued January 16, 2026. [2] The most detailed Chinese source is a MOFCOM press conference Q&A from January 22, 2026. [4] The two documents describe the same deal in structurally different language.

In the GAC backgrounder, Canada's EV commitments are introduced with the phrase "Canada intends to provide." That language was subsequently codified in law through SOR/2026-32. [1] China's agricultural commitments are introduced with the phrase "Canada expects," which appears three times in the document. [2] "Canada expects" that tariff suspensions will take effect. "Canada expects" resumed access for beef and other products. The Canadian backgrounder used less categorical language for China's future actions than for Canada's own intended actions. That drafting choice is consistent with China retaining control over the domestic measures required to implement its side, though it may also reflect standard diplomatic caution about stating another government's future administrative decisions.

MOFCOM's press conference uses different framing. China "will fully consider" Canada's requests "within the framework of established rules" and will issue a "final ruling based on facts and evidence." [4] This is conditional and process-dependent language. It positions China's concessions as outcomes of a legal and administrative process, not as political commitments made at a negotiating table.

China's public framing also treats its concessions as reciprocal responses to Canada's tariff reversals, not as new market access. From Beijing's perspective, these are WTO-consistent adjustments tied to Canada's decision to stop escalating. That framing positions them as structurally revocable: if Canada's posture changes, China's tariff adjustments can change with it.

In the January public summaries reviewed, neither government used treaty-style obligation language for China's agricultural measures. Canada says "expects." China says "will consider within rules." China subsequently implemented the non-seed tariff suspensions substantially as described, which matters more than January verb choices for assessing near-term outcomes. But the structural question is not whether China followed through in 2026. It is whose domestic law has already changed. Canada's EV quota exists regardless of what either backgrounder says. China's tariff suspensions depend on administrative decisions that can be reversed through the same administrative process that created them.

What the Deal Does Not Cover

The government's summary of the arrangement's agricultural benefits references "$4 billion in canola seed" and "$2.6 billion in other agricultural and seafood exports." [2] [15] These figures describe the pre-tariff-escalation market that was disrupted, not the value of new market access created. Several of the products counted toward this total remain blocked or unaddressed.

Canola oil is the most significant gap. China's 100% tariff, imposed in March 2025, remains in place and was not part of the arrangement. [2] Oil is the product where the most processing value is added in Canada. Seeds are crushed domestically into meal and oil; exporting raw seed sends the crushing margin offshore. The arrangement covers seed and meal but excludes oil, which means the product most aligned with Canadian value-added processing is the one that remains shut out.

Pork faces a 25% retaliatory tariff that was not addressed. [5] Beef was described as "accelerate the resumption" of pre-existing access, not as new market opening, and no timetable or regulatory instrument has been published. [2] Pet food and animal genetics exports were referenced in similar language.

The government's aggregate agricultural trade figure does not publicly disclose a product-by-product breakdown sufficient to distinguish between restored access, temporary access, and prospective access. Without that breakdown, it is not possible to determine how much of the cited value is covered by measures already in force and how much depends on future action.

The Three-Year Review

The arrangement includes one formal accountability mechanism: a three-year review clause. The GAC backgrounder states that "there will be an opportunity to review its progress and implementation in three years to assess and confirm if all expected Canadian benefits have materialised as anticipated." [2]

The clause confirms that one side's benefits require assessment. Canada does not need a review to determine whether its EV quota was enacted; it is published in the Canada Gazette. A review is needed because the benefits Canada was promised depend on the other side's ongoing conduct.

What the review clause contains is a date: three years from signing. What it does not contain, in any published document, is a set of benchmarks defining what "materialised as anticipated" means. It does not specify what happens if the review concludes that expected benefits have fallen short. There is no snapback mechanism that would restore Canada's pre-deal tariff levels. There is no renegotiation trigger. There is no binding arbitration provision. There is no published framework for how the review would be conducted or who would participate. [2]

The review may provide a political forum for assessing performance, but the public summary does not describe what remedies are available if either side is dissatisfied. Canada's commitments operate through domestic law that does not require review to confirm. The published review language addresses only "expected Canadian benefits," a category that depends on China's ongoing administrative decisions. A review without published benchmarks or remedies is a formal structure. Whether the unpublished arrangement contains stronger provisions is unknown.

The Architecture of the Arrangement

This is not the reciprocal, treaty-style architecture typical of a conventional free-trade agreement. In standard bilateral trade agreements, both sides make commitments of comparable legal durability, typically through instruments that bind both parties for the same term. The Canada-China arrangement uses a different structure: Canada's EV measure and China's agricultural measures were each implemented through domestic legal instruments with different forms and different expiry structures.

A bargain can be politically reciprocal even when each side implements it through different domestic legal procedures. Differences in implementation form do not necessarily establish differences in political commitment. But they do establish differences in what each side would need to do to reverse course. Canada would need to amend or repeal a published regulation. China would need to not extend a tariff suspension past its scheduled expiry.

In our assessment, that is the finding the documented record supports: the two sides' commitments differ in form and duration. Canada's are domestic regulation with no stated expiry, enacted in the Canada Gazette and enforced by the CBSA. China's non-seed suspensions expire in six months, and the canola seed tariff rests on a five-year MOFCOM determination whose relationship to the political arrangement is publicly unclear. The only publicly described accountability mechanism specifies no consequences. None of this alleges that either government acted improperly. Temporary tariff suspensions are a common diplomatic instrument, and the arrangement may deliver significant short-term value to Canadian farmers who faced near-zero exports to China in 2025. This analysis also does not estimate the EV quota's consumer benefits, effects on Canadian auto employment, or potential investment resulting from Chinese market entry. Those outcomes may be material, but they do not determine whether the two sides' commitments have matching duration and legal form.

The strongest case against this reading deserves full weight. Durability is not the only measure of value: a temporary restoration of billions of dollars in trade during an acute farm-market crisis can be worth more than a permanent concession with little near-term use. Canada retains the sovereign ability to amend or revoke its EV quota through the same regulatory process used to create it, so the absence of an automatic snapback does not mean the absence of Canadian leverage; China knows that political non-performance could prompt Canada to restore restrictions. Temporary concessions are standard diplomatic practice and are routinely extended when the underlying relationship is stable, and the three-year review clause exists precisely to create a framework for extension and deepening. On this view, the canola seed tariff, though formally issued through MOFCOM's process, would not have been resolved at 14.9% without the political context of the broader arrangement, and demanding binding treaty-level commitments from a preliminary diplomatic arrangement applies the wrong standard to the right outcome.

This is the third article in a series examining the Canada-China trade arrangement. The first documented what Canada gave up to restore its canola market. The second examined what actually arrived on the EV side. Together, the series documents what was traded, what was delivered, and whether the exchange is structurally balanced. On that last question, the documented record shows two sides using different legal instruments with different time horizons, implemented through an arrangement whose full text has not been published.

What Would Change This Assessment

The argument above rests on documented structure. Specific developments would weaken or overturn it:

  • If China extends the canola meal, pea, lobster, and crab tariff suspensions beyond December 31, 2026, the "expires in December" finding would need to be updated. An extension of less than twelve months would postpone the expiry concern without eliminating it. A multi-year extension would materially narrow the duration asymmetry. Indefinite or standing relief would substantially defeat the duration finding.
  • If the full text of the Preliminary Joint Arrangement is published or disclosed through Parliament, access-to-information, or authenticated leak, and reveals binding commitment language, enforcement provisions, or a dispute resolution mechanism not described in the GAC backgrounder, the "paper architecture" characterization would weaken.
  • If credible evidence emerges that the 5.9% canola seed anti-dumping rate was directly negotiated as part of the bilateral arrangement, the article's distinction between the anti-dumping determination and the temporary suspensions would lose its structural significance. Such evidence could include official statements connecting the rate to the bargain, negotiation records, parliamentary testimony, or investigation outcomes that departed unexpectedly from the evidentiary record.
  • If Canada and China publish benchmarks, remedies, an automatic snapback, or another enforceable response mechanism before or during the three-year review, the assessment that the review lacks publicly specified consequences must be revised.
  • If Canada adds an expiry date, reciprocity condition, or automatic review provision to the EV quota regulation, the duration asymmetry would narrow from the Canadian side.