The Receipt
Canada’s trade diversification push is real, reasonable, and older than the crisis that gave it urgency. The “1.5 billion consumers” figure the government cites dates to at least October 2018. [2] The FTA architecture it rests on (CETA, CPTPP, CUSMA) was negotiated and signed under the Harper and Trudeau governments. Most of the agreements presented as new were concluded or launched under predecessors. The strategy is continuous across three governments and two parties. The presentation is not.
The economics literature, validated across institutions, shows that distance, economic mass, and infrastructure orientation constrain the pace of trade diversification. Canada’s trading infrastructure was built to move goods south. The government’s own $5 billion Trade Diversification Corridors Fund and its target of doubling non-U.S. exports by 2035 acknowledge this structural timeline. [15] Services exports are genuinely diversifying. Goods exports are not, at least not broadly, and the headline gains are partly a gold-price artifact. The gap between the rhetoric and the structural evidence matters because it shapes how prepared Canadians are for what the trade war costs while the decade-scale work gets done.
On August 22, 2026, Prime Minister Carney walked away from trade negotiations with the United States and delivered the case for Canada going it alone. Twenty trade and security deals across five continents. 1.5 billion consumers with tariff-free access. The fastest trade negotiation in Canadian history. A promise to double Canada’s market access over the next six months through new deals with ASEAN and India. [1]
Wanting less dependence on a trading partner imposing 50% tariffs is rational. The diversification push is the right strategy. The question is whether the evidence underneath supports the timeline the rhetoric implies, and what happens to Canadian businesses and workers who make decisions based on the gap between the two.
This article catalogues what Canada has done, who did it, what the economics literature says about the structural limits, and what realistic diversification looks like according to the available evidence. It is not an argument against diversification. It is a sanity check calibrated to the documented record.
A Strategy Older Than the Crisis
Trade diversification has been continuous Canadian policy for nearly two decades. The instinct is even older. Mitchell Sharp’s “Third Option,” adopted by the Trudeau government in 1972, was the first formal strategy to reduce economic dependence on the United States. [34] Diversification has recurred as a Canadian policy objective across governments since, even as some, notably Brian Mulroney’s, simultaneously pursued deeper continental integration.
Stephen Harper made trade agreements a central plank of economic policy. His government launched the Global Commerce Strategy in 2007 and replaced it with the Global Markets Action Plan in 2013. [4] Under Harper, Canada launched CETA negotiations with the European Union in May 2009 and reached agreement-in-principle by October 2013. [5] The Canada-Korea Free Trade Agreement, Canada’s first in the Asia-Pacific, was launched in 2005, stalled in 2008, and concluded in March 2014. [6] India CEPA negotiations were launched in 2010. [7] The Trans-Pacific Partnership was negotiated under his government, and by October 2014, Canada had concluded 22 foreign investment protection agreements. [8]
Justin Trudeau’s government finished the marquee deals and formalized the diversification strategy. CETA was signed in October 2016 and provisionally applied in September 2017. [5] The CPTPP was salvaged after the United States withdrew, signed in March 2018, and entered into force on December 30, 2018. [9] CUSMA was negotiated in 2017–2018 and entered into force in July 2020. In November 2018, the government launched the Export Diversification Strategy: $1.1 billion over six years with a target of increasing overseas exports by 50% by 2025. [10] That target was genuinely met a year early. Overseas exports reached $296 billion in 2024, up from $195 billion in 2017. [11] Global Affairs found roughly 30% of the growth was concentrated in volatile or potentially unsustainable drivers including gold and international education; excluding those categories, Canada was only slightly below target. [11] [12] Indonesia CEPA negotiations were launched in June 2021 and concluded in December 2024. Ecuador FTA negotiations were concluded in January 2025. India CEPA talks collapsed in 2023 after the Nijjar affair and have only recently been relaunched.
Mark Carney’s government has signed deals its predecessors negotiated and added new ones. The Indonesia CEPA was signed in September 2025 (concluded under Trudeau). [13] The UAE foreign investment protection agreement was signed in November 2025 (negotiations launched in 2016). [14] The Ecuador FTA was signed in July 2026 (concluded under Trudeau). The UAE CEPA was concluded in July 2026 after a 47-day negotiation that the UAE calls the fastest under its CEPA programme and Canada calls among the fastest it has completed. It is the one agreement genuinely initiated under the current government. [16] The China Strategic Economic Partnership reached in January 2026, the India Strategic Energy Partnership and uranium supply agreement signed in March, and the various memoranda of understanding documented in our earlier reporting are not free trade agreements. [17] The $5 billion Trade Diversification Corridors Fund, the internal-trade liberalization framework established by Bill C-5, and the Critical Minerals Production Alliance coordinating supply chains with allied countries represent a genuine escalation in domestic infrastructure and institutional commitment over predecessors. [15] [35]
The following table summarizes the key claims, what they sound like, and what the documented record shows.
| The announcement | What it sounds like | What the record shows |
|---|---|---|
| “More than 20 trade and security deals” | 20+ new trade agreements | Mix of FTAs (3), FIPAs (1), security MOUs, strategic partnerships, and letters of intent. No public itemized list. Most are non-binding. |
| “1.5 billion consumers” | New market access | Government figure since at least October 2018. [2] Represents 15 pre-existing FTAs. Unchanged by any Carney-era deal. [3] |
| “Tariff-free access” | All goods enter partner markets duty-free | Government’s own materials use “preferential access.” FTAs include phase-outs, carve-outs, supply management exclusions, and rules of origin. |
| Indonesia CEPA | New trade deal | Negotiations launched June 2021 under Trudeau. Concluded December 2024. Signed September 2025 under Carney. [13] |
| Ecuador FTA | New trade deal | Negotiations concluded January 2025 under Trudeau, announced February 4, 2025. Signed July 2026 under Carney. [3] |
| UAE FIPA | New investment agreement | Negotiations launched 2016. Signed November 2025 under Carney. [14] |
| UAE CEPA (47 days) | Fast new trade deal | Genuinely Carney-initiated: launched November 2025. Record speed claimed by both sides in their own terms. Not yet signed or ratified. [16] |
| “Double that number…over the next six months” | Doubling FTA consumer coverage | Requires concluding ASEAN and India FTAs by approximately February 2027. India talks launched 2010. No six-month precedent for either. [7] |
| “Record investment” | New productive capacity flowing in | Nearly half the 2025 FDI headline was M&A: foreign companies buying existing businesses, not building new ones. [30] |
None of this is an argument that the deals are bad or that signing inherited agreements is illegitimate. Signing is a necessary government function. The finding is narrower: the strategy is continuous, credit is shared across three governments, and the structural direction was set years before the tariff crisis began.
What Nine Years of CETA Tell Us
If a single trade agreement could demonstrate what diversification looks like in practice, it would be CETA. Canada has had preferential access to roughly 450 million EU consumers since September 2017. Nine years of data with a wealthy, rule-of-law partner bloc that shares two official languages with Canada. If market access translates directly into diversification, CETA is the showcase.
The growth is real. Canada-EU goods trade grew approximately 46% from 2016 to 2021 and 76% in nominal terms from 2016 to 2025. [18] The largest export gains since 2017 were in ores, slag, and ash (up 163%), precious stones and metals (up 87%), and mineral fuels (up 63%). Aircraft fell 19%. Machinery fell 9%. [18]
But the most granular available data suggest utilization is uneven and has not accelerated since the tariff crisis began. A Global Affairs Canada study published in July 2026, examining Canada-France bilateral trade, shows the Canadian preferential utilization rate falling to 49.2% in 2024, down from a peak of approximately 70.5% in 2021. [20] The study attributes the decline partly to petroleum products transiting outside EU customs territory and notes the rate dropped “below the 50% warning threshold.” The EU-wide picture is moderate: the EU-to-Canada utilization rate stood at 63.2% in 2024. [21] Even after nine years of preferential access, operational barriers, compliance costs, and routing rules continue to prevent nominal tariff access from being fully converted into claimed preferences.
The CPTPP tells a similar story at a similar pace. Merchandise trade with the seven Indo-Pacific CPTPP markets grew 42.9% since the agreement entered into force in 2019, with merchandise exports up 24.3%. [22] Real growth, spread across six years.
This is what success looks like. It took nine years to reach moderate preference utilization with the European Union, and that utilization is now declining. Every new FTA that Canada signs (Indonesia, UAE, ASEAN, India) should be measured against this baseline. The question is not whether the agreements are worthwhile. It is how long they take to translate into the kind of broad-based trade diversification the rhetoric implies.
Fighting Gravity
There is a reason diversification is slow, and it is not a lack of political will. The gravity model of trade, one of the most consistently validated frameworks in international economics, holds that trade volume between two countries is proportional to their combined economic weight and inversely related to the friction of doing business across their shared border. [23] Distance, regulatory difference, and infrastructure orientation all count as friction.
The United States is both the world’s largest economy and Canada’s only land neighbour. Canada’s highways, railways, pipelines, and electrical grid were designed to move goods south. These are not policy choices that can be reversed by signing agreements. They are physical capital built over more than a century. As the Fraser Institute’s Jock Finlayson and Steven Globerman concluded in June 2026, “the favourable attributes of the U.S. market for Canadian exporters will make it very challenging” to significantly reduce U.S. dependence. [23]
This is not a single think tank’s opinion. C.D. Howe Institute research reaches the same conclusion from a centrist vantage: market access on paper does not automatically create exporters in practice, because the binding constraints are domestic, not foreign. Regulatory compliance, product certification, distribution networks, after-sales service, and local market knowledge all represent fixed costs that Canadian firms must absorb before a new FTA generates a single dollar of new trade. [24] The Asia Pacific Foundation and CERIUM have argued that diversification “can no longer be reduced to extra-U.S. export growth or trade-agreement counting.” [25]
The government’s own spending confirms the structural gap. The $5 billion Trade Diversification Corridors Fund targets ports, freight railways, and roads to move goods to non-U.S. markets. [15] The Arctic Infrastructure Fund is upgrading the Port of Churchill into a deep-water gateway for European trade. LNG Canada shipped its first cargo in June 2025; the proposed Ksi Lisims floating LNG project is targeting 2029 for first production. These are decade-scale infrastructure projects. Finlayson, for his part, conceded the Corridors Fund has “genuine merit” and should “enhance the competitiveness of Canadian-produced traded goods regardless of the markets we are doing business with.” [26] The investment is the right response to the structural constraint. The constraint is the reason it takes a decade.
When Trade Deals Collide
Canada’s diversification ambitions do not exist in a geopolitical vacuum. CUSMA Article 32.10, sometimes called the China clause, requires any CUSMA party to notify the others at least three months before commencing FTA negotiations with a non-market economy. The other parties can review the full text of any resulting agreement and, if they choose, terminate CUSMA on six months’ notice. [27]
The treaty text applies specifically to commencing “free trade agreement negotiations with a non-market country.” [27] Canada’s January 2026 China arrangement was formally described by both governments as a strategic partnership supported by memoranda of understanding, not a free trade agreement. Trade lawyer Barry Appleton has argued publicly for an expansive reading, contending that the arrangement functionally fit within 32.10’s prohibited scope and that required notice was not given. [28] A 2018 analysis from the Centre for International Policy Studies reached the opposite conclusion, calling 32.10 “American posturing, not a veto” and observing that it is largely duplicative of the general withdrawal clause in Article 34.6. [29] The legal bite of the clause remains disputed. It also becomes moot if CUSMA lapses or is terminated, and the joint review of the agreement formally opened on July 1, 2026.
What is less disputed is the direction of pressure. In Carney’s account of the collapsed August 2026 trade talks, the United States attempted to extend restrictions beyond non-market economies, seeking to limit Canada’s ability to pursue FTAs with market economies such as India, ASEAN members, and Mercosur countries without U.S. approval. [1] The White House did not confirm these specifics. If Carney’s characterization is accurate, the structural tension is clear: the more Canada diversifies, the more friction it creates with the trading partner that still buys more than 70% of its merchandise exports. [18]
And even without geopolitical friction, each new FTA creates compliance complexity for firms already navigating multiple preference regimes. Rules of origin under CUSMA, CPTPP, and CETA each have distinct requirements. For a mid-size Canadian manufacturer, deciding which regime to claim preferences under, and maintaining the documentation to prove compliance, is itself a cost of diversification that the consumer-headcount rhetoric does not capture.
Two Stories, Not One
The diversification picture splits cleanly into two narratives, and the rhetoric does not distinguish between them.
The services story is genuinely positive. Services exports hit a record of approximately $240 billion in 2025, roughly one quarter of total exports, and have tripled since 2010. Forty-seven percent of services exports go to non-U.S. markets, compared to just 28% for goods. [18] Services drove all of Canada’s net export gains since 2022. Financial services, software, engineering, and education do not require a pipeline or a container ship. They face fewer geographic constraints, and they are already less U.S.-dependent than goods. This is real, structural diversification, and it is the strongest evidence that the strategy is working.
The goods story is narrower than it looks. Non-U.S. export share for goods and services combined reached 32.8% in 2025, the highest in four decades. For merchandise alone, the non-U.S. share rose to approximately 28%. [18] But gold accounted for roughly one-third of non-U.S. merchandise export gains that year and 93% of the growth in exports to the United Kingdom. [18] The U.K. became Canada’s second-largest export market partly because of gold prices, not because of a structural trade shift. Stripping out gold and energy, the goods diversification picture is thinner than the headline suggests. TD Economics found in February 2026 that the non-U.S. gains, led by the United Kingdom, China and parts of Europe, were driven by “a narrow set of commodities” that did not fully offset declines in U.S.-bound industrial goods. [19]
Australia offers a cautionary comparison. Australia’s China goods-export share fell roughly 12 percentage points in a single year after Beijing imposed retaliatory trade bans. It looked like rapid diversification. But an analysis by the Australian Strategic Policy Institute found that partner shares “contracted once China dropped its discriminatory bans” and that China still bought more than a third of Australia’s goods exports. [31] Coerced redirection under trade bans is not cultivated, durable diversification.
The pattern of conflating different categories extends beyond trade. This publication previously documented that the government’s “record investment” headline combines mergers and acquisitions, government bond purchases, and actual business capital expenditure: three different forms of capital with different economic implications, described with a single word that implies productive capacity is being built. Nearly half the 2025 FDI headline was foreign companies buying existing Canadian businesses, not building new ones. [30]
There are genuine structural levers within Canada’s control. Removing interprovincial trade barriers could boost GDP by roughly 7%, or $200 to $210 billion, according to research by the IMF and the University of Calgary’s Trevor Tombe. [32] Bill C-5 established a mutual recognition framework for federal trade barriers, effective July 1, 2025, but provincial barriers, which account for most of the drag, have been slower to fall.
There is also early evidence that infrastructure built under previous governments is beginning to translate into geographic diversification. The first full year of expanded Trans Mountain pipeline capacity enabled crude oil exports to Indo-Pacific markets in 2025, a concrete example of decade-scale infrastructure investment producing the shift that trade agreements alone could not. [18]
The government’s own targets reveal the distinction between access and utilization. Carney says Canada can double the number of consumers nominally covered by trade agreements within six months, from 1.5 billion to roughly 3 billion, through ASEAN and India deals. [1] That is a measure of legal market access. The separate 2035 target of doubling actual non-U.S. export value is a measure of realized trade. Those are not competing timelines for the same outcome. They are two different achievements operating on very different clocks. BCG’s Centre for Canada’s Future assessed the 2035 goal as possible “by scaling roughly 50 priority goods and services, if key barriers are removed,” narrowing focus from approximately 1,200 export categories to about 50. [33] That assessment, and every credible institutional analysis that has examined the question, concludes that the United States will remain Canada’s largest trading partner for the foreseeable future.
What Would Change This Assessment
The structural argument in this article would weaken or fail under the following documented conditions:
- If the Canada-to-EU CETA preference utilization rate exceeds 75% while Canadian non-U.S. merchandise exports also broaden materially beyond commodities across multiple manufacturing categories, the claim that market access translates slowly into realized diversification would weaken.
- If the U.S. share of Canadian merchandise exports remains below 65% for two consecutive calendar years and the shift persists after excluding gold and energy, the claim of strongly binding structural concentration would need substantial revision.
- If Canada brings both an India CEPA and a Canada-ASEAN FTA from their current negotiating status to signed agreements by February 2027, the article’s claim that nominal market-access expansion necessarily moves on decade-scale timelines would fail.
- If by 2030 Canada’s U.S. share of goods exports falls below 60% on a three-year average, without the change being predominantly explained by gold, energy-price effects, or a collapse in U.S.-bound trade, the structural-concentration thesis would require major revision.