Donald Trump has begun implementing one of the central promises of his second presidential campaign. On 1 February, he ordered a 25% tariff on most imports from Canada and Mexico and an additional 10% tariff on Chinese imports, arguing that America’s three largest trading partners have failed to do enough about illegal immigration, fentanyl trafficking and economic imbalances. Canadian energy exports will face a lower 10% tariff, but the broader measures cover hundreds of billions of dollars in annual trade. Canada has already announced retaliatory tariffs on US goods, Mexico has promised countermeasures and China says it will challenge the measures through the World Trade Organization while responding as necessary. This is therefore no longer simply campaign rhetoric about economic nationalism. Trump is beginning an experiment with the world’s largest economy: can tariffs be used simultaneously to protect American industry, raise government revenue and coerce foreign governments without increasing prices and disrupting the integrated supply chains on which American producers themselves depend?
The first distinction is between tariffs as industrial policy and tariffs as bargaining instruments. These are related but different objectives. A protective tariff is designed to make imported goods more expensive so domestic production becomes relatively competitive. For that policy to work, the tariff usually needs to remain predictable long enough for companies to invest in American capacity. A coercive tariff works differently. Its purpose is to threaten another country with economic losses until that government changes behaviour, after which the tariff can be removed. The first mechanism depends on permanence. The second depends on credible reversibility.
Trying to use the same tariff for both objectives creates a contradiction. If Mexico believes tariffs will disappear immediately after making concessions on border enforcement, Mexican and American firms have less reason to relocate factories permanently. If companies believe tariffs are a durable part of Trump’s industrial strategy, Canada and Mexico have less reason to make temporary concessions because compliance may not remove them anyway. Industrial credibility requires stability; coercive credibility requires conditionality.
The economic incidence is equally important because tariffs are often described politically as taxes paid by foreign countries. Legally, the tariff is collected from the American importer bringing the product across the border. The ultimate cost then depends on bargaining power. A foreign producer may reduce its price to preserve access to the American market. The importer may absorb some cost through lower margins. Consumers may pay higher prices. Domestic competitors may increase their own prices because foreign competition has become more expensive.
The final distribution therefore is empirical.
But the tariff is not a cheque that Ottawa, Mexico City or Beijing simply writes to Washington.
This matters because Canada and Mexico are not ordinary foreign suppliers. After decades of integration under NAFTA and then the US-Mexico-Canada Agreement (USMCA), North American production operates through deeply interconnected supply chains. A car assembled in Michigan may contain components that crossed the Canadian or Mexican border several times during production. A tariff imposed each time value moves across borders can therefore increase the cost of something Americans consider domestically manufactured.
This is the supply-chain multiplication problem.
Traditional trade statistics assign an import to the country from which the final product enters.
Modern manufacturing distributes value creation across several countries.
Tariffs therefore can tax intermediate inputs used by American producers rather than simply final foreign products competing with them.
The automobile industry is particularly exposed. American, Canadian and Mexican plants were organised around the assumption that North America functions substantially as one production platform. Companies located particular stages where costs, logistics and specialised labour made them most efficient.
Reconstructing those networks inside the United States is possible in some sectors.
It is not instantaneous.
Factories require land, machinery, permits, suppliers and trained workers.
This creates the central short-run versus long-run distinction in tariff policy.
In the short run, supply chains are relatively fixed.
Tariffs therefore appear mainly as higher costs.
In the long run, firms can change sourcing and investment.
Domestic production may expand.
The economic case for tariffs therefore depends heavily on whether the long-run adjustment generates enough strategic and productive benefit to compensate for short-run consumer and producer costs.
Trump’s argument is essentially that previous governments placed too much weight on short-run efficiency.
Companies moved production abroad because labour was cheaper.
Consumers received cheaper goods.
Corporate supply chains became more efficient.
But communities losing factories bore concentrated costs that national GDP statistics did not fully capture politically.
This is the China shock argument extended to North America.
Trade produces aggregate gains while distributing those gains and losses unevenly.
A consumer saves $200 annually through cheaper imports.
A manufacturing worker loses a $60,000 job.
The consumer benefit is dispersed across millions of households.
The worker’s loss is concentrated geographically and personally.
Aggregate welfare can rise while particular communities deteriorate.
Economic nationalism begins from the claim that those distributional effects were treated as secondary for too long.
That criticism has substance.
The harder question is whether broad tariffs are the best remedy.
A tariff protecting a strategically important semiconductor supply chain can be justified through national-security externalities.
A universal tariff on products the United States has little comparative advantage producing may simply transfer income from consumers towards protected producers while reducing overall efficiency.
This is why targeting matters.
Industrial policy works best when government can identify a genuine market failure or strategic externality.
Broad protection assumes domestic production is desirable almost regardless of sector.
That is a much stronger claim.
Canada makes the problem particularly visible because the United States does not have an obvious strategic interest in economically separating itself from Canada. The two countries share one of the world’s closest defence relationships, enormous bilateral investment and deeply integrated energy markets.
Canadian oil flows into American refineries designed partly around heavier crude grades. Replacing that supply is not simply a matter of producing another barrel in Texas because different refineries are optimised for different inputs.
This explains Trump’s lower 10% tariff on Canadian energy.
Even the administration appears to recognise that interdependence creates constraints.
The broader political objective therefore extends beyond industrial protection. Trump has linked the tariffs explicitly to fentanyl and border security, invoking emergency economic powers to justify them.
This represents a significant expansion in how tariffs are used.
Trade policy becomes foreign-policy coercion.
The logic resembles sanctions.
America controls access to the world’s largest consumer market.
Foreign governments value that access.
Washington threatens to reduce it unless they alter behaviour.
This is market-access power.
The United States does not need to invade Mexico to influence Mexican border policy.
It can threaten economic costs large enough that cooperation becomes preferable.
Whether this works depends on the ratio of dependence.
Canada and Mexico send very large shares of their exports to the United States, making them highly exposed.
The United States trades heavily with both but possesses a much larger domestic economy.
This asymmetry gives Washington bargaining leverage.
Yet economic size is not the only variable.
Democratic governments also respond to national pride.
If Canadian Prime Minister Justin Trudeau appears to capitulate immediately under American threats, he faces domestic political costs. Retaliation can therefore be rational even when Canada knows a trade war hurts it more economically.
This is the audience-cost problem of economic coercion.
The stronger state asks: why doesn’t the weaker state simply concede?
The weaker government asks: what happens domestically if voters conclude foreign pressure determines national policy?
Nationalism can therefore make economically costly resistance politically rational.
Trump himself understands this mechanism because his political appeal depends heavily on refusing perceived foreign exploitation.
Canadian and Mexican leaders face their own versions of the same incentive.
Retaliatory tariffs are consequently designed partly for political targeting. Governments often place tariffs on products produced in electorally important regions rather than selecting goods randomly. The objective is to make influential domestic constituencies inside the initiating country feel the cost and pressure their government.
Trade wars therefore become political targeting through economic networks.
Canada does not need to damage the entire American economy equally.
It needs enough American businesses and voters to conclude the tariff policy is hurting them.
China learned this during Trump’s first trade war, targeting agricultural products produced heavily in Republican states.
The result demonstrates why tariffs can escalate even when economists on both sides recognise mutual losses.
Each government believes temporary pain can force the other to retreat first.
This is a war of attrition.
The side with greater tolerance for economic pain wins bargaining leverage.
But both sides become poorer while discovering who possesses more tolerance.
The inflationary consequences are particularly politically sensitive because Trump won partly by attacking the Biden administration over high prices. US inflation has fallen substantially, but the price level remains much higher than before the pandemic.
Tariffs now create a policy capable of increasing some prices deliberately.
That does not mean another 2022-style inflation surge is inevitable.
The effect depends on tariff duration, exchange rates, foreign producer margins, substitution and monetary policy.
But the direction of pressure is straightforward.
Imported products become more expensive unless somebody in the supply chain absorbs the tax.
Domestic competitors may gain room to increase prices too.
This creates a political contradiction between protection and affordability.
Voters can rationally support both domestic manufacturing and lower prices.
Policy cannot always maximise both simultaneously.
Trump’s political challenge will be convincing Americans that any higher short-term costs purchase something valuable enough to justify them: factories, jobs, border cooperation or strategic independence.
Without visible benefits, tariffs become another tax.
With visible industrial investment, they can become part of a broader economic restructuring.
The revenue argument is also weaker than it initially appears. Trump has discussed tariffs as a significant source of federal revenue and even suggested they could eventually substitute partly for domestic taxation.
Tariffs do generate government revenue.
But there is a Laffer-type constraint built into successful protection.
If tariffs actually cause imports to collapse and production to move domestically, the taxable import base shrinks.
A tariff cannot simultaneously eliminate imports and generate enormous permanent revenue from those same imports.
The two objectives conflict.
Maximum revenue generally occurs before maximum protection.
Again, policy needs to decide which objective dominates.
China presents a different strategic case from Canada and Mexico because the American relationship with Beijing includes military and technological competition. Reducing dependence on Chinese semiconductors, batteries, telecommunications equipment and critical minerals can therefore produce national-security benefits not captured by consumer prices alone.
The Biden administration accepted much of this argument.
It maintained several Trump-era China tariffs and added restrictions on advanced semiconductor technology.
The bipartisan shift is significant.
The old Washington consensus broadly assumed economic integration would create mutual prosperity and potentially encourage Chinese political liberalisation.
The new consensus increasingly assumes that economic dependence can become strategic vulnerability.
Trump deserves some credit for accelerating that intellectual shift.
Where his approach differs is scale.
Biden emphasised “small yard, high fence” restrictions around selected strategic technologies while subsidising domestic production.
Trump’s approach moves closer towards using tariffs across the entire economic relationship.
This raises the risk that de-risking becomes decoupling.
Selective diversification can improve resilience.
Complete economic separation between the world’s two largest economies would impose enormous costs globally.
The same logic applies even more strongly to Canada and Mexico.
North American integration itself can be a strategic advantage in competition with China.
A production network spanning the United States, Canada and Mexico gives American firms access to energy, minerals, labour and manufacturing capacity within a geopolitically friendly region.
Damaging that network while trying to reduce Chinese dependence could therefore be counterproductive.
This is why friend-shoring became attractive under Biden.
Move supply chains away from geopolitical rivals towards allies.
Trump’s tariffs blur that distinction by applying substantial pressure to allies and rivals simultaneously.
His theory is that allies can exploit America economically too.
The alternative theory is that preferential treatment for allies is precisely how economic coalitions become strategically useful.
This will become one of the central disagreements of Trump’s foreign policy.
Is American power maximised by extracting better bilateral terms from every partner individually?
Or by building blocs whose members accept some economic inefficiency because collective scale increases geopolitical power?
The first is transactional hegemony.
The second is institutional hegemony.
Trump strongly prefers the first.
The post-war American system was built largely around the second.
The USMCA itself illustrates the contradiction because Trump negotiated it during his first presidency as the replacement for NAFTA. It was supposed to create a revised but stable framework for North American trade.
Imposing broad tariffs only a few years later raises a credibility problem.
If negotiated trade agreements do not protect partners from sudden tariffs, what is the value of signing them?
Future governments may therefore demand stronger guarantees before making concessions.
This is the reputation cost of coercive flexibility.
Having the ability to change policy rapidly creates bargaining power today.
Using that ability frequently makes future commitments less valuable.
Again, power and credibility can trade against one another.
The next few days will show whether these tariffs actually remain in force for long. Trump may use them to obtain border or fentanyl concessions and then suspend them, demonstrating that the objective was primarily coercive.
If so, the economic damage could remain relatively limited while the political lesson becomes enormous: access to the American market can be conditioned on cooperation in areas unrelated to trade.
Other governments would notice.
Trade dependence would become geopolitical dependence more explicitly.
Countries might respond by diversifying away from the American market where possible.
That is the same weaponisation-of-interdependence paradox we saw with financial sanctions.
A network becomes useful as a weapon because others depend on it.
Using the weapon teaches others to reduce their dependence.
The United States possesses such an enormous market that diversification will be difficult.
But marginal changes accumulate.
Canada can deepen European and Asian trade.
Mexico can diversify investment.
China is already building alternative markets across the Global South.
Economic coercion therefore can produce adaptation over time.
This does not mean Trump should never use America’s market power.
It means the relevant calculation needs to include the long-term depreciation of leverage caused by using leverage today.
That is the strategic question beneath this weekend’s tariffs.
The United States possesses extraordinary economic power.
Trump intends to use it more aggressively.
The immediate targets are Canada, Mexico and China.
The broader audience is every country economically dependent on American markets.
The message is clear:
trade access is not simply commercial.
It is political leverage.
Whether that makes America stronger depends on what happens next.
If Canada and Mexico make concessions quickly, tariffs disappear and American objectives are achieved at limited cost, Trump will have demonstrated a powerful coercive tool.
If retaliation expands, prices rise and North American supply chains are disrupted without major concessions, the policy will demonstrate the limits of economic intimidation.
And if firms respond by permanently moving production into the United States, the industrial argument gains credibility even if consumers pay more.
These outcomes are not mutually exclusive.
That is why the policy should not be judged by whether tariffs are theoretically “good” or “bad”.
It should be judged by measurable results.
How much do prices rise?
How much new manufacturing investment occurs?
How much border or fentanyl cooperation changes?
How much retaliation occurs?
How long do tariffs remain?
How much government revenue is actually collected?
Those answers will tell us whether Trump’s economic nationalism is producing strategic leverage or merely redistributing the cost of interdependence.
For decades, American economic policy treated integration as something valuable enough that governments should avoid disrupting it unnecessarily.
Trump’s theory is almost the reverse.
Integration creates dependence.
Dependence creates leverage.
Leverage exists to be used.
The experiment has now begun.
The question is whether America’s trading partners are dependent enough that they change their behaviour before American consumers and firms decide the leverage is costing them too much.
