Are tariffs ever good?
The textbook says a tariff is a tax you levy on yourself. A serious movement says the textbook failed. Both can be partly right — and the answer turns on four narrow doors.
Als Debattengraph anzeigenThe pro-tariff revival
“To me, the most beautiful word in the dictionary is tariff. It’s my favorite word.”
— Donald Trump, Economic Club of Chicago, October 2024 — preceding the April 2025 “Liberation Day” reciprocal-tariff program
In April 2025 that favorite word became the largest peacetime tariff program in a century: a blanket 10% on nearly everything, “reciprocal” rates reaching far higher on dozens of trading partners. And this is not one man’s whim. Biden kept every China tariff Trump imposed in 2018. A decade-old intellectual movement has spent years arguing that the free-trade consensus failed — and the people it failed are voting accordingly.
Start with the object itself. A tariff is a tax on imports: it raises the price a domestic buyer pays for a foreign good. The populist promise rides on one claim about who hands over that tax — the claim that the foreign exporter pays it, so the revenue is money extracted from China rather than from Americans. That single claim is the most-measurable thing in the whole debate, and Stage 2 takes it apart. Here we hold a different question first: not whether the revival is right about who pays, but why a movement this serious decided the consensus was worth attacking at all.
“A nation that simply consumes what others produce, financing the gap with debt and asset sales, is not a healthy economy. Trade is supposed to be an exchange of goods, not a one-way transfer of wealth and productive capacity.”
— Oren Cass, American Compass (national-conservative case for tariffs as industrial policy)
Is the pro-tariff turn a serious correction or a populist confusion?
For seventy years, “tariffs are bad” was the closest thing economics had to a settled answer. A national-conservative movement now says the settlement was a mistake that hollowed out the country. Are they diagnosing a real failure — or just selling a slogan?
A real grievance, and what to do with it
“The economic consensus of recent decades held that the composition of a nation’s output didn’t matter — potato chips, computer chips, what’s the difference? It turns out the difference is enormous. A worker-centered trade policy has to care about what we make, not just what we consume.”
— Oren Cass, paraphrasing the American Compass case in The Once and Future Worker
Cass is making the strongest version of the national-conservative argument, and it deserves to be heard at full strength rather than caricatured. His claim is not that tariffs are free, but that the efficiency frame priced the wrong things: it counted cheaper consumer goods and never counted the productive base, the regional labor markets, or the political stability that a manufacturing economy underwrites. Markets optimize for the present; a country has to live in the future. On that view, accepting a measurable efficiency cost to keep the capacity to make things is not economic illiteracy — it is a different and defensible objective function. The open question is whether the tariff is the instrument that actually delivers it.
“Trade with China cost American workers in import-competing places dearly, and the labor market did not adjust the way the textbook promised. But the answer is not to relitigate trade — the jobs are not coming back from a tariff. The answer is to help the people and places that were left behind.”
— David Autor, on the policy lesson of the “China shock” research
Autor co-authored the research that proved the grievance, which makes his verdict harder to dismiss as elite denial. The magnitude of the shock — the 2.0 to 2.4 million manufacturing jobs lost between 1999 and 2011, the decade-long failure of those communities to recover — is laid out in the sibling walkthrough Is free trade always good? (Stage 2, the China shock), drawing on the integration history in Economic History Ch.18 and Ch.17. The damage is settled fact. What Autor denies is the causal claim the tariff program rests on: that reversing the trade flow now reverses the damage. The capacity is gone, the supply chains relocated, the workers aged out. A tariff in 2025 raises the price of imports without rebuilding what closed in 2005.
Where this leaves us
The revival is not a punchline. It rests on a genuine failure that the economics profession under-weighted for twenty years, and the anger driving it is earned. But a real disease does not validate every medicine, and the medicine on offer is a tax the populists insist someone else will pay. Whether that is true is not a matter of opinion. It is the most-measurable claim in the entire dispute, and the model that settles it is the one the slogans never show you.
Start with the simplest question, the one the slogan gets backwards. When America puts a tariff on Chinese steel, who actually pays it?
The textbook case against
“China is paying us billions of dollars in tariffs. Our farmers, manufacturers, and country are doing GREAT!”
— the recurring “China pays” framing of the 2018–2025 tariff program
The single most-repeated claim in the tariff debate is also the most-measurable. And it is false in a specific, checkable way. To see exactly how, you need the one diagram the slogan never includes.
A tariff raises the domestic price of an imported good from the world price $P_w$ to $P_w + t$. Follow the surplus. Consumers face the higher price, so consumer surplus falls. Domestic producers, now shielded, sell more at the higher price, so producer surplus rises. The government collects the tariff on whatever is still imported. Add those up and a gap remains: two triangles of value that consumers lose and that no one recovers — one from domestic production that costs more than the import it replaced, one from purchases that simply never happen at the higher price. That gap is the deadweight loss, the irreducible waste the tariff manufactures.
With the tariff raising the price by $t$, the deadweight loss is the sum of the production-distortion and consumption-distortion triangles:
$$DWL = \tfrac{1}{2}\,t \cdot \Delta Q_{\text{production}} + \tfrac{1}{2}\,t \cdot \Delta Q_{\text{consumption}}$$Every dollar of protection is bought with a cost that lands in no one’s pocket.
A tariff is a tax that makes imports more expensive. Consumers pay more; some of that extra money goes to domestic producers, some to the government — and some just evaporates. It is pure waste: the cost of making things at home that could have been bought cheaper abroad, plus the purchases people would have made but now can’t. The government’s tariff revenue is real, but it is money taken out of your own citizens’ wallets, not out of the exporter’s.
So who pays? In the base model, the importing country’s own buyers do — the tariff lands on them, not on the foreign exporter. But the textbook is honest enough to leave one door open, and it is the most important door in this walkthrough. A large country — one that buys enough of a good to move its world price — can use a tariff to push that world price down. The foreign exporter, desperate not to lose the giant market, eats part of the tariff by cutting the pre-tariff price. That captured discount is a terms-of-trade gain, and if it exceeds the deadweight loss, the tariff makes the imposing country richer at the rest of the world’s expense. This is the optimal tariff, and it is the strongest economic version of “China pays” — the one case where the slogan touches a real mechanism. It is also the hinge into Stage 3, where we push it to its limits.
The full partial-equilibrium apparatus — the incidence diagram, the deadweight-loss triangles, and the optimal-tariff condition — lives in Economics Ch.2 §2.6 (International Trade). The comparative-advantage baseline that makes free trade the default — Smith’s 1776 demolition of mercantilism, Ricardo’s 1817 step — sits in the lineage at History of Economic Thought Ch.3 (Classical political economy); the doctrine the revival partly rehabilitates, mercantilism, is one chapter back. The sibling walkthrough Trade theory from Hume to gravity traces how that baseline was built; this walkthrough takes it as given and asks when a tariff escapes it.
“China is paying us billions of dollars in tariffs.”
— the “China pays” claim, restated through the 2018–2025 program
Does China pay the tariff?
The whole political case for tariffs rests on a claim about who hands over the money. The textbook says the importer’s own consumers. The optimal-tariff door says “sometimes the exporter, partly.” The 2018 data says which one was true.
The baseline and the one door it leaves open
“Free trade, one of the few totally agreed-upon propositions in economics, has the rare distinction of being a doctrine that is true but not obvious — and a doctrine that practical men, hardened to most subtleties, find almost impossible to believe.”
— Paul Krugman, “Ricardo’s Difficult Idea” (1996)
Krugman — who won his Nobel partly for showing where the simple free-trade story breaks — is here defending the baseline, which makes the defense weightier. Comparative advantage is not a guess; it is a theorem, and the gain from trade survives every generalization Ricardo could not have imagined. For a small country with no power over world prices, the verdict is total and admits no exception: a tariff is a pure self-inflicted tax, deadweight loss with nothing on the other side of the ledger. That is the bedrock the revival has to dislodge, and on most goods it cannot.
“A large country can improve its terms of trade with a tariff. The optimal tariff is positive. This is not a heterodox claim — it is in every graduate trade textbook, and it is the one case where protection raises national welfare.”
— the optimal-tariff theorem, the textbook’s own concession
This is the door, stated by the mainstream against itself. The optimal-tariff result is uncontested theory: a country big enough to move world prices can capture a terms-of-trade transfer that beats the deadweight loss. It is the honest exception the slogans grope toward without understanding. The whole question of the next stage is what that door actually licenses — whether it survives retaliation, whether it justifies breadth, and whether any of the other arguments for protection can clear the same bar. The mercantilist instinct the revival rehabilitates, traced at History of Economic Thought Ch.3, was never wholly empty; it was just wrong about almost everything except this.
Where this leaves us
For a small country, the verdict is closed: a tariff is a tax with no offsetting gain, and “China pays” is false on the goods America taxed. But the textbook itself leaves one door open — the optimal tariff, where a large country shifts the world price and makes the foreigner pay part of the bill. The pro-tariff revival’s strongest economic argument lives behind that door. So, it turns out, does its biggest problem.
There is one case where the textbook itself says a tariff can make you richer — and three more arguments it can’t fully dismiss. Walk through all four and the same shape appears every time: a real exception, and a condition almost no one meets.
The legitimate-exception debate
“We’re pursuing a ‘small yard, high fence’ approach — narrowly scoped protections around a small number of technologies with national-security implications. The goal is not to decouple from the global economy. It’s to ensure that a handful of foundational technologies don’t end up in the hands of those who would use them against us.”
— Jake Sullivan, U.S. National Security Advisor, Brookings Institution, April 2023 (the “new Washington consensus” speech)
The most serious case for protection in 2025 isn’t economic at all. It is that some dependencies are too dangerous to price — that there are goods you should be able to make even if buying them abroad is cheaper, because the day you need them, the seller may be your adversary. This is the pro-tariff argument the mainstream has actually moved toward. So take the whole case at its absolute strongest, four arguments deep, and ask where the conditions truly hold.
There are exactly four arguments under which a tariff can raise the imposing country’s own welfare. Each is real as theory. Each comes with a condition that is rarely satisfied. Take them in turn — one at-strength claim, one qualification, each.
1. The optimal tariff, pushed to its limit. Stage 2’s in-model exception: a large country with market power captures a terms-of-trade transfer that can exceed the deadweight loss. The qualification is fatal in practice. The optimal tariff is a beggar-thy-neighbor policy — the gain comes straight out of the trading partner’s hide — and partners retaliate. Model two large countries each setting their best tariff against the other and you get a Nash trade war: both impose positive tariffs, both shrink trade, and both end up worse off than under free trade. It is the prisoner’s dilemma of trade policy, and it is exactly why the WTO exists — to let countries jointly precommit to the cooperative outcome neither can reach alone. The game-theoretic apparatus is Economics Ch.6 §6.5 (Cournot, Stackelberg, and strategic interaction). The exception is genuine only for a large country acting unilaterally and unretaliated-against — a condition the existence of the WTO and the reality of retaliation make rare.
2. Infant industry. When building an industry generates learning-by-doing that spills beyond the firm — the social return exceeds the private return — the market under-invests, and temporary protection can be welfare-improving. The argument is theoretically valid, and the historical claim it rides on is real: the United States and Germany both industrialized behind tariff walls in the nineteenth century. The qualification is the conditions. Genuine externalities, a credible sunset clause, and a capture-resistant state that enforces export discipline — all three, or the policy rots into permanent shelter. For every South Korean POSCO that grew into a world-beating producer, there are dozens of Latin American import-substitution industries that consumed subsidies forever and never grew up. And the historical claim is murkier than the slogan: Douglas Irwin shows the nineteenth-century US tariff was mostly about revenue, not strategy, and that growth accelerated as barriers fell — the record is in Economic History Ch.8 §8.3 (Industrialization beyond Britain). The framework confrontation at full strength — Hamilton, List, and Ha-Joon Chang against Smith and Ricardo — is owned by the sibling walkthrough Smith vs. List on national economy; here it is one door among four.
3. National security and strategic goods. Dependence on a geopolitical rival for critical goods — semiconductors, rare earths, pharmaceutical precursors, defense inputs — is a genuine externality that market prices do not capture. The cheapest supplier in peacetime can be the chokepoint in a crisis, and the COVID supply-chain shock and the concentration of advanced chip fabrication in the Taiwan Strait made that abstraction concrete. Targeted protection of genuinely strategic goods can be justified even at a measurable efficiency cost — this is the strongest post-2020 exception, and the one the profession has actually moved toward. The qualification is the whole game: it licenses targeting, not breadth. “National security” is the single most-abused justification in the trade-policy book — every protected industry from sugar to steel has claimed it. The argument is for a small yard with a high fence, not a wall around the entire economy. The adjacent live debate over the modern industrial-policy program — the CHIPS Act, the Inflation Reduction Act, the spending side rather than the tariff side — is adjudicated in the sibling walkthrough Is industrial policy back, and is it justified?
4. Retaliation and bargaining leverage. A credible tariff threat can extract concessions — market access, intellectual-property protection, subsidy cuts — worth more than the deadweight loss it costs to make the threat believable. As a repeated-game, credible-commitment argument it is coherent. The qualification is the empirical record, and it is not encouraging: the 2018–2020 trade war’s “Phase One” deal was largely unfulfilled, the costs landed immediately on American farmers and importers, and the strategic gains stayed speculative. The bargaining empirics are owned by the sibling walkthrough Is free trade always good? (Stage 3, trade wars); one door here, no re-derivation.
Four doors. Each opens onto a real room. Read them together and the pattern is unmistakable: every exception is a scalpel, precise about the conditions under which it cuts. The development-policy framing of the learning-externality and strategic-goods cases sits in Economics Ch.20 §20.8 (Trade and development), and the East-Asian developmental record behind the infant-industry door is the spine of Economic History Ch.17 §17.3.
“Small yard, high fence: narrowly scoped protections around a small number of technologies with national-security implications.”
— Jake Sullivan, Brookings, April 2023
Is “national security” a real reason for tariffs, or a universal excuse?
The strongest case for protection in 2025 is the one economists have actually warmed to: some dependencies are too dangerous to leave to the market. It is also the case every rent-seeking industry has learned to imitate. Where is the line?
Four exceptions, argued at strength
“Under oligopoly with increasing returns, a government subsidy or protection can shift profits from foreign firms to domestic ones. The strategic-trade insight is that with imperfect competition, free trade is no longer guaranteed to be optimal.”
— the Brander-Spencer strategic-trade argument; see History of Economic Thought Ch.10 (Counter-revolution)
This is the most theoretically respectable case for protection, and it comes from inside the modern mainstream — Krugman, Brander, and Spencer built it in the 1980s. Where markets are oligopolistic and run on scale economies, the perfect-competition result that free trade maximizes welfare no longer holds, and a well-targeted intervention can shift oligopoly rents homeward. Combine that with the national-security externality and you have the steelman of strategic protection at full strength: there are real conditions under which a tariff or subsidy genuinely helps the country that imposes it.
“The power of producing wealth is therefore infinitely more important than wealth itself. A nation must sacrifice and give up a measure of material prosperity in order to gain culture, skill, and powers of united production.”
— Friedrich List, The National System of Political Economy, 1841
List is the infant-industry tradition’s sharpest voice, writing against British free-trade orthodoxy and arguing that Britain itself had used a century of protection to build the industries it now urged everyone else to expose. Alexander Hamilton’s 1791 Report on Manufactures made the American version. The East-Asian developmental record vindicated the argument’s shape in fragments — Korea, Taiwan, and Japan really did build world-class industries behind temporary walls. The systematic Smith-vs-List confrontation, with both traditions argued at full strength and the conditions under which List was right spelled out, is the work of the sibling walkthrough Smith vs. List on national economy; the door is named here to mark the depth of the dispute, not to settle it.
“The exceptions to the case for free trade are real but limited. The optimal tariff invites retaliation; infant-industry protection rarely meets its conditions; the strategic-trade argument is fragile to the structure of competition. The base rate of successful protection is low, and breadth is never justified by any of them.”
— Douglas Irwin, Free Trade Under Fire, 5th ed., 2020
Irwin, the leading historian of trade policy, grants every exception its theoretical due and then makes the move that matters: he counts. The optimal tariff dissolves under retaliation. Infant-industry protection works in a handful of cases and fails in most, and even the celebrated cases owe more to high savings, human capital, and export discipline than to the tariff itself. Strategic-trade policy reverses its own prescription if the competition turns out to be Bertrand rather than Cournot — you can subsidize when you should have taxed. The theory is sound; the base rate of success is low; and not one of the four exceptions, at its strongest, justifies a tariff on everything.
Where this leaves us
Four doors, four real exceptions — and four conditions the blanket-tariff program satisfies exactly none of. The optimal tariff needs market power and an opponent who won’t retaliate. The infant-industry argument needs genuine externalities, a credible sunset, and a state that will let a failing firm fail. The national-security argument needs a target, not a wall. The bargaining argument needs a track record better than 2018’s. Each exception is a scalpel, precise about where it cuts. The revival is reaching for a hammer.
So the exceptions don’t save the blanket-tariff program. But that leaves the hardest question of all unanswered: if the populists are this wrong about the cure, why are they this right about the disease?
The verdict: when, if ever
“The tariffs will make Americans poorer. That is not a controversial claim among economists; it is arithmetic. The harder truth is that the people cheering for them were failed by a policy regime that promised them adjustment and delivered abandonment.”
— the calibrated mainstream verdict on the 2025 tariff program, after Irwin and the China-shock literature
Two things are true at once, and the honest answer holds both. The economists are right that the tariffs will make America poorer. The populists are right that something broke, and no one fixed it. A verdict that drops either half is propaganda for one side.
No new model here — the apparatus is the taxonomy from Stage 3, run once against the blanket-tariff program. Does it clear any of the four doors? The optimal-tariff door requires market power the United States lacks on most goods and an opponent who won’t retaliate, which every trading partner promptly did. The infant-industry door requires genuine learning externalities, a sunset, and export discipline; an across-the-board tariff names no industry and imposes no discipline. The national-security door licenses targeting, and a tax on everything is the opposite of targeting. The bargaining door requires a record better than the unfulfilled Phase One deal. The program clears nothing. It is, in the precise sense of Stage 2, a small-country-style self-inflicted tax — a deadweight loss the country imposes on itself and calls a victory.
And yet the grievance underneath it is not a fiction. The China-shock losers were concentrated, persistent, and never compensated; the “winners compensate losers” assumption was a model convenience the political system never honored; the pure-efficiency framing systematically under-weighted the distributional and political cost. The right response to that grievance is domestic adjustment, compensation, and place-based policy — not a wall. The development-policy framing of that response sits in Economics Ch.20 §20.8, and the baseline the verdict keeps is the incidence model of Economics Ch.2 §2.6. The backlash itself — how the unaddressed distributional damage became a political movement — is the territory of the sibling walkthroughs on inequality and the immigration debate.
“A 10% tariff across the board, higher on countries that treat us unfairly — that’s how we bring the jobs and the factories back home.”
— the blanket-tariff case, as made for the 2025 program
Are across-the-board tariffs the right tool for a real problem?
The grievance is real and the consensus failed the losers — grant all of it. The question is whether a tax on the entire economy is the instrument that fixes it, or the one that adds a second injury on top of the first.
A real grievance, and the right remedy for it
“The losers from trade were real, they were concentrated, and the compensation the models assumed would protect them was never delivered at anything close to adequate scale. The political backlash was not irrational. It was a bill coming due.”
— the China-shock literature (Autor, Dorn & Hanson) on the distributional debt
This is the grievance at full strength, and it has to be granted before any verdict on the cure is credible. The distributional damage was not an asterisk; it was the central fact the efficiency frame buried. Communities lost their economic base and got a promise of retraining that arrived underfunded and mistargeted when it arrived at all. Oren Cass and the national-conservative project read that betrayal correctly — a profession that promised compensation and delivered abandonment forfeited the right to be surprised when the abandoned reached for the bluntest instrument on offer. The case-up treatment of deliberate deglobalization as a political choice runs through the sibling walkthrough the Brexit case.
“The diagnosis is right and the prescription is wrong. Tariffs are a tax on your own consumers that will not rebuild the industries that closed. The remedy for a labor-market shock is labor-market policy — adjustment, insurance, place-based investment — not a trade war.”
— the mainstream remedy critique (Irwin, Krugman) on the 2025 program
The remedy critique is the verdict’s spine. Granting the grievance does not grant the policy — a real disease still requires the right medicine, and a tariff treats a labor-market injury with a consumption tax. Tools should match problems: a concentrated regional collapse calls for wage insurance, relocation support, and place-based investment, all of which target the harmed directly; a blanket tariff targets no one and harms everyone a little to subsidize a few producers. The adjacent live debate over the spending-side version of the strategic-protection argument — the CHIPS Act and the Inflation Reduction Act as industrial policy rather than as tariffs — is adjudicated in the sibling walkthrough Is industrial policy back, and is it justified?
Where this leaves us
Are tariffs ever good? Yes — rarely, and only when a real exception meets conditions that the blanket-tariff program never satisfies. A large country with market power and no fear of retaliation. An infant industry with genuine externalities and a state that will let it fail. A strategic good too dangerous to depend on a rival for. A bargaining chip wielded by someone with a better record than 2018. Outside those narrow doors, a tariff is a tax you levy on yourself and call a victory. But the grievance the tariff revival rides is not a fiction. The China-shock losers were real, the compensation never came, and the profession that promised it owns part of the bill. The honest answer is the conditional — and the conditional includes a debt the free-trade consensus still hasn’t paid.
Where this leaves us
We started with a favorite word and a blanket tariff, and with a serious movement insisting the free-trade consensus had failed. Stage 2 settled the most-measurable claim: on the goods America actually taxed, the importer’s own consumers paid, and “China pays” was false — with one honest exception the textbook itself grants, the optimal tariff for a large country with market power. Stage 3 took the pro-tariff case at its absolute strongest, four arguments deep, and found the same shape in each: a real exception governed by a condition rarely met. The verdict is the conditional, and it fits on a single line.
The four legitimate exceptions, and the line that runs through all of them:
- Optimal tariff / terms-of-trade. Real for a large country — dissolves under retaliation in a Nash trade war.
- Infant industry. Real under genuine externalities, a credible sunset, and export discipline — conditions rarely all met.
- National security / strategic goods. The strongest post-2020 exception — but it licenses a small yard with a high fence, not a wall.
- Retaliation / bargaining leverage. Coherent in theory — poor in the 2018–2020 record.
- The line: every exception is a scalpel that justifies targeted protection of a named good; none justifies breadth. A blanket tariff invokes the strongest argument to defend the broadest policy, and gets it backwards.
The honest answer lives in the conditional, and it carries a debt. Tariffs are almost always a tax the country levies on itself; the legitimate exceptions are real but narrow, and the blanket revival clears none of them. Yet the grievance driving the revival is genuine — the China-shock losers were never compensated, and the free-trade consensus owns part of the backlash for the remedy it assumed and never built. The next time someone tells you “tariffs bring the jobs back” or “free trade is always optimal,” you have the four doors and the one line to push past both slogans to the conditions each ignores.