Two Greek figures inspect a cargo ship, port crates and a customs scroll.

The Forced-Labour Tariffs

After the Supreme Court struck down the emergency-powers tariffs in February 2026 and a temporary surcharge ran out in July, a new legal basis was needed. The trade representative found it in sixty investigations into whether trading partners keep out goods made with forced labour. The duty took effect on 24 July 2026: Republicans in Congress welcomed it, Democrats called it a pretext, and two small importers went to court.

2.5 points is the whole difference in duty between a country with a forced-labour ban and one without ^[1]
99.4 percent of American imports come from the sixty economies the duty covers ^[1]
0.8 billion dollars of goods a year are held at the border on suspicion of forced labour ^[5]

The proposal

An additional duty of 10 or 12.5 percent on imports from sixty economies, set by whether each bans, or has committed to ban, imports made with forced labour.

The debate

The administration and the Republican chair of the Ways and Means Committee present the duty as pressure on partners to ban goods made with forced labour. Democrats agree that forced labour is a real problem but see the justification as a way to keep the global tariffs after the court ruling. The importers who sued argue that the trade law was never meant for a duty on nearly all imports. Congress has not voted on it.

“The Trump Administration’s investigations have already compelled a number of countries to put forward proposals and take concrete steps to implement forced labor import bans, and I am glad to see this proposal recognizes those countries’ efforts in their proposed tariff rates.”

argues for: Jason Smith, Republican, Missouri Press release, 23 July 2026
“President Trump recognizes that decades of moral suasion have not eradicated forced labor from global supply chains.”

argues for: Jamieson Greer, United States Trade Representative Press release, 23 July 2026
“Today's forced labor justification is too convenient to be taken seriously. Forced labor is a real and pervasive problem in our supply chains and demands serious enforcement.”

argues against: Richard Neal, Democrat, Massachusetts Press release, 23 July 2026

The analysis

The Forced-Labour Tariffs

5% of the weight is for, 95% against.

An argument's weight combines three questions: how big the effect is, how certain it is and how much the good at stake matters. The larger the slice, the more the argument weighs.

For

Against

The forced-labour tariffs are a poor measure on the current analysis, although the forced labour of 27.6 million people worldwide is a real problem ^[4].

Very likely worse

Very likely worse: Under the assumptions used, the proposal brings more than it costs in fewer than 1 of 100 model runs.

Overall effect: very large

How strongly the proposal works overall, in five steps from very small to very large, comparable across all debates. The ring is always full, even when the overall effect is small; this step says how large it is.

The analysis comes down against the duty, as Democrats do, but for a different reason. Uncertainty over the legal basis, the centre of their criticism, is among the smallest costs; what weighs is that a near-universal tariff works like a general tax on imports. American buyers pay it, poorer households from a larger share of their budget, and trade that stops and retaliation by partners add losses that nobody receives. The forced-labour purpose, which supporters stress, enters only through the discount for countries with a ban and is the strongest argument for the measure. No measurement supports it, and border data suggest that detection, more than law, limits what reaches American shelves.

In detail

All arguments

Arguments for3 arguments

Forced labour that leaves the supply chain

1points

The lower tier is open to economies that ban imports made with forced labour, have committed to such a ban in a reciprocal trade agreement, or run a partial regime that keeps out certain forced-labour goods. That discount is the one part of this measure aimed at the problem it is named after. Seventeen qualify; whether 2.5 points moves the other thirty-eight, and whether a commitment on paper frees anyone, is the question.

Value 9 · Basic rightsImpact 0.3Plausibility 3.5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is people who are not held in forced labour: not paid, not free to leave, often with documents taken and a debt that never shrinks. This site places that in the class it uses for the constitutional core — below life and health, above the working order of institutions and above money. It is priced as a deprivation of liberty rather than as a loss of wages, because what is taken is the ability to leave rather than an amount of income. That the people concerned are not Americans does not lower the weight; this site prices what a measure does, not who it happens to. The value is high because the stream is liberty, set below life and health because it is recoverable.

About 27.6 million people are in forced labour worldwide, 23.7 million of them in forced labour imposed by private actors rather than by a state, and that private forced labour generates roughly 236 billion dollars of illegal profit a year ^[4]. Only part of that produces goods that cross borders — most forced labour is in domestic service, construction and agriculture for local markets — and roughly 3.5 million people is used here for the traded share, in a range from 2 to 6 million. The United States takes about 23 percent of the world's at-risk imports by value although it is only 13 percent of global imports, which puts roughly 800,000 of those people in supply chains that end in America ^[5]. What the measure does to that number runs through one narrow channel: a 2.5 percentage point discount for economies that ban such imports, commit to a ban in a reciprocal trade agreement, or keep out certain forced-labour goods under a partial regime ^[1]. Because a commitment or a partial regime is enough, the discount can be earned before any comprehensive ban is enforced, so the share of people it frees is smaller than the share of governments it moves. Five percent of the 800,000 is used here, in a range from half a percent to twenty percent — 40,000 people a year. A year in forced labour is valued at the rate this site uses in every debate for a year a person spends without command of their own time: all 5,840 waking hours of it at the value of an hour under compulsion, which makes it weigh as much as a year of life. The American federal rate for wrongful imprisonment, 50,000 dollars a year, lies just above it and the German one, about 32,000 dollars, below. On the scale of this evaluation that gives a Normalised Impact of 0.32, in a range from 0.03 to 1.42. Violence and injury suffered in forced labour would be a separate stream and are not counted here, for want of a number. The Impact is small against the tariff that carries it, because the discount that does the work is a fifth of the duty and applies to a country's laws rather than to any particular shipment.

▸ Show calculation ▾ Hide calculation

	People in forced labour producing goods that cross borders Setting, range 2 to 6 million: most forced labour is in domestic service, construction and agriculture for local markets [4]	of 23.7 million in privately imposed forced labour, out of 27.6 million in forced labour worldwide	3.5 million people
×	Share in supply chains that end in America the United States takes 23 percent of the world's at-risk imports by value although it is 13 percent of global imports [5]	23 %	800,000 people
×	Share the discount frees Setting, range 0.5 to 20 percent: the discount is 2.5 percentage points against a duty of 10 or 12.5, it applies to a country's laws rather than to any shipment, and a trade-agreement commitment or a partial regime earns it before a full ban is enforced [1]	5 %	40,000 people a year
×	Value of a year of freedom regained This site's rate for a year of deprivation of liberty: 5,840 waking hours at 6.85 euro-equivalent, the value of an hour under compulsion — equal to one year of life. Range 27,375 to 43,100 euro-equivalent (about 32,000 to 50,000 dollars), the German and the American statutory compensation for wrongful imprisonment	40,000 euro-equivalent each	1,600 million euro-equivalent
÷	Normalised Impact scale of this evaluation	5 billion euro-equivalent a point	0.32

Score 0.32 Impact × 9 Value × 3.5 Plausibility ÷ 10 = 1 of 100

Nothing here has been measured, and the design gives reason for doubt in both directions. The counterfactual is the sectoral duties alone, without the sixty-economy action, and no comparable action has ever been taken, so there is no precedent to read. The chain has three links: a discount induces a government to ban such imports or only to commit to a ban or a partial regime, which already earns the lower rate ^[1]; the ban is then actually enforced, which the tariff schedule does not require; and enforcement removes forced labour from production rather than moving it elsewhere. The third link is where the evidence that does exist points against: American customs blocks about 0.8 billion dollars of goods a year on forced-labour suspicion out of imports of 3,300 billion, which suggests that detection is the binding constraint rather than the legal authority ^[5]. The confounder that matters is that the seventeen economies at the lower tier qualified through bans, trade-agreement commitments or partial regimes that were in place or negotiated before the final action and would largely have stood anyway, so part of what looks like an effect is a pre-existing state of affairs. Reverse causation does not arise. The Plausibility is low because the chain that carries this argument has three links and none of them has been observed.

Evidence: Mechanism · plausibility at most 6 Evidence of effect: Mechanistic · at most 6 Placement: Chain open · plausibility 3–3.5

Open: Whether any of the thirty-eight economies at the higher tier adopts a ban or a commitment in the next two years, and whether the commitments already made are followed by enforced bans, is directly observable and would settle the first two links of the chain.

Factories that stay open

0.36points

A tariff raises the price of the imported version and makes the domestic one competitive, which is the oldest argument for one. The 2018 and 2019 American tariffs were studied closely enough to test it. Manufacturing employment fell rather than rose.

Value 6 · OutputImpact 0.3Plausibility 2

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is not the jobs themselves — a worker who moves from one industry to another has not made the country poorer — but the cost of the move: months without work, skills that do not transfer, a house in a town where the plant was the employer. This site prices that at the level it uses for economic output and systems. Where the displacement is concentrated in particular places, as manufacturing displacement is, the cost is larger than an average would suggest, because a whole local labour market absorbs it at once. That concentration is a fact about the size of the effect rather than its weight, and it is in the Impact. The value sits in the middle-upper part of the scale, at the level this site uses for economic output.

A duty of nine percentage points on 3,300 billion dollars of imports shifts some demand to domestic producers. If that shift is two percent of import value, it supports something like 250,000 manufacturing positions that would otherwise not exist. Those are not net new jobs — the workers come from somewhere — so what is counted is not the wages but the displacement avoided: workers who would otherwise have gone through the loss of a job in a town with one large employer. Roughly 100,000 people a year are assumed to avoid that, in a range of 33,000 to 300,000 if the shift happens at all (whether it does is a question of Plausibility), and the cost of such a displacement is put at about 17,000 dollars each, drawn from what the research on American import competition finds about earnings that do not recover. The result is about 1.7 billion dollars a year. Set against this, and not counted here, is the same effect in reverse for firms that use imported inputs, which the next paragraph explains is what the evidence actually found. The Impact is small even at face value, because what a tariff can protect is a displacement cost rather than a wage bill.

▸ Show calculation ▾ Hide calculation

	American goods imports a year [3]		3,300 billion dollars
×	Share of demand shifting to domestic producers Setting, range 0.8 to 5 percent if demand shifts at all: a nine point duty against import demand that responds roughly one for one	2 %	66 billion dollars
=	Workers who avoid a displacement Setting, range 33,000 to 300,000 if demand shifts at all: the positions are not new jobs, so only the displacement avoided is counted	about 100,000 a year of the 250,000 positions involved	100,000 people a year
×	Cost of a displacement avoided Setting, range 5,000 to 40,000 euro-equivalent (about 5,800 to 46,000 dollars): drawn from what the research on American import competition finds about earnings that do not recover [6]	15,000 euro-equivalent each (about 17,000 dollars)	1,500 million euro-equivalent
÷	Normalised Impact scale of this evaluation	5 billion euro-equivalent a point	0.3

Score 0.3 Impact × 6 Value × 2 Plausibility ÷ 10 = 0.4 of 100

This is the one claim in the debate that has been tested on American data, and it came out the other way. The 2018 and 2019 tariffs were imposed at different rates on different products at different times, which allows exposed manufacturing industries to be compared against unexposed ones, and the research using that variation found manufacturing employment slightly lower rather than higher in the protected industries ^[6]. The reason it gives is arithmetic: the same tariffs raised the price of imported inputs and drew retaliation against exports, and those two effects were larger than the protection. The confounder that would matter — that exposed industries were declining anyway — is what the product-level comparison absorbs, and reverse causation does not arise, since the tariff schedule was set in Washington rather than by any industry's employment. The reply available to this argument is that the 2018 tariffs were narrow while this one is near-universal, so input costs rise for everyone and the relative protection is what counts. That is a real difference and it is why the figure is not scored at zero. The Plausibility is low because the effect claimed here was looked for in American data under a proper comparison and the opposite was found.

Evidence: Study · plausibility at most 8 Evidence of effect: Quasi-experimental · at most 8 Placement: Effect did not materialise · plausibility 1.5–2

Open: The new duty applies at two rates across sixty economies, which gives usable variation. Comparing employment in industries sourcing mainly from the 10 percent tier against those sourcing from the 12.5 percent tier would settle this within two years.

Less depends on one supplier

0.21points

A duty that falls on everyone still falls hardest on whoever supplies most, and it makes a second source worth paying for. What that is worth is the cost of the disruption it prevents, multiplied by how likely the disruption is.

Value 6 · OutputImpact 0.1Plausibility 3.5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is output that continues when a supply route closes: factories that keep running, shelves that stay stocked, hospitals that have the equipment they ordered. It belongs to the class this site uses for economic systems and prosperity. It is not counted at the level of health or security even where the goods are medical, because what is priced is the continuity of supply rather than any particular patient's outcome, and pricing both would count the same disruption twice. What the second source costs to maintain is inside the tariff's own cost and is not counted again here. The value sits in the middle-upper part of the scale, at the level this site uses for economic output.

This is an argument about a rare event, so what is counted is the expected value rather than the event. A serious disruption to a dominant supply route — a blockade, an export ban, a conflict — is put at a two percent chance in any year, in a range from half a percent to five, and the cost if it happened at 200 billion dollars of lost and delayed output. That is an expected four billion dollars a year of exposure. A near-universal duty makes a second source worth paying for and is assumed to cut that exposure by fifteen percent, in a range from five to thirty: 600 million dollars a year. The probability sits in the quantity rather than in the Plausibility, which is the treatment this site uses for rare events with large consequences. The Impact is the smallest on the pro side and would be considerably larger for anyone who thinks a disruption more likely than one year in fifty.

▸ Show calculation ▾ Hide calculation

	Cost of a serious disruption to a dominant supply route Setting, range 80 to 400 billion dollars	lost and delayed output	200 billion dollars
×	Chance of it in any year Setting, range 0.5 to 5 percent: the probability belongs in the quantity, not in the Plausibility	2 %	4 billion dollars a year
×	Share of the exposure the duty removes Setting, range 5 to 30 percent: a broad duty makes a second source worth paying for	15 %	0.6 billion dollars a year
÷	In euro-equivalent	1.16 dollars to the euro	0.52 billion euro-equivalent
÷	Normalised Impact scale of this evaluation	5 billion euro-equivalent a point	0.1

Score 0.1 Impact × 6 Value × 3.5 Plausibility ÷ 10 = 0.2 of 100

What is scored here is not whether a disruption happens but whether the reasoning that produces the figure is sound, and there are two weak points. The counterfactual is the same import pattern under the sectoral duties alone, and no measurement exists of how much a broad tariff actually diversifies sourcing rather than simply raising the cost of the same sourcing. The 2018 and 2019 experience is the closest evidence and it is mixed: sourcing did shift away from the most affected origin, but a great deal of it shifted to third countries that assembled the same components, which is a change of address rather than of dependence ^[6]. That counter-mechanism is named and unresolved, and it is what holds this argument down. The confounder is that firms were already diversifying for reasons of their own after 2020, so part of any shift is not attributable to the tariff. Reverse causation does not arise. The Plausibility is low because the link between a broad tariff and genuine diversification has not been shown and the evidence that exists suggests relabelling rather than reshoring.

Evidence: Mechanism · plausibility at most 6 Evidence of effect: Mechanistic · at most 6 Placement: Chain open · plausibility 3–3.5

Open: Component-level trade data can distinguish a genuine change of supplier from a change of final assembly point. Two years of it under the new duty would settle whether dependence actually falls.

Arguments against4 arguments

A tax that takes more from those with less

14points

A tariff is collected from importers and paid by whoever buys the goods. Goods are a larger share of a small household budget than of a large one, so the same duty takes a bigger bite lower down. About 270 billion dollars a year changes hands.

Value 5 · Household budgetsImpact 4.7Plausibility 6

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is money, priced at the middle of the scale. The duty moves it from American buyers to the American Treasury, and because both ends are domestic the transfer is booked once, as the difference between what a dollar is worth at each end, rather than twice as a cost to one and a benefit to the other. The whole sum does not appear anywhere in this evaluation, which is deliberate: counting 270 billion dollars as a cost and again as revenue would say something false about what the measure does. That the money is collected at the border rather than on a payslip changes nothing about its weight. The value is the middle of the scale, and the regressive shape of the tax is carried entirely by the Impact.

American goods imports run at about 3,300 billion dollars a year ^[3]. Fifty-five of the sixty economies pay a flat 10 or 12.5 percent; the European Union, Taiwan, Japan, South Korea and Switzerland pay one of the two rates net of their existing most-favoured-nation duties, and goods compliant with the North American agreement, goods already under the sectoral duties and a list of exempt products are excluded ^[1]^[2]. Once those exemptions, the net-of-duty treatment and the lower tier's larger partners are weighted in, the new duty adds a trade-weighted average of roughly nine percentage points over the position that would otherwise apply, in a range from six to twelve. Import volumes fall as prices rise, so collections come to about 270 billion dollars. The money moves from households to the federal budget, and the two ends differ in what a dollar is worth there. Goods take about twice as large a share of spending in the bottom fifth of American households as in the top, which puts the buyers' weight at roughly 1.1 against the budget's 1.0. Only that difference of 0.1 is a loss: a Normalised Impact of about 4.7 on the scale of this evaluation, in a range from 2.4 to 8. The Impact is the largest in this debate even after netting, because a tenth of a very large number is still a large number.

▸ Show calculation ▾ Hide calculation

	American goods imports a year [3]		3,300 billion dollars
×	Additional duty over the position that would otherwise apply Setting, range 6 to 12 points: trade-weighted across the 10 and 12.5 percent tiers and the five net-of-duty cases, net of the sectoral duties already in force and of the exemptions for North American-agreement goods, sectoral-duty goods and listed products [1]	9 percentage points	297 billion dollars
×	Collections after import volumes fall import demand responds roughly one for one to a price rise of this size	91 %	270 billion dollars
÷	In euro-equivalent	1.16 dollars to the euro	233 billion euro-equivalent
×	Difference in what a dollar is worth goods take about twice as large a share of spending in the bottom fifth of American households as in the top; only the difference is a loss, not the whole transfer	1.1 minus 1.0	23.3 billion euro-equivalent
÷	Normalised Impact scale of this evaluation	5 billion euro-equivalent a point	4.7

Score 4.7 Impact × 5 Value × 6 Plausibility ÷ 10 = 14 of 100

The mechanism is a matter of statute and arithmetic: a duty is collected at a stated rate on a measured volume. The counterfactual is the tariff position without this action, which is the sectoral duties alone, and both are published schedules. What is estimated is how much of the duty reaches consumers rather than being absorbed by foreign exporters in lower prices, and here the American evidence is unusually clear: studies of the 2018 and 2019 tariffs using product-level variation found the duty passing through almost entirely to American prices, with foreign exporters absorbing close to none of it ^[7]. That is why no absorption is assumed here. The confounder that would matter — that exchange rates move to offset the duty — is real and unresolved, and it would reduce the figure. The regressivity weight is a judgment rather than a measurement, and the range around it is wide for that reason. Reverse causation does not arise. The Plausibility is at the top of what a projection can carry: the duty and the volume are measured and only the weighting is judged.

Evidence: Study · plausibility at most 8 Evidence of effect: Quasi-experimental · at most 8

Trade that simply stops happening

13points

Part of a tariff is a transfer and part of it is neither paid nor received: exchanges that were worth making at the old price are not worth making at the new one. Nobody gets that money because it never exists. It is the part economists agree on.

Value 6 · OutputImpact 3.8Plausibility 5.5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is output that does not happen: goods not made, not shipped, not bought, and the domestic production that replaces some of them at a higher cost than the import it displaces. It belongs to the class this site uses for economic systems and prosperity. It is a genuine loss rather than a transfer, which is what separates it from the argument above — no household gains what the Treasury loses here, because there is nothing to gain. Nothing about who bears it enters, because it falls on producers and consumers on both sides of the border in proportions nobody can trace. The value sits in the middle-upper part of the scale, at the level this site uses for economic output.

Two losses arise, and both are measured off the same fall in imports. A nine point duty on 3,300 billion dollars of imports, with import demand responding roughly one for one to price as the transfer argument assumes, removes about 297 billion dollars of trade a year. On that trade, buyers lose the difference between what the goods were worth to them and what they cost, and domestic producers who step in do so at a higher cost than the import they replace; both sit inside the same triangle — half the duty times the trade that stops — which comes to about 13 billion dollars a year. The second loss is that the trade which stops was already paying duty: the ordinary tariffs of about three percent that remain in place under the sectoral position alone, in a range from two to five, are forgone on every dollar of trade lost, about 9 billion dollars. Together about 22 billion dollars a year, in a range from 12 to 41 billion, the width reflecting how strongly trade responds to price. The estimate rises more than proportionally with the duty, so a twelve point tariff would cost well over half as much again as a nine point one. What is not counted here is the same effect in reverse — that some of the displaced trade goes to third countries rather than disappearing — which would reduce the figure and is one reason the range runs low. The Impact is second only to the transfer and, unlike the transfer, none of it comes back to anybody.

▸ Show calculation ▾ Hide calculation

	American goods imports a year [3]		3,300 billion dollars
×	Imports that stop import demand responding roughly one for one to a nine point price rise, as in the transfer argument	9 %	297 billion dollars a year
×	Value of the exchanges that stop happening the standard measure of what a tax on an exchange destroys rather than moves; the domestic production that replaces part of the imports at a higher cost is inside this figure, not added to it	half of 9 percentage points	13.4 billion dollars a year
+	Duty already paid on the trade that stops Setting, range 2 to 5 percent: the ordinary duties that remain under the sectoral position, forgone on every dollar of trade lost	3 % × 297 billion	22.3 billion dollars a year
÷	In euro-equivalent Setting, range 10 to 35 billion euro-equivalent (about 12 to 41 billion dollars), set by how strongly trade responds to price	1.16 dollars to the euro	19.2 billion euro-equivalent
÷	Normalised Impact scale of this evaluation	5 billion euro-equivalent a point	3.84

Score 3.84 Impact × 6 Value × 5.5 Plausibility ÷ 10 = 13 of 100

This is the least contested proposition in the debate and also one of the least directly measured. That a tax on an exchange reduces the number of exchanges is not in question, and the counterfactual — the same trade without the duty — is what every estimate of the 2018 and 2019 tariffs compared against. Those estimates, using product-level variation in exposure, found import volumes falling sharply and roughly in line with what the standard calculation predicts, which is the support this argument rests on ^[7]. What is not measured is the elasticity at a near-universal duty of this size: every American estimate comes from narrower tariffs, and a broad one behaves differently because substitution toward untaxed origins is not available. That works in both directions — less substitution means less trade diverted and more genuinely lost — and it is why the range is wide. The confounder that would matter is that other things changed in 2018 and 2019, including retaliation, which the product-level comparison partly absorbs. Reverse causation does not arise. The Plausibility is above the middle: the mechanism is settled and the size at this scale of duty has never been observed.

Evidence: Study · plausibility at most 8 Evidence of effect: Quasi-experimental · at most 8

What sixty economies do back

4.3points

A duty on sixty trading partners invites sixty responses. In 2018 and 2019 the retaliation landed mostly on American agriculture and cost billions in lost exports. This action is several times larger and reaches nearly every partner at once.

Value 6 · OutputImpact 1.4Plausibility 5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is American output that finds no buyer: harvests sold below cost or not at all, production lines built for export markets that close, and the towns where those are the employer. It belongs to the class this site uses for economic systems and prosperity, the same as the trade lost in the argument above and for the same reason. What the exporters lose is not counted as a transfer to foreign producers, because the point is the output that is not produced rather than who produces it instead. Federal compensation payments to affected farmers, which followed the last round, are a transfer within the American budget and are not counted here at all. The value sits in the middle-upper part of the scale, at the level this site uses for economic output.

American goods exports run at about 2,100 billion dollars a year. In 2018 partners answered an American action on 12.7 percent of imports with retaliatory duties on 127 billion dollars, 8.2 percent of American exports, raised by about 13 points ^[7]. An action on nearly all imports is taken here to draw retaliation on a fifth of exports, a setting in a range from 5 to 45 percent, with duties of about ten points: 420 billion dollars of exports facing a new barrier. Exporters do not lose all of that: some sell elsewhere, some absorb the duty, some keep the sale. What is lost is the margin on the trade that stops: the affected exports are taken to fall by a tenth, foreign demand for American goods responding about one for one to the duty, as measured for the 2018 retaliation ^[7], and a fifth of that lost value is margin rather than cost saved, giving about 8.4 billion dollars a year, in a range from 2.3 to 23 billion. The width is unavoidable because retaliation is a political choice rather than a mechanical response, and because a near-universal American action gives sixty governments the same reason to act at once, which has no precedent. The Impact is a third of the trade destroyed on the import side, which is the ordinary proportion: a country's exports are smaller than its imports and retaliation is usually narrower than what provoked it.

▸ Show calculation ▾ Hide calculation

	American goods exports a year [3]		2,100 billion dollars
×	Share facing retaliatory duties Setting, range 5 to 45 percent: in 2018 retaliation covered 8.2 percent of American exports in answer to an American action on 12.7 percent of imports; this action reaches nearly all imports [7]	20 %	420 billion dollars
×	Affected exports that stop Setting, range 3 to 20 percent: foreign demand for American goods responded about one for one to the 2018 retaliatory duties, and targeted exports fell by about 10 percent [7]	10 %	42 billion dollars a year
×	Margin lost on the trade that stops Setting, range 5 to 50 percent: some exporters sell elsewhere, some absorb the duty, some keep the sale	20 %	8.4 billion dollars a year
÷	In euro-equivalent	1.16 dollars to the euro	7.2 billion euro-equivalent
÷	Normalised Impact scale of this evaluation	5 billion euro-equivalent a point	1.4

Score 1.4 Impact × 6 Value × 5 Plausibility ÷ 10 = 4.3 of 100

That tariffs draw retaliation is a matter of record rather than of theory — it happened in 2018, in 2019, and in every comparable episode — so the direction is not the question. The counterfactual is American export performance without this action. The closest measurements use product- and destination-level variation in the 2018 retaliation: exports of targeted products fell by about 10 percent, with foreign demand responding about one for one ^[7], and a monthly product-level estimate puts the loss of agricultural exports at 13.5 to 18.7 billion dollars a year ^[8]. Manufacturing industries more exposed to retaliation also lost employment relative to others ^[6]. What these cannot supply is the size here. This action is several times larger, reaches nearly all partners at once, and is framed on a ground — forced labour — that some partners may find harder to retaliate against openly than a straightforward trade dispute. The confounder that matters is that the same partners are negotiating other things with the United States simultaneously, so retaliation is bundled with concessions in ways no model captures. That is unresolved and it is the main reason this figure carries a range of an order of magnitude. Reverse causation does not arise. The Plausibility is at the middle: retaliation is certain in direction and its size here has no precedent to be read from.

Evidence: Precedent · plausibility at most 8 Evidence of effect: Quasi-experimental · at most 8

Three legal bases in five months

0.53points

The emergency-powers tariffs were struck down in February, replaced within hours by a balance-of-payments surcharge, which expired in July and was replaced by this. About 100 billion dollars has been certified for refund. A firm planning a five-year supply chain has nothing to plan against.

Value 6 · OutputImpact 0.2Plausibility 4

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is investment that is postponed or not made: a warehouse not built, a supplier contract not signed, a production line kept in its current location because moving it might be wrong within a year. It belongs to the class this site uses for economic systems and prosperity. What is priced is the output that investment would have produced, not the money itself, which is not destroyed but merely held. Nothing here is a judgment about whether the tariff should exist; a stable tariff and no tariff are both plannable, and it is the changing that costs. The value sits in the middle-upper part of the scale, at the level this site uses for economic output.

American business fixed investment runs at roughly 3,500 billion dollars a year, of which perhaps thirty percent is in activities exposed to trade policy: manufacturing, logistics, wholesale and the supply chains behind them. Research on policy uncertainty finds investment in exposed sectors falling by one to two percent when the rules are unsettled, and one and a half percent is used here, giving about 16 billion dollars of investment deferred each year. What that costs is not the investment but its return, put at eight percent, or 1.26 billion dollars a year, in a range from 0.35 to 3.5 billion. Deferred is the right word: most of it is made eventually, which is why the figure is the return on a delay rather than the value of the investment. The 100 billion dollars of tariff refunds working through the system is a transfer rather than a loss and is not counted here ^[3]. The Impact is the smallest in this debate, which is the honest size of a cost that shows up as delay rather than as damage.

▸ Show calculation ▾ Hide calculation

	American business fixed investment a year		3,500 billion dollars
×	Share in trade-exposed activities manufacturing, logistics, wholesale and the supply chains behind them	30 %	1,050 billion dollars
×	Investment deferred while the rules are unsettled Setting, range 0.5 to 3 percent: research on policy uncertainty finds one to two percent in exposed sectors	1.5 %	16 billion dollars
×	Return on a year's delay most of the investment is made eventually, so what is lost is the return on the delay rather than the investment	8 %	1.26 billion dollars a year
÷	In euro-equivalent	1.16 dollars to the euro	1.1 billion euro-equivalent
÷	Normalised Impact scale of this evaluation	5 billion euro-equivalent a point	0.22

Score 0.22 Impact × 6 Value × 4 Plausibility ÷ 10 = 0.5 of 100

That firms defer irreversible commitments when the rules might change is one of the better-established findings in investment economics, and the counterfactual is the same firms under a settled schedule. What has not been measured is this episode. The uncertainty here is unusual in kind: not that the tariff might rise or fall but that its legal basis might be struck down entirely, which happened in February and could happen again to this one. The chain is short — unsettled rules, deferred commitment, output later than it would have been — and every link is visible. The counter-mechanism is real and only partly answered: a firm facing an uncertain tariff may invest sooner rather than later, to secure domestic capacity before the rules tighten, and the net direction across firms has not been established for this case. Reverse causation does not arise. The Plausibility is below the middle: the mechanism is well established in general, unmeasured here, and the direction is not certain across all firms.

Evidence: Mechanism · plausibility at most 6 Evidence of effect: Mechanistic · at most 6 Placement: Chain closed, unevidenced · plausibility 4–5

Open: Capital expenditure guidance from trade-exposed firms, compared across the February and July changes of legal basis, would measure the deferral directly and could carry P to 6.

How robust the result is

Under the assumptions used, the proposal brings more than it costs in fewer than 1 of 100 model runs.

gains more than it costscosts more than it gains

Each dot stands for 200 of the 20,000 model runs.

The 270 billion dollars a year the duty raises from American buyers and hands to the Treasury is a transfer between two domestic parties. It is booked once, as the difference between what a dollar is worth at each end, rather than twice as a cost and a benefit. Anyone who counts the revenue as a separate gain gets a very different total. The comparison throughout is with the sectoral duties alone, the tariff position that would otherwise apply after the emergency-powers tariffs were struck down in February 2026 and the balance-of-payments surcharge expired in July. A year in forced labour is valued at the rate this site uses for a year a person spends without command of their own time. Counted this way, the measure lies ahead in none of the runs.

Show the distribution of the model runs

20,000 simulated futures

Very likely worse · 0 %

In fewer than 1 of 100 simulated futures the proposal brings more than it costs.

No effect is known exactly; every argument has a range. The calculation draws a possible size for each argument from its range 20,000 times and counts how often the benefits exceed the costs.

Net effect in points per future. Left of zero the drawbacks outweigh the benefits, right of zero the benefits win. Eight in ten futures fall between −44 and −18 points.

Scale and assessment

Scale of this evaluation: Normalised Impact, unitless and calibrated to this topic. For comparison, one point here corresponds to about 5 billion euro-equivalent a year.

Size class:large

Effect very large (−150) · 5 % of the effect is positive

Areas show individual argument weights. The positive share of the effect nets out money transfers between the two sides; the probability of a positive net effect is determined separately.

How we score →

Review and sources

AI evaluation · not yet reviewed by a human

This evaluation was produced and sourced by an AI model; a human review is still pending. Figures and conclusions may still change. The review log is at the foot of the page.How review works →

Last reviewed by Claude Opus 5.5 · October 29, 2026 · 8× AI, not yet reviewed by a human

October 29, 2026AI reviewClaude Opus 5.5record updated
Page texts for the new page layout (draft G) with quotes from primary sources; numbers unchanged.
October 21, 2026AI reviewClaude Fable 5.1record updated
Money amounts in argument texts stated in dollars per style-en.md; scores, chains and balance unchanged
October 20, 2026AI reviewClaude Fable 5.1record updated
Summary and scale note rewritten to the style of 20 Oct 2026; numbers unchanged
October 16, 2026AI reviewClaude Opus 5.5record updated
All seven arguments now carry an Impact range, from the setting ranges in the derivation combined in quadrature or from the range stated in the Impact text; pro-2 range covers only the case in which demand shifts.
October 14, 2026AI reviewClaude Opus 5.5record updated
Englische Debatten, Befunde SO-17/SO-18: niedrige Zollstufe auch über Abkommenszusage oder Teilregime erreichbar, Vergeltungsbelege auf Fajgelbaum u. a. (2020) und Grant u. a. (2021) umgestellt; Zahlen unverändert.
September 20, 2026AI reviewClaude Fable 5.1re-scored
Neuer Anker freiheitsentzug-jahr (Julian 20.09.): Jahr Zwangsarbeit 43.000 → 40.000 € (Band 27.375–43.100 €); pro-1 i 1,72 → 1,6, Score 5,42 → 5,04; r 0,050 → 0,048 (deutlich schlechter, unverändert).
September 14, 2026AI reviewClaude Fable 5.1re-scored
Gesamtprüfung 08.09.: Maßnahmentext auf 17+38+5 Net-of-MFN-Fälle und Ausnahmen korrigiert; con-2 Doppelzählung des Produktionsdreiecks durch Zollkeil-Rechteck ersetzt (i 20 → 19,2); ILO 23,7 Mio als privat auferlegt; r 0,05 → 0,05 (deutlich schlechter).
September 6, 2026AI reviewClaude Opus 5First evaluation
Created for the English side: the tariff transfer booked once as a weight difference, with the forced-labour purpose priced through the two-tier discount alone.

Evaluations are produced with AI support and reviewed on a schedule for new developments; human passes are marked separately.How we review →

Show all 8 sourcesHide sources

Global Trade Alert: Forced-Labour Section 301 Tariffs on 60 Economies Take Effect on 24 July: An Overview of the Final Action. globaltradealert.org
Holland & Knight: And the (Tariff) Beat Goes On: New Section 301 Forced-Labor Tariffs Imposed on 60 Countries. hklaw.com
Penn Wharton Budget Model: Effective Tariff Rates and Revenues. budgetmodel.wharton.upenn.edu
International Labour Organization: Profits and Poverty: The Economics of Forced Labour. ilo.org
Walk Free, Global Slavery Index: Importing Risk. walkfree.org
Flaaen and Pierce, Board of Governors of the Federal Reserve System: Disentangling the Effects of the 2018-2019 Tariffs on a Globally Connected U.S. Manufacturing Sector. federalreserve.gov
Fajgelbaum, Goldberg, Kennedy and Khandelwal, Quarterly Journal of Economics: The Return to Protectionism (2020): product-level variation in the 2018 American tariffs and the retaliation against them. nber.org
Grant, Arita, Emlinger, Johansson and Xie, Applied Economic Perspectives and Policy: Agricultural Exports and Retaliatory Trade Actions: An Empirical Assessment of the 2018/2019 Trade Conflict (2021). doi.org

More debates in this area

Unclear · 48 %

Size: medium

School Scholarship Credit

Let every taxpayer send up to 1,700 dollars of federal income tax to scholarship organizations for school children instead of the Treasury, from January 2027.

Very likely better · 100 %

Size: very large

Lifting The Payroll Cap

Charge the Social Security payroll tax on earnings above the taxable maximum of 184,500 dollars, which today stop being taxed altogether.

Very likely worse · 0 %

Size: large

Keep the Tips Deduction

Make the deductions for tipped and overtime income permanent instead of letting them lapse after 2028.

Very likely better · 100 %

Size: medium

California Billionaire Tax

Charge everyone who lived in California on 1 January 2026 with a net worth above 1 billion dollars a one-time tax of 5 percent, and spend nine tenths of it on health care.

All debates in Taxes, Tariffs and the Federal Budget →