The 100% Drug Tariff Where the 100% Hits Almost Nothing

In about three weeks, on July 31, 2026, US Customs and Border Protection will begin enforcing a new tarifftariffA tax a government charges on goods brought in from other countries, usually a percentage of the item's value. on imported medicines that carries one of the largest headline numbers in modern American trade policy: 100 percent. The authority is Section 232Section 232A US trade law that lets the president impose tariffs on imports judged a threat to national security; it is the same authority used for the 2018 steel and aluminum tariffs., the same national-security trade law that produced the 2018 steel and aluminum tariffs. The instrument is Proclamation 11020Proclamation 11020The formal presidential order, carrying the force of law, that created this pharmaceutical tariff; signed April 2, 2026 and published April 9, 2026., signed April 2, 2026 and published a week later. The target, on paper, is patented brand-name drugs and their active ingredients entering the country.

A 100 percent tariff means the tax equals the full value of the good: a $1,000 vial would owe $1,000 in duty. If that rate genuinely applied to the roughly $213.8 billion in pharmaceuticals the US imported in 2025, it would be one of the heaviest sector tariffs the country has ever levied. But read the proclamation's own carve-outs and rate caps line by line, and something strange appears. Almost every dollar of drugs actually crossing the border is routed into a lower lane before the 100 percent ever touches it.

The gap between the number in the headline and the number CBP will collect is the story, but the tariff is not free.

We are not asking whether drug prices will rise (too broad) or whether this is good policy (not our lane). We are asking one narrow, answerable question: when the meter turns on July 31, how much does it actually meter?

The 'trade-weighted effective ratetrade-weighted effective rateThe average tariff actually paid across all imports, weighting each product's rate by how much of it is imported, so a very high rate on a tiny slice barely moves the average.' is just the average tariff actually paid across all imports, weighting each product's rate by how much of it is imported. A punishing rate on a tiny slice of trade barely moves that average. Our claim, stated so it can be proven wrong: the effective rate at full phase-in lands around 8 percent, high single digits, plausibly into the low teens, not mid-teens, and the slice of import value facing the full 100 percent is a low-single-digit rounding error. That is a claim about the rate, not about the money. Eight percent on a branded base near $100 billion is still billions of dollars in duty.

Proclamation 11020 is structured as a tiered schedule, not a flat wall. There are four core lanes, plus one small sub-case. Lane one is 0 percent: generics, biosimilars, US-origin returns, and a broad list of excluded classes (nuclear medicines, plasma-derived products, fertility drugs, cell-and-gene therapies, antibody-drug conjugates, and chemical-biological-radiological-nuclear countermeasures), together with branded drugs made by the companies that signed pricing deals and appear in Annex II. Lane two is a 15 percent cap. Lane three adds 20 percent for the Annex III companies that have committed to build US plants. Lane four is the 100 percent default, which applies to whatever the first three do not catch. The sub-case: the UK sits at 10 percent rather than 15.

The single most important sentence in the document is the one that defines the 15 percent lane. It sets the rate at 15 percent 'for products of' the EU, Switzerland, Liechtenstein, Japan and Korea (10 percent for the UK), unless a lower rate applies. Read carefully, that is an origin-based ceiling, not a reward for a deal. It is available to any importer of a drug made in those places, whether or not the manufacturer signed anything. Because the largest branded suppliers to the US are exactly those places, the origin cap, not the generics carve-out, is what holds the effective rate down.

That distinction matters, because the generics exemption gets most of the attention. Generics and biosimilars are indeed carved out entirely, and they fill roughly nine of every ten US prescriptions, but by import value they are cheap, so removing them takes a smaller bite out of the $213.8 billion base than the headcount suggests. The branded, patented dollars are where the money is, and those are exactly the dollars the 100 percent rate was written to hit. The reason it mostly does not is the origin cap: the branded blockbusters arrive from Ireland, Germany, Switzerland and Belgium, and the moment they do, the ceiling drops from 100 to 15.

To answer the question we did not guess at an average, we rebuilt the import base and dropped each dollar into the lane the proclamation assigns it. Start with the $213.8 billion total for 2025 and the country-of-origin breakdown: Ireland $42.6B, Germany $26.1B, Switzerland $19.4B, India $15.1B, Belgium $12B, and the long tail beyond. India is overwhelmingly generics, so it moves to the 0 percent lane. The four EU/Swiss sources, about $100 billion combined, all sit under the 15 percent origin cap by default.

But 'under the cap' is not the same as 'taxed at the cap.' Two things pull branded value below 15 percent, all the way to zero. First, where the manufacturer is one of the Annex II companies that signed a pricing agreement, its products pay nothing until 2029. Second, several high-value biologic categories, cell-and-gene therapies, antibody-drug conjugates, plasma products, are excluded classes regardless of origin. Netting those out, we estimate the branded value that actually pays 15 percent at about $76 billion, not the full $100 billion of origin-eligible flow. The rest of the branded EU/Swiss value drops into the 0 percent lane alongside the generics.

What is left for the 100 percent default is small: patented products that are simultaneously non-generic, made outside the capped origin countries, made by a non-signer, and not in an excluded class. On our reconstruction that is about $3.8 billion, a low-single-digit share of the base. Do the weighted arithmetic across all five lanes and the effective rate is (0x120 + 10x5 + 15x76 + 20x9 + 100x3.8) / 213.8, or about 8.2 percent.

The single biggest source of uncertainty is how much branded value the signer companies pull into the 0 percent lane, a judgment call, not a published figure. Push that estimate up and the effective rate slips toward the mid-single digits; push it down and it climbs toward the low teens. In no realistic version does it approach the 100 percent headline, and in none does it fall to zero.

It would be easy to misread all of this as 'the tariff collects almost nothing.' It does not. 'Almost nothing' applies to the 100 percent rate specifically, the punishing number in the headline touches only a sliver of value. The tariff as a whole still bills real money. A roughly 15 percent charge on something like $76 billion of branded imports is on the order of $11 billion in duty a year, before counting the smaller 10 and 20 percent lanes.

That duty is paid at the border by the importer, not by the exporting country. From there it flows into the same channel every other input cost does: toward the hospitals, insurers and pharmacy-benefit managers who buy branded drugs, and ultimately toward patients through premiums and list prices. How much is absorbed versus passed on is a separate question we are not answering here. The point is only that a modest average rate on a large base is not a rounding error in dollars, even when it is a rounding error relative to '100 percent.'

The 100 percent is a lever to pull pricing concessions loose. The 15 percent is what the Treasury actually banks.

The low July reading is not the steady state. It is the front end of an escalation written into the proclamation itself. The 0 percent lane for Annex II signers is scheduled to expire on January 20, 2029, at which point that value gets pulled up toward the caps or the 100 percent default. The 20 percent onshoringonshoringMoving manufacturing back into the United States; companies that commit to build US plants get a lower tariff rate for now. lane for Annex III companies is scheduled to rise toward 100 percent on April 2, 2030. The structure is designed so the pressure builds over years, not so it lands all at once on day one.

A final discipline note on cause and effect. Several large drugmakers agreed to most-favored-nation pricing concessions in the same window that this tariff was negotiated, and it is tempting to say the tariff produced them. We cannot show that. The concessions were negotiated alongside the tariff, under the same administration and the same threat, but the proclamation does not prove the counterfactual, what those companies would have agreed to without it. What we can say is narrower and firmer: on July 31 the 100 percent headline will touch a low-single-digit slice of import value, the trade-weighted rate collected will be around 8 percent at full phase-in, and both numbers are scheduled to climb.