Analyzing the effects of tariffs on prices and inflation

The Federal Reserve published two research notes about tariffs and inflation on the same day, April 8, 2026.

One, from the Board of Governors, found that tariffs raised core goods prices by 3.1 percent and explained the entirety of excess inflation in that category.¹ The other, from the Minneapolis Fed, found that the pattern of price increases across goods categories was inconsistent with tariffs being the cause, and that the categories where the story does hold together account for at most 0.2 percentage points of core inflation.²

Same question, access to the same data, different answers. And released on the same morning from different parts of the same organization.

The question is a politically charged one, and one that is hard to arrive at a definitive answer for. As in many cases in economics, there are a lot of confounding variables and it can be hard to identify them and measure their individual effects.

Three questions

1. Who pays the tariff? A duty is charged to the importer at the border. Does the foreign exporter cut their price to absorb it, or does the cost stay on the American side?
2. Do the tariffs reach the store shelf? If tariffs stay on the American side, does the importer pass the cost on to consumers or absorb it in margins?
3. Does it raise inflation? If consumer prices do rise, is that a one-time step up in the price level, or a sustained increase in the rate at which prices change?

Who pays the tariffs?

Let's start by trying to tease out the actual size of the tariffs.

There is some complexity because announced rates overstate what gets paid, for a few reasons: there are exemptions and carve-outs, announced tariffs that never take effect, importers switching to suppliers in countries that are not tariffed, and incomplete enforcement. A reasonably straightforward way to calculate the rate is to divide customs duties by the value of goods imported and you get the rate importers really paid.³

Once we do the math, the effective tariff rate is immediately visible. The rate sat near 1.7 percent through 2015 to 2017. The first trade war (in 2018 and 2019) lifted it only to about 3.1 percent, and it was still 2.4 percent in the first quarter of 2025. It then jumped to 11.1 percent by the fourth quarter of 2025, and has since eased back to 8.8 percent as of the second quarter of 2026.

0246810

201620182020202220242026

8.8 percent of goods imports as of Apr 2026US Bureau of Economic Analysis, Federal Government Current Tax Receipts: Customs Duties · US Census Bureau, Imports of Goods, Balance of Payments Basis

Three things follow from that line.

1. The 2025 tariffs are much larger. Roughly three and a half times the first trade war, which makes 2018 a useful guide but not a complete one.
2. The first trade war never ended. The rate never fell back to its starting point of about 1.6 percent. It sat between 2.4 and 3.3 percent for the six years after, so 2025 was stacked on that floor rather than raised from nothing.
3. The rate peaked in late 2025. That matters once the question shifts from "did tariffs raise prices?" to "are tariffs still raising inflation?"

So now we can dig in to who actually pays the tariffs.

Luckily, the answer is unusually clear: American importers, and the people they sell to pay the tariffs. Amiti, Flanagan, Heise and Weinstein at the New York Fed matched import prices to tariff changes through November 2025. Once they did so, they were able to establish that foreign exporters absorbed 6 percent of the burden in the first eight months of 2025, 8 percent in September and October, and 14 percent in November. Nearly 90 percent of the tariff cost increases stayed with American firms and their customers.⁴

That is the same basic result researchers found in the first trade war. Amiti, Redding and Weinstein found complete pass-through of the 2018 tariffs into domestic prices of imported goods, with foreign export prices essentially unchanged.⁵ Clausing and Obstfeld's August 2026 survey gathers five separate studies that show the burden being put on American buyers, with estimates ranging from about 90 percent of the tariff to more than the entire tariff.⁶

The theory that foreigners pay the tariffs hasn't seen much convincing evidence yet. If it turns out to hold at all, it will be through the effect tariffs have on exchange rates - and exchange rates move for too many reasons to state that reliably.

Do the tariffs reach the store shelf?

A tariff to the importer does not directly or immediately result in consumer inflation. Firms have inventories, contracts, pricing calendars, competitive constraints, and margins, all of which provide downward pressure on the final price of a product.

A review of the impacts from 2018 makes that point clearly. Cavallo, Gopinath, Neiman and Tang compared prices at the point of import with prices in stores. Border prices rose by nearly the full tariff, but retail prices rose much less. The gap came out of margins, which means a large part of the burden landed on American businesses rather than American shoppers.⁷

A similar outcome has been observed with the 2025 round of tariffs:

• Minton, Ray and Somale¹ at the Board of Governors use a distributed lag regression across 59 core goods categories and find full dollar-for-dollar pass-through about seven months after a tariff change. Through February 2026, they estimate a cumulative effect of 3.1 percent on core goods PCE prices and 0.8 percent on core PCE overall.
• Amiti, Heise and Weinstein⁸ estimate that about 26 percent of the tariff increase reaches consumer prices. Of that effect, 64 percent is direct and 36 percent is indirect, with the indirect channel running through imported inputs and domestic producers raising markups when import competition gets more expensive. That indirect channel takes nine to twelve months.
• Cavallo, Llamas and Vazquez⁹ track prices in near real time and find incomplete pass-through after six months, with a cumulative CPI effect of roughly 0.8 percentage points by early 2026.

Those three estimates do not agree on the scale of the impact to consumers, but they do agree that consumers are impacted. In all cases it was found that some amount of the tariff cost eventually reaches consumers, but that it reaches them slowly, taking the better part of a year to arrive.

That is enough to raise the price of a tariffed good. The impact on inflation is a different question.

Do the tariffs raise inflation?

Inflation is a rate, not a level. It measures how fast prices are climbing, usually against the same month a year earlier. A one-time increase moves the price level up, but once the increase is complete, prices stay. Because inflation measures the amount of change, a one-time increase will show up one time in the data for a one-time change, and the overall inflation rate will only reflect that change once.

Concretely, for a good that goes up in price once and then holds:

Year	Price	Inflation
2003	$100
2004	$100	0%
2005	$100	0%
2006	$105	5%
2007	$105	0%
2008	$105	0%

Inflation shows up once to reflect the price increase, but then returns to 0% and stays there once the price increase stops.

Since a tariff is a tax on imports, for the importer it raises the cost of the affected goods. If that cost reaches consumers, the price level increases and that change is reflected in the inflation rate. But once the adjustment is complete, there is no more inflation. A higher price level is not the same thing as a continuing higher inflation rate.

Since the implementation of the second round of tariffs, Fed Chairman Powell's base case has been exactly that: a one-time shift in the price level.¹⁰ The caveat is that "one-time" does not mean "instant." If the direct channel takes seven months and the indirect supply-chain channel takes another nine to twelve, the same one-time shock can keep measured inflation elevated for a year or more. While it is passing through the data, it looks like ordinary inflation.

There is also a serious counter-argument to the pure "one-time shift" school of though. Halbersleben, Jordà and Nechio at the San Francisco Fed studied 16 advanced economies over roughly 40 years and estimated the inflation response to a 10 percent tariff increase. Headline inflation falls by about 1 percentage point in the first year (mostly through lower energy prices), which is speculated to be attributable to demand destruction. Goods inflation then peaks 1.2 percentage points above baseline in year two, and services inflation rises 0.6 points by year three and remains 0.5 points elevated in year four.¹¹

That result should be considered in context. It is an average across decades of mostly smaller tariff changes, and the current post-"liberation day" episode moved the realized rate from 2.4 percent to 11 percent in three quarters. But it is a useful reminder that tariffs do not arrive in a vacuum. They raise import costs and reduce purchasing power at the same time. This can mean forces that counter each other, increasing the difficulty in measuring individual effects.

The current data points to a fading tariff impulse

Current inflation data seems to indicate that tariffs look like a real contributor to goods prices, but only serve as a weak explanation for ongoing inflation by themselves.

As of July 2026, core PCE was rising 3.34 percent year over year and core CPI was rising 2.47 percent.¹²

If tariffs were the dominant current driver, the cleanest place to see it would be core goods. That is the part of the basket most directly exposed to imported merchandise. Core goods CPI is rising 0.78 percent, compared with a 2015 to 2019 average of minus 0.33 percent. That is meaningfully above the pre-pandemic norm, but core goods are only about a quarter of core CPI. Even crediting tariffs with the entire excess in core goods inflation, the arithmetic gets you to roughly 0.26 percentage points of core inflation.

That sits close to the Minneapolis Fed estimate and well below the Board of Governors estimate, but it is incomplete. It credits tariffs with all excess core goods inflation, including anything caused by non-tariff shocks. But it also misses the indirect channel into services that Amiti and coauthors describe.

The broader inflation mix is harder to square with a simple tariff story. Shelter is rising 3.16 percent, slightly below its pre-pandemic average of 3.29. PCE services are running at 3.69 percent against a pre-pandemic 2.28, and services are not directly tariffed. Total PCE goods inflation jumped from 1.8 percent in February to 4.8 percent in May, but that was rooted in changes caused by energy pricing increases, not trade policy.

The St. Louis Fed's August update reaches the same basic conclusion. Before February 2026, tariffs accounted for a large share of the overshoot above the 2 percent target. Since March, other forces, especially energy, have done more of the work.¹³

So the current data supports a middle position. Tariffs pushed some prices higher, and the pass-through was large enough to matter. But the evidence for tariffs as the main ongoing inflation engine has weakened as the effective tariff rate has rolled over and the inflation pressure has moved elsewhere.

What we can say now

Four things are settled well enough to influence policy.

Americans pay. Across two distinct tariff regimes and more than a dozen studies, foreign exporters absorbed almost none of the cost. The New York Fed's work at the tariff-code level leaves nearly 90 percent of the burden with American firms and their customers. Nothing in the literature seriously contests this.

Tariffs raised the prices of tariffed goods, slowly and incompletely. Margins absorbed part of it and inventories delayed the rest. The direct channel took about seven months, and the indirect one ran nine to twelve months behind it.

The aggregate inflation effect was real and bounded. Credible Federal Reserve estimates span 0.2 to 0.8 percentage points of core inflation. The range is wide, but the ceiling is illustrative: even the most generous estimate leaves most of the overshoot to something else.

The impulse has passed its peak. The effective tariff rate topped out at 11.1 percent in the fourth quarter of 2025 and was down to 8.8 percent by the second quarter of 2026. The estimated contribution to inflation stopped rising in early 2026. Since March, the St. Louis Fed puts energy ahead of trade policy in explaining what is left of the overshoot.

Together those four describe a one-time increase in the price level, stretched over a year or more, rather than a new inflation regime.

Sources

1. Detecting Tariff Effects on Consumer Prices in Real Time, Part II (Minton, Ray and Somale, Board of Governors, April 2026). Full pass-through at seven months; 3.1 percent on core goods PCE, 0.8 on core PCE.
2. Tariffs can't explain rising goods inflation (Mehrotra and Waugh, Minneapolis Fed, April 2026). The cross-category mismatch, published the same day as the note above.
3. Customs duties and goods imports (FRED). The effective tariff rate above is these two series divided, quarter by quarter.
4. Who Is Paying for the 2025 U.S. Tariffs? (Amiti, Flanagan, Heise and Weinstein, New York Fed, February 2026). The incidence split, at the ten-digit tariff-code level.
5. The Impact of the 2018 Tariffs on Prices and Welfare (Amiti, Redding and Weinstein, 2019). Complete pass-through into domestic prices in the first trade war, with foreign export prices unchanged.
6. Who pays for tariffs? Insights from recent research (Clausing and Obstfeld, PIIE, August 2026). A survey of seven studies of the 2025 round.
7. Tariff Pass-Through at the Border and at the Store (Cavallo, Gopinath, Neiman and Tang, 2021). Border prices against store prices, and the margin compression between them.
8. The Anatomy of Tariff Pass-through into Consumer Prices (Amiti, Heise and Weinstein, NBER, July 2026). The 26 percent figure, and the direct/indirect decomposition.
9. Tracking the Short-Run Price Impact of U.S. Tariffs (Cavallo, Llamas and Vazquez). Near-real-time price tracking, and the incomplete pass-through at six months.
10. Economic Outlook (Chair Powell, September 2025). The source of both the "one-time shift in the price level" base case and the caveat that this "does not mean 'all at once.'"
11. The Effects of Tariffs on the Components of Inflation (Halbersleben, Jordà and Nechio, San Francisco Fed, March 2026). Sixteen countries, forty years, and the first-year disinflation result.
12. Core PCE, core CPI, core goods CPI and shelter CPI (FRED). Every inflation figure above is computed from these directly.
13. Tariff Effects on Inflation Stabilize in Recent Months (St. Louis Fed, August 2026). The most recent read, and the handoff from tariffs to energy.