Tariffs Added 2.9 Points to Goods Inflation, About 0.6 to All Prices
The 2.9 points measures 67 goods categories, about a fifth of the basket, against less-exposed goods in February. Across a whole budget it works out to about 0.6 points (0.8 if all goods behaved alike), or roughly $320 to $460 a year for an average household by Daybreak Wire's arithmetic.
Two lines run across one New York Fed chart. The solid one, actual 12-month inflation in a sample of consumer goods, ends February 2026 at 2.0 percent. The dashed one, the authors' estimate of that same inflation with no tariffs, ends at minus 0.9 percent. The gap is the 2.9 percentage points quoted all week, and it describes 67 categories of goods, measured against less-exposed goods, as of a month that is now seven months old.
Mary Amiti and Sebastian Heise of the New York Fed and David Weinstein of Columbia University published the estimate on Tuesday, Oct. 6, in the New York Fed's Liberty Street post, which summarizes the staff report behind it. The paper calls its findings preliminary. The post's headline sentence says that by February 2026, tariffs had contributed 2.9 percentage points to goods price inflation
.
Post by @NickTimiraos
A percentage point here is a unit of the rate itself. The 2.9-point gap is a difference in the yearly change, which is not the same as goods costing 2.9 percent more. The underlying rule of thumb is smaller: For every percentage point increase in average tariffs, consumer goods prices increase by about a quarter of a percent after one year.
What the 2.9 is measured against
The sample is 67 non-oil categories from the Consumer Price Index, weighted by 2022 spending. Their weights, the paper says, sum to 20.2 percent of the total consumption basket, and cover 75 percent of consumer goods expenditures.
Services, about two-thirds of the basket, are excluded.
Which 67? CNBC reported that The researchers didn't say which 67 types of goods they evaluated
, and Breitbart makes the same complaint.
The second limit is one the authors print themselves. The method compares heavily tariffed goods with less-exposed goods on each date, so anything that moved all prices together cannot be separated out, the familiar missing-intercept problem. Hence this sentence in the staff report:
"The counterfactual therefore treats these relative price effects as absolute ones, and should be read as illustrative of the magnitude of the tariff-induced price increases for the goods in our sample."
Mary Amiti, Sebastian Heise and David E. Weinstein, New York Fed economists (Weinstein of Columbia), staff report
Two readings have drifted from that. Raw Story wrote on Oct. 7 that goods inflation was increasing to just under 3% by early 2026
, when the paper's actual goods inflation in the sample was 2.0 percent and 2.9 is the tariff contribution. CNBC's headline says inflation on many items was entirely due to tariffs
, while its own text frames the counterfactual as prices pulling back by almost 1%
.
From a goods index to a household budget
Spread across everything a household buys, the same effect is about 0.6 points (Daybreak Wire arithmetic: 2.9 times the sample's 20.2 percent share). If all goods (27.1 percent of the basket) behaved identically, it rises to about 0.8, an assumption and not a finding.
| Yardstick | Effect of tariffs | Where it comes from |
|---|---|---|
| 67 goods categories, 12-month inflation, February 2026 | 2.9 points | New York Fed Staff Report 1201 |
| Same effect weighted by the sample's 20.2 percent of the basket | about 0.6 points | Daybreak Wire arithmetic: 2.9 x 0.202 |
| If all goods (27.1 percent of the basket) behaved the same | about 0.8 points | Daybreak Wire arithmetic: 2.9 x 0.271; an assumption |
| Core PCE prices through February 2026 (different method) | 0.8 percent | Federal Reserve Board note, April 8, 2026 |
Other methods land in the same band. A Federal Reserve Board note from April by Robert Minton, Madeleine Ray and Mariano Somale found tariffs boosting core PCE prices as a whole by 0.8 percent
. Fortune reported in May that Dallas Fed researchers put March core inflation at 0.80 percentage points lower absent tariffs.
In dollars: the Bureau of Labor Statistics put average household spending at $78,535 in 2024, so the sampled categories are about $15,864 a year. Treating the 2.9 as a 2.9 percent price effect on that gives roughly $460; the level effect of about 2 percent forecast for August gives $317. Call it a back-of-envelope $320 to $460 a year. The first figure is a simplification, since the 2.9 is a rate gap and the level effect peaked near 3 percent.
Rough, because it mixes 2024 spending with the CPI's 2022 weights, covers only the sampled goods, and treats every household as the average one. It also sits well below the Tax Foundation's $1,000 estimate, cited by Fortune, for the 2025 tariffs, which measures all tariffs on all imports, a different question.
"Direct" means the higher price of imported consumer goods themselves; the rest is U.S. producers paying more for imported parts, and domestic firms raising markups as imports get dearer.
On precision: for a 10 percent tariff on all imports, the staff report estimates prices 2.6 percent higher after twelve months, with a 95-percent confidence interval of [1.45, 4.38]. It prints no interval for the 2.9, so one bold number sits on a band that is probably wide. And which products? The paper lists only a few examples, such as apparel and bakery items. The Fed Board note gives a wider sense of the spread, across 59 core goods categories rather than the New York Fed's 67: theoretical effects range from close to zero for books, newspapers, and computer software to about 8% for some appliances and information processing equipment.
Where the dissent lands and where it stops
Breitbart Business Digest answered on Wednesday, Oct. 7, and its strongest point is the denominator: In the first place, the paper does not show and does not even claim that tariffs pushed up overall inflation by 2.9 percentage points.
Translate it to the whole basket, it says, and you get roughly six-tenths of a point, which would put February's 2.4 percent overall CPI near 1.8 without tariffs. Does anyone really believe we'd be below two percent inflation if not for tariffs?
It also argues that pricier washing machines leave households less for other goods, so there might be no effect on overall inflation at all. It concedes that this paper is substantially stronger than the earlier tariff research.
Several points survive. The six-tenths conversion matches the arithmetic above. The authors concede the relative-effect limit, the full list of 67 categories is unpublished, and the staff report says its estimates do not account for the possible tariff impact on the exchange rate, the monetary-policy response, and a host of other general equilibrium forces.
The offset question is open.
What does not carry is the rhetorical question. The paper never says tariffs explain headline inflation, so the fact that headline inflation fell does not refute a relative-price estimate about 20 percent of the basket. The 1.8 percent is a product of applying a goods-only number to everything. The Fed Board and Dallas Fed estimates sit in the same range, though none settles the economy-wide offsets.
This team's work has met rougher reception. On Feb. 18, National Economic Council director Kevin Hassett said of an earlier New York Fed paper by overlapping authors: I mean, the paper is an embarrassment.
As of Thursday evening, no White House response to the Oct. 6 post had turned up in the coverage reviewed.
A year-long lag is not a year of rising inflation
Coverage stressed that effects take a year to arrive, which is true and easy to misread. At retail, about half of the direct effect is in place after three months and all of it by six. The indirect effect from U.S.-made goods more than doubles between six and twelve months.
The year is how long the ripple takes, not a year of fresh increases.
The authors' own forecast, which assumes tariffs stay at end-of-September levels except the announced January 2027 increase on Canadian cars, trucks and auto parts, separates the two ideas. The effect on the price level peaked near 3 percent in February and is forecast to ease to about 2 percent by August. The contribution to the 12-month change is forecast to fall to around zero by August, turn negative as 2025's big increases drop out, and turn slightly positive by mid-2027 as the Canadian tariffs pass through.
"Taken together, the two panels show that tariffs have left consumer goods price levels higher, while their effect on inflation fades."
Mary Amiti, Sebastian Heise and David E. Weinstein, New York Fed economists (Weinstein of Columbia), Liberty Street post
Prices about 2 percent higher stay that way until something pushes them down, while the inflation rate asks only how fast they are changing. Shelves cost more, yet the tariff push on the yearly rate is largely spent.
The government's numbers fit that. Core goods, the closest published gauge though not the New York Fed's basket, ran 1.0 percent over 12 months in February and 0.7 percent in August. Meanwhile all items rose 3.4 percent in the August CPI, released Friday, Sept. 11, with energy up 16.3 percent and gasoline up 27.4 percent. Energy is the biggest mover in the headline. (Earlier Daybreak Wire coverage of the 10-year Treasury yield is a separate story.)
And the February snapshot cannot score today's tariffs. The Supreme Court struck down the emergency-power tariffs that month, and a 10 percent surcharge replaced them; CNBC notes products from many countries now often face tariffs of about 10%.
By tax adviser UHY's account, that surcharge ran from Feb. 24 to July 24, when new Section 301 tariffs on 60 economies took over, and a 50 percent tariff on specified Canadian goods began Aug. 19.
Any "tariffs added X points" headline deserves four checks: which basket, against what baseline, as of what date, and level or rate. The 2.9 is goods, relative to less-exposed goods, February, rate.
The most useful fix is one nobody has the right to demand and the Fed could easily supply: the list of the 67 categories and their weights, which would let readers check half this argument themselves. The next real test is the September CPI on Wednesday, Oct. 14, which will show whether core goods are still cooling.
