Donald Trump has been announcing changes to American tariffs with bewildering speed. What is the tariff on imports from China right now? 54%? 104%? 145%?

Actually, it’s none of those. Despite the rates swirling around in news headlines, Mr Trump’s tariffs are actually a mishmash of rules, imposed via executive orders, on top of America’s 17,000-line tariff menu. Some, such as the “fentanyl” duties, are sweeping. Others, such as levies on steel and aluminium, are targeted. Some items, such as smartphones, are exempt. We’ve parsed all of these rules and estimated what tariffs on each country would be, were importing patterns unchanged from 2024. As of September 26th, China still faces one of the highest rates, at 36.5%, followed by India at 33.6%.

The impact of all this on Americans’ shopping bills is hard to measure, and not just because the tariffs keep shifting. One reason is that the consumer price of an imported product reflects not just the cost of the good itself but also retailers’ rents, wages, transport costs, profit margins and more. Another is that tariffs on “intermediate” goods, such as raw materials used to make other things, affect consumer prices indirectly. Even service industries depend on imported goods—think of lawyers using smartphones or computers made abroad. If their costs rise, they may pass some of the increase on to customers. And, because the tariffs do not apply to goods already in transit, it will take time before the new taxes start to bite.

Tariff effect on domestic cars and trucks*, % of consumer spending on category

1.

Fentanyl tariffs on China

2.

Tariffs on Canada and Mexico

3.

Tariffs on automobile imports

4.

China tariffs put on hold

5.

Tariffs on steel and aluminium doubled

6.

“Reciprocal” tariffs go into effect

7.

Japan trade deal takes effect

*New vehicles. Owing to levies on parts

Despite this daunting complexity, it is still possible to calculate tariff burdens as a share of consumer spending in each category of goods and services. Take new domestic passenger vehicles. Even though American-made cars face no direct tariff, they contain many foreign components: engines from Mexico, say, seats from China, brakes from Germany and raw steel from Canada. Every time Mr Trump changes his tariffs on any of those countries or products, he changes the cost of producing Buicks, Dodges and Fords.

Even at its peak, the overall tariff burden on domestic cars was a manageable 4%. But for other types of spending—particularly on imports of finished goods from China—it has been extreme. Electric personal-care appliances, such as hairdryers and toothbrushes, were saddled with a whopping 59% effective rate in late April. Below, you can explore the tariff burden for all spending categories in the Personal Consumption Expenditures (PCE) index—the Federal Reserve’s preferred measure of inflation—on every day of Mr Trump’s term, broken down by the contribution from each country. Given the president’s penchant for backtracking on trade threats, these numbers reflect only levies that have already taken effect, and thus exclude the bulk of the tariff barrage he announced on July 31st. They account for tariffs on imported inputs feeding through into domestic production costs, and for all costs besides merchandise, such as labour. They also assume that production patterns remain unchanged from 2024, and do not account for any changes in firms’ behaviour in response to tariffs.

How much are these tariffs actually costing American consumers? For any individual spending category, there is no way to separate the impact of import taxes from those of other factors that also affect prices. But for the economy as a whole, we can produce an estimate by comparing inflation rates for different types of spending over time.

To produce the following graphic, we first compare inflation in each spending category during the final six months of Joe Biden’s presidential term with the overall average in that period. For example, from August 2024 to January 2025, the price of eggs increased by 24 percentage points more than the PCE index as a whole did. Next, we repeat this exercise for Mr Trump’s current term: from January to August of this year, inflation in eggs was 13 points below the average. Thus, the gap in inflation between eggs and the overall PCE basket fell from +24 points to -13, a decline of 37 points. Finally, after calculating this “difference in differences” for all spending categories, we group them into three buckets: those whose relative inflation rates have increased since Mr Trump’s inauguration; those that have seen decreases; and those that are similar.

If tariffs are indeed generating additional inflation, we would expect that the bin with increasing relative inflation rates should include lots of categories with high tariff rates, whereas the one with decreasing rates should contain mostly those with low tariff burdens. This pattern does exist. For now, however, it looks fairly weak. So far, we find that each percentage point of additional tariff burden is associated with 0.3 points of change in relative inflation. Given that current tariffs amount to about 1% of all consumer spending, this implies that the overall PCE index would be 0.3 points lower had those levies not been imposed.

Methodology

Our calculations are based on the method laid out in a paper by Omar Barbiero and Hillary Stein of the Boston Fed. The first step is grouping each of the 18,000 Harmonised Tariff Schedule (HTS) codes for individual products—some as specific as aluminium wheels with diameters of 57-63cm—into 402 broader “commodity” categories published by the Bureau of Economic Analysis (BEA), such as “other motor vehicle parts manufacturing”. We derive these mappings using the US International Trade Commission’s Commodity Translation Wizard and the BEA’s “PCE bridge” tool.

Next, we determine the contribution of imported intermediate goods to the cost of all the products that use them. To measure this, we use the BEA’s input-output (IO) tables, which list the value from each commodity category that feeds into each of 402 industry categories. A mathematical tool called the “Leontief inverse” allows us to account for second, third, and fourth-order effects: outputs that are later used as inputs to make other things, which in turn may become inputs into yet more complex goods. Finally, we use the PCE bridge to map the resulting figures onto 212 spending categories in the PCE price index. Our resulting estimates of how much each country’s exports wind up in each PCE category line up well with the published numbers in the Fed paper.

Finally, we look up the tariff rate applied to each pairing of a country and an HTS product code on each day, using ChatGPT to parse the Customs and Border Protection message board that records every change to tariff policies. Multiplying these rates by the dollar value of imports of each HTS code from each country in 2024, we wind up with the tariff burden in each PCE category, expressed as a share of total consumer spending.

Sources:

CBP; USITC; ”The impact of tariffs on inflation”, by O. Barbiero and H. Stein, 2025; The Economist