Margin absorption
Which sectors are absorbing cost shocks and which are passing them on
MacroRoom·Measure by MacroRoom, from BLS
A supermarket pays 10% more for apples than it did a year ago. Does the price on the shelf go up 10%? Often it goes up less, and the difference comes out of what the shop keeps on each sale. That difference is the wedge: how much less — or more — shelf prices rose than the prices producers charged, over twelve months. A negative wedge means shoppers were shielded from part of the increase. Where the money went is a separate question, and the retail margin is the evidence.
- Sectors absorbing
- 9 of 15
- Where the shelf price rose less than producer costs over the past year, August 2026.
- Median pass-through
- 0.76
- Historically, about 76% of a producer price move reaches the shelf in the typical sector.
- Widest gap
- −14.0 pt
- Gasoline (all types): producer prices up +38.2%, shelf up +24.2%.
- Sectors covered
- 16
- Each matched across producer price, retail margin and shelf price by hand.
DataMonthly through August 2026
Who is absorbing right now
Each sector at its own latest month. Bars to the left mean the shelf price rose less than producer costs. Hover for the breakdown; click a bar to load that sector below. Sectors marked † are more than a month behind.
One sector, over time
Producer price, retail margin and shelf price, all as 12-month changes. Drag across the chart to zoom.
How much gets through, on average
How much of a producer price move reached the shelf, by sector, over the full period or any window you choose. Bars past 1.0 mean shelf prices moved more than producer prices — the opposite of absorption.
Methodology
Three series per sector, matched by hand: the producer price index, the retail trade margin, and the consumer price index. All are 12-month log changes, which removes the stable seasonality in these unadjusted series. The wedge is the shelf change minus the producer change.
The margin is a price, not a markup
The retail margin series measure the price of the retailing service — what the seller charges for moving a good to the shelf — not a percentage markup. A margin can rise while the wedge is negative. The two are not the same statement and the chart does not treat them as one.
Pass-through is not one number
Pass-through is the slope of the shelf change on the producer change, ordinary least squares with a constant. The published estimate is over the full period, 2009-01 to 2026-08; the same slope on months where producer prices rose and where they fell is in the download. The custom window is a reader's tool: it is estimated in the browser from the same series, over whatever range is chosen. Twelve-month changes overlap, so a short window holds few independent observations; the month count in the tooltip is the measure of weight. No standard errors are shown.
What it does not say
- This describes what margins and prices did. It does not say what caused inflation, and it does not attribute a cause to any firm or sector.
- Six sectors — the food group and alcohol — share a single margin series for food and alcohol retailing. Their margin figures are identical by construction, and the variation is at the retail-industry level, not the product level.
- Two producer series do not match their consumer series in scope. Meat, poultry, fish and eggs maps to a producer index covering meat alone. Fruit and vegetables maps to a farm-level index while the consumer series includes processed goods.
- The margin series begin in March 2009, so on 12-month changes the panel effectively starts in 2010, except toys, whose industry margin series runs from 2008. There is no reading for the 2008 crisis.
- October 2025 consumer prices were never published; twelve-month changes spanning that month are absent rather than interpolated.
What was retired
An earlier version of this work tested whether the direction of margins could identify the source of inflation — whether a price rise was driven by costs or by markups. That claim was retired under kill criteria set before the results were seen. The classifier is descriptive, not causal, and this monitor does not revive it. The pass-through slopes above are descriptive, not a re-run of that test: they follow a rule fixed before the first rebuild, and no significance is attached to them.
Inputs and sources3 series per sector
The series exist publicly. What does not is the mapping between them — which producer index, which margin index and which consumer index describe the same good. That crosswalk is ours, was built by hand, and is published below.
Download
We built this measure, so we host its data.