US inflation: every measure on one chart

What is actually pushing it up

Split the whole shopping list into four parts. The numbers below show how much of the year of inflation to September 2026 each part is responsible for. They are in percentage points, so if inflation is 3 percent and housing shows 1.1, housing accounts for 1.1 of those 3 points. A minus number means that part pulled inflation down instead.

• Housing +1.14
• Services, not housing +0.77
• Goods +0.14
• Food and energy +1.67

The four parts add up to +3.72 points, against inflation of 3.7%.

A part matters for two reasons: how much its own price moved, and how big a slice of your spending it is. Housing is about a third of the whole list, so it dominates this picture even in a quiet year. Something small like plane tickets can double in price and still barely register here.

Every measure, side by side

Every measure here is read at the same moment, over four different lengths of time. The left column is the one people quote. The shorter columns pick up a change sooner, and they are jumpier for the same reason. A measure that looks calm over a year can already be moving in the last three months, which is often the first hint that something has turned.

What gets thrown away, and why

Core inflation removes the same two things every month, food and energy, whether or not they were the odd ones out. The measures below work differently. They line every price up by how far it moved this month and then cut off both ends of that line, which is all that "trimming" means. The list of what gets cut is rebuilt every month, and nothing is banned for good.

• Core CPI, for contrast

  The familiar one, and the odd one out on this list, because it does not look at what prices did at all. It removes the same two groups every single month by name, food and energy, whether or not they were the odd ones out that month. What you get is a quick, steadier read that everyone quotes, and the rest of this list exists because "always the same two" is a rough rule.

• Median CPI

  Rank every price in the basket by how much it moved this month, weighted by how much people actually spend on it, then keep the single one sitting at the halfway point and throw away everything on both sides of it. One reading survives and the rest go. That makes it the steadiest line on this page, and one category going wild can barely move it.

• 16% trimmed-mean CPI

  The same ranking, but it keeps the whole middle of the pack and averages it instead of keeping one point. It is built from 45 spending groups, and what comes off is the bottom 8 percent and the top 8 percent, measured by share of spending rather than by counting items. That leaves 84 percent of the basket, so it is calmer than core while still using most of what households buy.

• Trimmed-mean PCE

  The same idea applied to the wider basket, the one the 2 percent goal is written on. It takes 24 percent off the bottom and 31 percent off the top, by share of spending, leaving 45 percent. The two ends are deliberately different sizes: that split was chosen by testing which one tracked the underlying trend best. Use it to judge the trend on the same basket the target is set on.

• Sticky-price core CPI

  Not a trim at all. It sorts the basket by how often each price changes rather than how far it moved, keeps only the slow half, and takes food and energy out as well. Anything that reprices more often than once every 4.3 months is gone. A shop that only reprices once a year has to guess at future inflation when it does, and that guess is what this measure picks up.

• Flexible-price core CPI

  The other half of that same sort: only the prices that change often, again without food and energy, so anything that reprices less often than once every 4.3 months drops out. It shows what is reacting to the economy right now. It turns early and it is jumpy.

Sticky and flexible are the odd ones out here. The trims ask how far a price moved this month, so what gets cut changes every time. The sticky split asks how often a price changes at all, which is a settled fact about the item. Petrol is always flexible, even in a quiet month.

Which prices count as slow, and which as fast

The number is how many months typically pass between one price change and the next. Anything slower than 4.3 months counts as slow.

For the wider basket you can see exactly which items got cut, because the month's included and excluded list is published as a spreadsheet every release. For the CPI trims only the rule is published, not the monthly casualty list. The pattern is the same either way, and things with famously jumpy prices land in the cut ends far more often than something steady like rent.

Dropping food and energy every time would be simpler, but a fixed list makes two claims it never states out loud: that what it removes is always noise, and that what it keeps is always signal. Neither is true. A one-off jump in something you kept, like a tax rise on tobacco, goes straight into the number, where a trim would have thrown it out.

What each one actually means

Short answers in plain words. You do not need any of this to read the chart, but it is here when a name on the list makes no sense.

The numbers

This page shows the last ten years. The full history, every component, all four windows and the machine-readable feed come through the Helious API, and through MCP at api.helious.io/mcp if you want your own tools to read it directly.

How this page is built

FAQ