Inflation Is 3.4%. It Is Also 25%. Both Numbers Are From the Same Report.
The headline inflation rate averages categories moving 46 times faster than each other. Airline fares are up 25.6% and education is up 0.6% — in the same month, from the same BLS release.
The headline CPI-U figure for the 12 months to July 2026 is 3.36%. But that is an average across categories ranging from 0.55% (education and communication) to 25.55% (airline fares), so the headline number describes a spending pattern almost nobody has.
Get new posts in your inbox
No spam, unsubscribe anytime.
About the Author
Sujit KarkiFinance Researcher & Market Analyst
Independent finance researcher and market analyst with expertise in macroeconomics, equity markets, and personal finance. I help regular investors make better-informed decisions through rigorous, data-driven analysis.
Headline inflation for the 12 months to July 2026 is 3.36%.
Within the same release, airline fares rose 25.55% and education and communication rose 0.55% — a 25-point spread and a 46-fold ratio.
Gasoline is up 24.64% and energy overall 14.73%, while medical care is up 1.70%.
Seven of the eleven component categories measured sit below the headline figure. The average is pulled up by a few fast movers.
Your own inflation rate depends on your spending mix, and can differ from the published number by many points in either direction.
Every month a single number gets reported as the inflation rate, and every month
a large number of people read it and think it is wrong, because it does not match
what happened to their money.
They are not imagining it. The headline number is an average, and the things it
averages are moving at wildly different speeds — not by a little, and not
occasionally.
One report, twelve answers
These figures are all from the same Bureau of Labor Statistics release, the same
index basis, the same month:
The headline figure is 3.36%. The fastest-moving category is 46 times the
slowest. The spread from top to bottom is 25 percentage points.
Same month, same report — the extremes
Category
12-month change
Against the headline
Airline fares
25.55%
7.6x the headline
Gasoline
24.64%
7.3x
Energy
14.73%
4.4x
All items (headline)
3.36%
—
Medical care
1.70%
half the headline
Education and communication
0.55%
one sixth
Why the average describes nobody
The headline rate weights each category by how much a statistically average
household spends on it. That is the right way to build a national index. It is
also why the result matches almost no individual.
Consider two people in the same city in the same month:
One drives to work, flies to see family twice a year, and rents. Their basket
leans on gasoline (24.6%), airline fares (25.6%) and shelter (3.2%). Their lived
inflation is well into double digits.
One works from home, owns outright, and spends heavily on healthcare and a
child's education. Their basket leans on medical care (1.7%) and education
(0.55%). Their lived inflation is close to nothing.
Both read "inflation is 3.4%" and one of them concludes the statistics are lying.
Neither is wrong about their own experience. Seven of the eleven component
categories here sit below the headline — the average is dragged upward by a
small number of fast movers, which is exactly the shape that makes a mean
unrepresentative.
Pro Tip
This is also why "inflation is coming down" and "everything is still expensive"
are both true at once. A falling rate means prices are rising more slowly, not
that they are falling. Cumulative increases stay in the price level, and in
categories like airline fares they compound from an already elevated base.
Work out your own rate
You do not need a model for this, just an honest look at where your money goes.
Take your last three months of spending and split it into these
categories. Rough percentages are fine — this is not accounting.
Multiply each share by that category's rate from the chart above.
Add them up. That is your personal inflation rate, and it is the number
your budget actually experiences.
Someone spending 20% on transport and fuel, 30% on shelter and 15% on food gets
a very different answer from someone spending 5%, 45% and 10%. The exercise takes
ten minutes and is more useful than any headline, because it tells you which of
your own line items is doing the damage.
The practical follow-on: the fast-moving categories are where substitution
actually pays. A 25% rise in airline fares is worth restructuring travel around.
A 1.7% rise in medical care is not worth reorganising your life over, however
much the total bill stings.
What this does not mean
Three honest limits, because this argument gets misused in both directions.
It does not mean the official figure is wrong or manipulated. It is a
correctly constructed weighted average, published transparently with every
component available for anyone to check — which is exactly how this post was
written. The problem is in how a single number gets reported, not in how it is
computed.
It does not mean your rate is necessarily higher. The categories running
hottest here are energy and travel. If you neither drive nor fly much, your
personal rate this year is likely below 3.36%, not above it. The spread cuts
both ways and people tend to assume only one direction.
Category rates are not stable. Energy is the most volatile component in the
index and its 14.7% could be materially different next month. A personal rate
computed today is a snapshot, not a trend.
Methodology
Every figure comes from the BLS public API. The computed table is published with
this post as cpi-categories.json; the script that
fetches and computes it is
scripts/cpi-categories.mjs.
Data. CPI-U, US city average, not seasonally adjusted — the same basis the
headline figure is published on, so the comparison is like for like. Twelve
series: all items, plus eleven components. Twelve-month change is the latest
published month against the same month a year earlier.
Period. Latest available at time of writing is July 2026.
Reproducing it.npm run cpi:categories in the repository re-fetches from
BLS and prints the current table. The figures above will age; the command will
not.
A gap worth knowing about. October 2025 is missing from the published CPI
series entirely — the data jumps from September to November that year. It does
not affect the 12-month changes reported here, which compare July to July, but it
is a real hole in the record and it appears across the whole CPI family.
Disclosure. The commentary was drafted with AI assistance; the data
selection, the calculation and the conclusions are mine.
Sujit Karki||5 min read