On September 11, the Bureau of Labor Statistics reported that consumer prices rose 3.4% over the 12 months ending in August. If that number didn’t match your experience, you’re not imagining things. Over that same year, gasoline rose 27.4% and airline fares jumped 23.4%, while used cars got 2.3% cheaper and prescription drugs and other medical care goods fell 2.7%. The 3.4% is an average of all of that, weighted by how a typical urban household spends money.
You aren’t a typical urban household. Nobody is, exactly. Knowing how the Consumer Price Index gets built helps you read the headlines, and it also helps you figure out what inflation is doing to your own budget, which may be a lot more or a lot less.
What the CPI measures
The CPI tracks the change in prices that consumers pay for a fixed basket of goods and services. It isn’t a price level. It won’t tell you that groceries cost more in Boston than in Omaha. It tells you how much prices changed compared with an earlier period.
The data collection is enormous. According to the BLS technical note for the August 2026 release, prices come from 75 urban areas, about 6,000 housing units, and roughly 22,000 retail and service businesses, including supermarkets, department stores, and hospitals. BLS staff gather prices by visiting in person, by phone, and online. Sales taxes tied to a purchase are included.
The main index is the CPI-U, which covers all urban consumers, more than 90% of the U.S. population. There’s also the CPI-W, which covers households that earn most of their income from hourly and clerical work, about 30% of the population. In August, the CPI-W rose 3.5% over the year, a little faster than the CPI-U. A third version, the chained CPI-U, rose 3.3%.
Weights decide everything
Price changes are only half the math. The other half is how much each item counts, and BLS calls those weights “relative importance.” They come from the Consumer Expenditure Surveys, which record what households actually spend money on. Since 2023, BLS has updated the weights every year using spending from two years earlier.
The single biggest piece is shelter. In the December 2025 relative importance table, shelter made up 35.6% of the CPI-U. Food at home was 8.3%. Gasoline was 2.9%, and all energy combined was about 6.3%.
That’s why gas prices up 27% didn’t push the overall index up 27%. A rough back-of-the-envelope calculation using those weights suggests gasoline alone added somewhere around three quarters of a percentage point to the annual rate. That’s a big share of 3.4%, but a small share of 27%.
The odd part: owners’ equivalent rent
Most of that shelter weight isn’t rent. The largest single item in the whole CPI is something called owners’ equivalent rent, at 26.2% of the index, compared with 7.8% for actual rent paid by tenants.
Owners’ equivalent rent estimates what homeowners would pay to rent their own homes. BLS uses it because a house is partly an investment, and the CPI tries to measure the cost of living in one, not the cost of buying one. It makes sense on paper. In practice it means a homeowner with a fixed-rate mortgage has a quarter of the CPI basket tied to a cost that never shows up in their bank account. Their mortgage payment didn’t go up 3% this year. Their property taxes and insurance might have, but that’s a separate story.
Headline, core, and the number the Fed watches
News reports usually cite two figures. The headline number (3.4% in August) includes everything. Core CPI strips out food and energy, because those prices swing hard from month to month. Core rose 2.4% over the year, which tells you that most of the gap between the two came from energy. The energy index was up 16.3%.
You’ll also see monthly changes. August’s all-items index rose 0.4% on a seasonally adjusted basis, after 0.1% in July. Seasonal adjustment smooths out patterns that happen every year, like apparel prices dropping after the holidays. BLS notes that the unadjusted numbers are the ones most relevant to what consumers actually pay.
For what it’s worth, the Federal Reserve’s 2% inflation target is based on a different measure, the personal consumption expenditures price index, not the CPI. The two usually move together but rarely match.
Where the CPI hits your money directly
Several formulas that decide what you get paid, and what you earn, are written around the CPI.
Social Security’s annual cost-of-living adjustment is based on the CPI-W, comparing third-quarter averages. The September CPI, due October 14, 2026, completes that quarter, so the 2027 COLA gets announced that day. The Social Security Administration publishes the formula and the result.
Series I savings bonds use the CPI-U to set the inflation part of their rate twice a year. Federal tax brackets and the standard deduction are adjusted using the chained CPI. Many union contracts, pensions, and leases are tied to CPI as well.
Your savings account isn’t directly tied to it, but inflation still matters. If your high-yield savings account pays less than the inflation rate, your money is losing buying power even as the balance grows.
Why your personal inflation rate is different
Take August’s actual 12-month changes and apply them to two households with very different budgets.
The first rents an apartment and drives to work. Say rent is 35% of spending, gasoline is 8%, groceries 12%, restaurants 6%, electricity 4%, and everything else 35%. Apply the August figures (shelter up 3.0%, gasoline 27.4%, food at home 2.2%, food away from home 3.4%, electricity 3.8%, and core’s 2.4% for everything else) and this household’s inflation comes to roughly 4.7%.
The second household owns a home with a fixed-rate mortgage and works remotely. Say the mortgage is 28% of spending with no change at all, gas is just 2%, food and electricity are the same as the first household, and everything else is 48%. That household’s rate comes out around 2.3%.
Same economy, same month, and one family’s costs rose about twice as fast as the other’s. These are simplified examples, but the pattern holds. Commuters felt this year’s gas spike. Frequent flyers felt airfares. Someone who bought a used car got a small break.
Estimating your own number
You can do a rough version of this with your bank and card statements. Add up a few months of spending in the big categories: housing, transportation fuel, groceries, eating out, utilities, insurance, and everything else. Turn each into a percentage of the total. Then multiply each share by the matching 12-month change from the latest BLS release and add them up.
It won’t be precise, and that’s fine. The point is to see which categories are driving your costs. If most of your inflation is coming from gasoline, a raise that matches the headline CPI may still leave you behind. If your big costs are locked in, you may be doing better than the news suggests. Either way, you’ll be budgeting with your own number instead of the national one.