Why inflation is not the same in every city
Pick an area to see which spending groups moved most over 12 months, 5 years or 10 years. The reasons follow.
U.S. average, September 2025 to September 2026
+3.7 %
Fastest: energy +20.4 %; slowest: medical care +1.5 %.
| Energy | +20.4 % |
| Transportation | +7.9 % |
| Other goods and services | +4.1 % |
| Apparel | +3.4 % |
| Group | Change |
|---|---|
| All items | +3.7 % |
| Housing | +3.3 % |
| Shelter | +3.2 % |
| Food and beverages | +2.4 % |
| Food at home | +1.8 % |
| Energy | +20.4 % |
| Transportation | +7.9 % |
| Medical care | +1.5 % |
| Apparel | +3.4 % |
| Recreation | +2.8 % |
| Education and communication | +2.3 % |
| Other goods and services | +4.1 % |
| All items less food and energy | +2.6 % |
Prepared by Radif Partners · How the figures are built · BLS indexes through September 2026
Over the latest 12 months published, inflation in the 23 metro areas ranged from +1.3 % in Houston to +5.2 % in Urban Hawaii, a spread of 4 percentage points around the national 3.7 %. Four things explain most of such gaps. Housing comes first: shelter is about 36 % of the national basket, and local rents move with local markets, so shelter inflation runs from Houston (-0.1 %) to Urban Alaska (+4.9 %) today. Energy comes next, because gasoline and electricity prices, and the share of each in the budget, differ by region. Then the sample: each metro index rests on far fewer prices than the national one and, as the BLS warns, carries substantially more sampling error. Finally the calendar: 20 metros are priced every other month, so their latest 12-month figure can end a month earlier than the national one. Over 10 years, local differences shrink but do not vanish.
Housing: the local part of the basket
Most of what a household buys has a national or even global price: a phone, a car, an airline ticket, a box of cereal. Shelter does not. Rents and the rental equivalent of owned homes depend on the local market, and shelter is by far the heaviest item of the CPI-U, about 36 % of the national basket in the December 2025 weights. When rents in one metro rise twice as fast as in another, that difference flows almost directly into the all items index. Over the latest 12 months, shelter inflation was highest in Urban Alaska (+4.9 %), St. Louis (+4.7 %), New York (+4.5 %) and lowest in Atlanta (+0.9 %), Tampa (+0.2 %), Houston (-0.1 %).
Shelter also moves slowly. The BLS measures rent on six panels of housing units, each visited every six months, and the index follows the rents of all those units, not only the new leases. A change in market rents therefore shows up in the index gradually, over several releases. The rent inflation page follows shelter in each metro.
Energy: one price, many budgets
Gasoline, electricity and natural gas are not priced the same from one place to another, and the share of each in household budgets is not the same either. The local energy index follows both. In 2026, with national energy prices up 20.4 % over 12 months, energy inflation by metro ranged from Tampa (+9.4 %) to Washington, DC (+26.4 %).
Which groups moved most in your area?
Urban Hawaii, September 2025 to September 2026
+5.2 %
| Energy | +20.9 % |
| Recreation | +9.8 % |
| Other goods and services | +9.5 % |
| Transportation | +7.6 % |
The sample: noise in small numbers
The national CPI rests on prices collected from about 6,000 housing units and 22,000 retail establishments in 75 urban areas. A metro index uses only the share of those collected in that metro. In the words of the BLS, each local index has a much smaller sample size than the national or regional indexes and is subject to substantially more sampling and other measurement error. One consequence is that the rankings shuffle: of the five metros with the highest 12-month rate today, 1 was also among the top five a year earlier. Small groups such as apparel or medical care swing even more, and a change of several points in one release can be noise.
This is why the BLS urges contract writers to consider the national or regional index rather than a local one. It is also why a single month of a metro index should not be read as a turning point.
The calendar: different months, different news
Only New York, Los Angeles and Chicago are priced every month. Of the others, half are priced in odd months and half in even months. When the national index reports September, an even-month metro still reports August, and a price jump that happened in September appears in its index only one release later. Two metros can look far apart on a given day simply because their latest figures end in different months. The CPI by city table shows the month of each figure.
Over ten years, the gaps that remain
Noise cancels out over long periods; real differences do not. Since the 2015 annual average, among the metros whose index goes back that far, prices rose most in Seattle (+52.8 %), San Diego (+51.8 %), Miami (+50.8 %) and least in Washington, DC (+35.1 %), Houston (+32.5 %), Urban Alaska (+30.4 %). Those gaps reflect housing markets above all, and they are the ones that matter for a long contract or a pension paid in one place.
What the differences do not mean
A metro where prices rose fast is not necessarily an expensive one, and a slow one is not cheap. The CPI measures change from a base period, not the level of prices: the BLS says plainly that an area index does not show whether prices or living costs are higher or lower there than in another area. A city can be expensive and stable, or affordable and rising fast. To compare what two cities cost today, you need price level statistics such as regional price parities; to compare how fast they changed, the inflation calculator is the right tool.