Methodology

Last updated 2026-10-22. Applies to BEA data for 2008–2024 and Zillow data through September 2026.

What the price levels measure

The core numbers are the U.S. Bureau of Economic Analysis's Regional Price Parities (RPPs). An RPP compares the price of the same basket of goods and services across places in the same year. The US average is 100. An RPP of 110 means prices are 10% higher than the national average; 90 means 10% lower.

BEA publishes an overall figure ("all items") and four parts:

• Housing: rents paid by tenants and the rental value of owner-occupied homes. BEA prices both with local rents, comparing similar homes.
• Goods: physical products, including food, clothing, vehicles and fuel, medical goods and recreation goods.
• Utilities: household electricity, gas, water and sewer.
• Other services: all other services, including medical care, education, eating out, transportation and recreation services.

RPPs are price levels within a year. They are not inflation rates. When a place's index rises from one year to the next, it means prices there rose faster than in the country as a whole.

Official, derived and estimated

Every figure on the site carries one of three tags.

• official Published as is by BEA or Zillow. We only round it.
• derived Calculated by us from official figures with a formula on this page (for example, the difference between two cities, a rank or a five-year change).
• estimated Calculated by us using an assumption that the source does not publish (only the housing-share adjustment below).

We do not show prices of individual items such as a meal or a haircut. No free official source publishes them for every metro area.

Comparing two places and salary equivalence

For places A and B:

• Relative cost: RPP_B ÷ RPP_A − 1, for the overall index and for each part.
• Equivalent salary: salary in A × RPP_B ÷ RPP_A, rounded to the nearest $100.

This keeps the same basket of goods and services affordable in both places. It uses an average household's spending and ignores taxes, which can matter a lot between states. It is a before-tax comparison.

The housing-share slider

The default comparison uses BEA's overall index, where housing is one part of an average budget. People who spend more of their budget on housing, such as renters in expensive cities, feel housing differences more. The slider lets you set your own housing share h.

BEA combines the parts using each area's own spending pattern and does not publish a single national split. We estimate the average housing share w by fitting BEA's overall index as a weighted sum of its four parts across all metro areas (least squares, weights summing to 1). For 2024 this gives w = 18.3%. The estimate is stable from year to year: 2015: 16.8%, 2016: 17.2%, 2017: 16.6%, 2018: 16.2%, 2019: 16.9%, 2020: 17.9%, 2021: 17%, 2022: 17.1%, 2023: 17.5%, 2024: 18.3%.

The same fit cannot separate goods, utilities and other services reliably (they move together), so we do not use those weights. Instead:

• Everything except housing: NH = (RPP_all − w × RPP_housing) ÷ (1 − w)
• Your index: h × RPP_housing + (1 − h) × NH
• Equivalent salary: salary × index_B ÷ index_A

When h equals w, this gives exactly BEA's overall result. At any other setting the result is estimated. It is least exact in the most expensive housing markets, where housing takes a bigger share of local spending than the national average. The same "everything except housing" figure is used in some of the written findings, which say when they rely on it.

History charts

History charts show each year's RPP from 2008 to 2024, as revised by BEA in its latest release. BEA does not publish figures for a few areas in early years (Enid, OK before 2013; Poughkeepsie-Newburgh, NY before 2016; Twin Falls, ID before 2015); those years are left blank. Single-year jumps in small metro areas can partly reflect sampling noise in the Census survey BEA uses for rents.

Rankings and similar places

• Rank 1 is the most expensive (or highest value). Ties share a rank.
• "Cheaper than X% of metros" is the share of the other metro areas with a higher index.
• Similar metro areas: we put each metro's four part indexes on a common scale (subtracting the average and dividing by the standard deviation across metros) and list the metro areas with the smallest combined distance. This finds places with the same mix of costs, not just the same overall number.

Rent and home values

• Typical asking rent: Zillow Observed Rent Index (ZORI), all homes and apartments, smoothed. It measures asking rents for new leases, which change faster than what existing tenants pay.
• Typical home value: Zillow Home Value Index (ZHVI), homes in the middle third of values, smoothed and seasonally adjusted.
• One-year and five-year changes compare the latest month with the same month one and five years earlier. "From peak" compares the latest month with the highest month since the series began.
• Price-to-rent ratio: typical home value ÷ (12 × typical monthly rent) derived. It is a rough rent-or-buy signal: it ignores mortgage rates, taxes, insurance and maintenance, and the two indexes cover somewhat different homes.

Gas prices

• Source: the U.S. Energy Information Administration's weekly retail gasoline prices (regular grade, all formulations), in dollars per gallon including taxes. EIA surveys a sample of stations every Monday. It publishes the US average, 9 regions, 9 states and 10 cities; other states are shown through their EIA region, and the page says so.
• The site re-reads EIA's latest-weeks table a few times a day so new weeks appear without waiting for a site update. New values are only accepted if the weeks already on the site match EIA's table exactly and the new prices are plausible; otherwise the last checked data is used.
• Weekly changes compare with the previous week; yearly and five-year changes compare with the same week 52 and 260 weeks earlier. derived
• Monthly and yearly figures are simple averages of the weekly prices in that month or year; yearly comparisons use full years only (50 weeks or more). derived
• Record highs are the highest weekly average in EIA's series, in dollars of the time (not adjusted for inflation).
• Gas cost calculator: gallons = miles ÷ mpg; cost = gallons × price. A commute is there and back each day; a month is 52 ÷ 12 weeks.
• These are area averages for one week, not the price at a particular station today.

Matching Zillow areas to BEA metro areas

BEA uses the federal government's 2023 metro area definitions and identifies areas by code. Zillow names its areas ("Austin, TX") and still uses older boundaries for some. We match on the main city and state, then fix mismatches by hand. 4 metro areas have no comparable Zillow area, and we show them without rent and home values rather than borrowing a neighbor's figures.

How the written findings are produced

The short findings on each page are generated from that page's numbers by fixed rules: which part of the budget is furthest from the US average, how much each part contributes to the overall gap, how the index and rank moved since 2019, how rents and home values changed against the US, and which places are most similar. A finding is only included when the data supports it, and they are ordered by how notable they are. We check automatically that no sentence repeats across pages. We do not claim causes the data cannot show.

Which pages search engines see

Every metro area has a page, but at launch only the 50 largest metro areas (plus 3 named in our launch plan), all states and the ranking pages are offered to search engines. A page must have complete price data, at least ten years of history, rent or home-value data, and at least four findings. Comparison pages are offered only for a set list of pairs (all pairs among the 15 largest metro areas plus common moves). Every other comparison still works but is marked noindex. We will widen this as the site proves useful.

Limits

• Metro areas only. Neighborhoods inside a metro area can differ by more than two metro areas do.
• An average household's basket. Your own costs depend on household size, housing, commute and habits.
• No taxes. State and local income, sales and property taxes are not included.
• Timing. BEA's latest year is 2024, published in early 2026. Zillow data are monthly.
• No childcare yet. The official source is county-level and ends in 2022; we will add it only when we can show it without guessing.

Updates

BEA releases new RPPs once a year, usually in the winter; Zillow updates monthly. Each update re-downloads the source files, checks them against known published values, and rebuilds every page. Corrections are listed on the corrections page.