Methodology & Data Sources

MoveMath estimates what a household’s life costs in each US metropolitan area after taxes, and the salary needed to keep the same budget after a move. This page documents every dataset, its vintage, how the engine uses it, and what it leaves out. The engine is open in its logic: every line item in every budget names its source and vintage, and the same TypeScript code produces the server-rendered pages and the interactive tools.

Geography

Places are the 387 Metropolitan Statistical Areas in the 50 states and DC under OMB’s July 2023 delineation (OMB delineation file), with Census Vintage 2025 population estimates and Census Gazetteer coordinates. The six Puerto Rico metros are excluded because our income-tax data does not cover Puerto Rico. Each metro has a principal city (first city in the OMB title) whose state sets income tax, and a principal county: the county holding the largest share of that city’s population in the Census Vintage 2025 place-within-county estimates (Census subcounty estimates, principal cities from OMB List 2). The principal county sets the HUD rent area, childcare prices and county-based local taxes — for example District of Columbia, DC for Washington, DC-VA-MD-WV, St. Louis city for St. Louis and Baltimore City for Baltimore. Where a city spans several counties (New York’s five boroughs, Kansas City, Oklahoma City) the county with the largest share is used. Metro names list states in OMB title order (“New York, NY-NJ”). Rural (non-metro) areas are not covered.

Income taxes (2026)

Taxes come from the verified 2026 tax engine and data MoveMath shares with our sister site TakeHome: federal brackets, standard deduction and Child Tax Credit; Social Security (per earner, up to the wage base) and Medicare including the Additional Medicare Tax; each state’s brackets, deductions, exemptions, credits and recapture rules; state employee payroll taxes (disability, paid leave, unemployment where employees pay); and mandatory local income taxes. Federal figures cite 15 IRS and statutory sources; state files cite 161 state revenue-department sources.

• Annual liability, not withholding. We compute the tax owed on a full year of wages with the standard deduction. No itemized deductions, no EITC or other credits besides the Child Tax Credit, no pre-tax retirement or health deductions.
• Filing status. Two adults file jointly; one adult files single or, with children, head of household if selected. All children count as Child Tax Credit dependents. The refundable part of the credit that may come at filing is not counted as income.
• Local tax. We apply the mandatory local income tax for a resident of the principal city — or, in Maryland and Indiana, the principal city’s county (Baltimore City for Baltimore; the St. Louis 1% earnings tax for St. Louis). The District of Columbia has no separate local tax: its income tax is applied as the “state” tax for Washington, DC. In our data 32 metros have a computable local tax (for example New York, NY-NJ; Philadelphia, PA-NJ-DE-MD; Detroit, MI; Denver, CO; Baltimore, MD). For 40 more, a local tax exists but its rate is not in our verified data (Pennsylvania earned-income taxes, many Ohio and Michigan cities, Cleveland); those budgets carry a note and exclude it rather than guess.

Housing

Rent: HUD Fair Market Rents (FY2027)

Source: FY 2027 Fair Market Rents — County Level Data; Federal Register 91 FR 56156, effective 2026-10-01. An FMR is the 40th-percentile gross rent (rent plus tenant-paid utilities, excluding phone, cable and internet) for standard-quality units. We use the FMR of the metro’s principal county’s FMR area and label it as such. Because FMRs — and the alternative Census median gross rent — include utilities, renters get no separate utility line.

Bedrooms default to one bedroom for the adult(s) plus one bedroom per two children, capped at four. This follows the two-persons-per-bedroom guideline common in HUD voucher subsidy standards (24 CFR 982.402 requires “the smallest number of bedrooms needed to house a family without overcrowding”), with adults in their own room. Users can override it.

Census ACS (2020-2024)

Source: Census ACS 5-Year Summary File 2024 (table-based, 5-year estimates 2020-2024). For each metro and county: median household income (B19013), median gross rent (B25064), median home value (B25077) and median real estate taxes paid (B25103). The effective property-tax rate = median real estate taxes ÷ median home value; it is applied to the home price the user enters. Values are in 2024 inflation-adjusted dollars.

Top-coded medians. The Census publishes median real estate taxes only up to “$10,000 or more” (and “under $200” at the bottom). That cap is reached by the New York metro and 17 counties (Bergen County, NJ; Essex County, NJ; Hunterdon County, NJ; Monmouth County, NJ; Morris County, NJ…). We show those medians as “$10,000+” and never treat the cap as an estimate. For a metro whose median is top-coded, owner budgets use the effective rate of the metro’s principal county when that county’s median is not top-coded — for New York, Kings County, NY (principal county) at 0.71%, the county already used for its rent, childcare and city tax — and the metro page shows the metro-wide rate only as a lower bound (at least 1.63%). Where no uncoded figure exists (top-coded counties in the cost-of-owning tool), the rate and property tax are shown as “at least” $10,000 ÷ median value and the county is left out of rankings. We do not substitute a state-level rate: our ACS file has no state medians, and a state figure would mix very different places. Renter budgets never use property tax.

Owning

Monthly principal and interest use the standard amortization formula on the price, down payment, rate and term entered (defaults 20% down, 6.5% for 30 years are example inputs, not market data). Homeowners insurance is a user input because no official dataset of premiums by place is in our sources; the $1,800 default is a placeholder. PMI, HOA dues, maintenance and closing costs are not modeled. Owners get an electricity line (see Energy).

Childcare

Source: U.S. Department of Labor Women’s Bureau, National Database of Childcare Prices (data file), 2022, the latest year published. County median weekly prices for full-time care by age band — infant (0–23 months), toddler (24–35), preschool (36–54) and school-age — in centers and home-based family childcare. Annual cost = weekly price × weeks (default 52) × children in each band. By default prices are expressed in current dollars by multiplying by the change in the BLS CPI-U for day care and preschool (series CUUR0000SEEB03) from the data year’s average to the latest month; users can turn this off. The NDCP flags describe how DOL collapsed six-month age bands into each group (all bands matched, the most common band price, or the highest of several modes); they are not sample-size or pooling indicators.

• Older years. 25 metros’ principal counties have no 2022 price (Philadelphia, PA-NJ-DE-MD; St. Louis, MO-IL; Pittsburgh, PA; Kansas City, MO-KS…). For them we use DOL’s latest published year for that county (2020 and 2021), label every figure with that year, and adjust it with the BLS factor from that year’s average (for example × 1.2945 from 2020).
• No data. 21 metros’ principal counties have no NDCP price in any year from 2018 to 2022 — 12 in Indiana, 4 in New Mexico and 5 in Missouri (Indianapolis, IN; Albuquerque, NM; Fort Wayne, IN…). Their budgets show childcare as “not included” with a note, and comparisons involving them warn that the salary needed is biased by the missing cost. We do not borrow a neighboring state’s price.
• Checked outliers. Riverside County, CA’s center infant price ($404.68 a week in 2022) is far above Los Angeles County’s ($256.98). We checked both rows in DOL’s raw file: both are the published figures (flag 1), Riverside is in line with San Bernardino, Orange and San Diego counties, and Los Angeles is the low outlier. Both are used as published.

Food

Source: USDA Food Plans: Monthly Cost of Food Reports, August 2026 report. Users pick the Thrifty, Low-Cost, Moderate-Cost or Liberal plan. Members are priced as: each adult = average of the male and female adult rows (20–50 years in the Thrifty plan; 19–50 years in the Low-Cost, Moderate-Cost and Liberal plans, which USDA publishes with 14–18 and 19–50 age groups); infant = child 1 year; toddler = child 2–3; preschool = child 4–5; school-age = average of child 6–8 and 9–11. USDA’s household-size adjustments are applied (1 person +20%, 2 persons +10%, 3 persons +5%, 4 persons +0%, 5 persons −5%, 6 persons −5%, 7+ persons −10%). USDA plans are US averages, so we multiply by the metro’s BEA price parity for goods ÷ 100. This is our method, not USDA’s; Alaska and Hawaii use the same approach because USDA publishes only Thrifty reference-family costs for them.

The Thrifty plan is not always the cheapest: for a child 4–5 it costs $188.30 a month versus $178.80 on the Low-Cost plan ($43.50 vs $41.30 a week). We checked this against USDA’s report: it is in the source, because the Thrifty Food Plan was re-evaluated in 2021 with new dietary guidelines and consumption data while the other three plans are still 2007 market baskets updated for inflation.

Energy and fuel

• Electricity (owners): the state’s 2025 average monthly residential bill (revenue ÷ customers ÷ 12, released 2026-10-07) from EIA final 2025 annual electric sales and revenue data (Form EIA-861), Annual State Historical Spreadsheet HS861 2010- (Total Electric Industry, Residential); with custom usage, kWh × the state’s 2025 average price from the same table. Renters get no electricity line because rents include utilities.
• Natural gas (owners, custom usage only): state residential price from EIA Natural Gas Annual 2024 — Table 23: Average price of natural gas delivered to consumers by state and sector × the Mcf a month you enter. Our sources do not include average gas use per home, so gas is not added by default — a known understatement for homes heated with gas.
• Gasoline: miles ÷ mpg × the EIA weekly retail price (EIA weekly prices): the state’s own series for the 9 states EIA publishes, otherwise the East Coast sub-region, “West Coast less California”, or the PADD region. Miles and mpg defaults (12,000 miles per car, 25 mpg) are example inputs. Other car costs are not modeled.

Sales tax

State-level general sales tax rates as of 2026-10-13 for 51 states and DC, each from the state revenue department (the source link for each state is in our data file). The rate applies to “other goods” and, where the state taxes groceries, to food at the state grocery rate. Local sales taxes are not modeled, so real sales tax is higher in most places. California, Utah and Virginia’s stated rates include their mandatory statewide local add-ons. Idaho taxes groceries at the full 6% state rate (offset in part by an income-tax Food Tax Credit we do not model). Legislated future changes are shown as notes only and not applied: District of Columbia 7% from 2027-10-01; South Dakota 4.5% from 2027-07-01.

Other goods and services (estimate)

Everything else a household buys is represented by two user-adjustable monthly baselines at US-average prices — other goods (default $500) and other services (default $500) — multiplied by the metro’s BEA Regional Price Parities for goods and for other services (BEA MSA RPPs, 2024; released 2026-02-19). The defaults are illustrative, not measured spending; the point is to show how local prices change the same basket. A “fixed costs” input adds amounts that do not vary by place (health premiums, car or loan payments).

Salary needed

For place A we compute the annual surplus: gross income − all taxes − all modeled costs. For place B we search for the gross income at which B’s surplus equals A’s, using bisection to within $1: try an income, run the full tax engine for B, subtract B’s costs, and halve the search range. Costs do not depend on income and no marginal tax rate reaches 100%, so B’s surplus rises with income and the answer is unique. Two-earner households keep A’s income split. If B’s surplus with no wage income already matches A’s, the answer is $0. Our test suite checks convergence within $10, monotonicity across tax brackets and wage caps, and that A → B → A returns the starting salary.

Moves and migration

Moving pages exist only for the 398 largest metro-to-metro flows in the Census Bureau’s Metro Area-to-Metro Area Migration Flows: 2016-2020 American Community Survey (ACS 5-year 2016-2020), matched to current metro codes — the latest metro-to-metro file the Census Bureau has published. State pages show state-to-state migration flows (2024 ACS 1-year). Each figure has a 90% margin of error.

Limitations

• Not modeled: health insurance and medical costs, car payments, insurance and maintenance, phone and internet, debt payments, moving costs, local sales taxes, and wage differences between places.
• Metro-level figures hide differences between neighborhoods and suburbs; childcare uses one county per metro.
• Data vintages differ: childcare prices are from 2022, Census figures cover 2020-2024, price parities are for 2024, rents are FY2027. Each is the latest official release.
• Results are planning estimates, not tax, legal or financial advice.

Found an error? See our editorial policy and contact us.