Common mistakes when comparing cities with average COL numbers

“Denver is 15% more expensive than the US average” is a sentence about a basket, not about you. The basket's housing weight is not your bedroom count. Its income assumption is not your job. It has no state tax. If you sort a spreadsheet of those percentages you will pick a shortlist that is internally consistent and still wrong for a nurse with two kids, or a remote designer, or anyone who would not rent the average unit.

Leftover replaces the percentage with a dollar figure for one household. Leftover vs a cost-of-living index is the longer argument. The short version: sort /results, not a COL column.

People remember “Texas has no income tax” and forget Philadelphia's city wage tax, Illinois's flat rate, and New York City's local income tax. They also forget that the raise from dropping a state tax is monthly, and so is the rent gap. A 4% state tax on a $100,000 salary is about $300 a month after a standard deduction — less than the difference between a one-bedroom in Chicago and a one-bedroom in San Diego.

Run the compare with both tax codes, not with a mental adjustment. Philadelphia vs Pittsburgh exists because the local wage tax is the story. The tax and rent tradeoffs guide walks the no-income-tax states that still lose leftover to rent or property tax.

Census ACS median rent describes the whole occupied stock, including leases signed eight years ago and rent-stabilised units. Asking rent for a new lease is what you will pay. In slow-turnover markets the gap is huge — the methodology cites New York at about 41% below asking. In fast-building Sun Belt metros the two numbers sit closer. If you use ACS medians to rank cities, the expensive, sticky markets look affordable and the places with a lot of new apartments look worse than they are.

Uprootmap uses Zillow asking rent for that reason, with HUD Fair Market Rent as a floor. Set your bedroom count. Then go look at listings. If they disagree with the page, believe the listings and subtract the difference from leftover.

A $180,000 software salary in San Jose does not follow you to Omaha unless your employer said it does. Local-market pay in the model scales the job by the metro's wage index. Remote holds the number you typed. Mixing the two — keeping the coastal salary in your head while ranking inland cities on “what jobs pay there” — produces a fantasy leftover.

If the offer letter says the salary is location-adjusted, use local pay. If it says the salary is unchanged, use remote. If you do not have an offer yet, run both and do not average them. The checklist puts this switch first on purpose.

Housing and childcare dominate leftover once kids exist. A three-bedroom plus one full-time daycare slot can erase a no-income-tax raise. Ranking metros “for us” while the form still says one adult and one bedroom is how people end up in love with a city whose family leftover is negative.

Set adults, kids, kids under five, and bedrooms to the household that will actually move. Then open the kids topic on a city page — Raleigh schools is the pattern — for enrollment and staffing, not for a quality score. Quality is off-site work.

New York leftover looks better for a single renter than a COL index expects because the model can drop a car. Dallas leftover looks worse than the cheap rent suggests because it cannot. If you compare those cities on rent and groceries alone, New York loses a race it sometimes wins on leftover, and Dallas wins a race it sometimes loses.

The daily-life tile on each city page says whether a household can skip a car. Read it before you treat transport as a wash. Two $500 car payments are $1,000 a month, which is a state tax bracket.

Rents move. Tax schedules move once a year. A tweeted COL table from 2022 will still rank Austin as a bargain and will not know what happened to asking rent. This site's rents and home values show a refresh date on the about and methodology pages. If leftover is a close call, check that date, then check current listings.

Also do not trust a leftover number you computed last month with a different household still in the URL. Look at the address. If it still has someone else's salary, that is someone else's city ranking.

Leftover is a filter, not a decision. It will not tell you whether you will like January in Minneapolis or July in Phoenix. It will not license you to teach in that state or find a hospital that is hiring. Use it to throw out the metros that do not clear the month, then do the rest of the work with listings, licensing boards and people who already live there.

IRS “who moved” pages — who moved to Austin, San Francisco → Austin — are a useful sanity check. If households in your income band are already making the trip, you are not inventing the market. If they are not, leftover may be missing a constraint (licensing, industry concentration, housing supply) that the model does not see.

Read a city page with this walkthrough, pick four metros with the checklist, and treat every number as an estimate. The disclaimer and terms say the same thing in plainer legal language: this is a model, not a quote.