Methodology
What's actually left — and what it's really worth.
For a chosen occupation and place, we estimate your gross pay, subtract everything that leaves a paycheck, then subtract the real costs of a life. But a single “leftover” number hides the truth, so we go further — separating the money that looks spare from the money that's actually durable.
One correction worth stating: saving isn't a cost. A dollar into a 401(k) or the principal portion of a mortgage is wealth you keep, not money spent — so we count it as kept and split the headline into liquid cash vs. wealth being built. A diligent saver shouldn't rank worse than someone who saves nothing.
Pay
Wages come from the U.S. Bureau of Labor Statistics Occupational Employment and Wage Statistics (OEWS) — 830 detailed occupations across the nation, all 50 states + DC, and ~390 metros. Because the wages are a couple of years older than current rents and taxes, we bring pay forward to current dollars (BLS Employment Cost Index) so numerator and denominator share a vintage. Experience level maps to the 25th percentile (entry), median (mid), and 75th percentile (senior) — read that as the spread of pay in that place today, a cross-section, not a forecast of your raises. A second income is added to the household and taxed jointly.
Everything that leaves the paycheck
Federal income tax uses current IRS brackets and the standard deduction for your filing status. Payroll covers Social Security, Medicare (plus the surtax), and — where they exist — state disability / paid-leave deductions (e.g. California SDI). State income tax uses each state's real progressive brackets, and we add local income taxes where they're widespread (Ohio, Maryland, Pennsylvania…) plus big-city rates for NYC and Philadelphia. We also subtract pre-tax withholding — 401(k), HSA, FSA — which both shrinks your taxable income and reduces the cash that actually reaches your account. That last part is the piece most calculators miss.
The cost of a life — measured, not guessed
Every cost is a toggle, computed from your household (size, children, who's in paid care, car, renter/owner, commute, coverage) and the place. The location facts are measured directly and carry a state data chip: rent — real observed market rent (Zillow ZORI), not the below-market HUD voucher ceiling, electricity & gas (EIA), health (premium share + typical out-of-pocket), and taxes. Health is more than a premium — we add deductibles, copays, prescriptions and dental; on the Marketplace we apply the ACA premium tax credit by income. Car insurance starts from the state average (NAIC) but then personalizes it for the things that actually move a quote — driver age, gender (where legal to rate on), and annual mileage from your commute — and is fully editable, because real premiums swing more than 2× between drivers. Beyond the obvious bills we model the leaks calculators ignore: a replacement reserve (the annualized cost of re-buying tires, brakes, phones, mattresses), subscriptions, a social tax(gifts, weddings, holidays), and a convenience premium (delivery, rideshare) that rises with a punishing commute. Where a number is a national estimate rather than local data, the chip says so — and you can type your own figure over any modeled line; it recomputes everything live.
Three kinds of “left over”
A calculator that says you have $1,400 spare is lying by omission. We subtract in layers. Free cash flow is what's left after necessities — the number most tools stop at. Durable cash flow is what survives once you also fund a replacement reserve and a modest emergency set-aside (toward a 3-month buffer). Lifestyle cash flow is what's left after actually living — subscriptions, going out, the social calendar. The gap between the first and the last is where the disconnect between a spreadsheet and a real life lives. We also flag when a household is “acting broke”: technically positive, but a single ~$1,500 repair would erase more than a month of free cash flow.
Beyond the dollars
Money isn't the whole story. Resilience scores how exposed you are when life goes wrong — runway if you're laid off, whether a 10% rent hike pushes you under, single vs. dual income, and a haircut for volatile pay (a volatile $120k behaves like a stable $95k). Mobility scores room to move up — pay headroom (the spread from entry to senior), employer density and industry concentration from BLS QCEW (how many places you could jump to). And time costnets your commute and the unpaid-but-mandatory hours (a lunch break too short to use) out of your effective wage — the “$43/hour” job that's really $36 once you count every hour it takes.
Finding your career
The OEWS list knows occupations by bureaucratic titles (“Customer Service Representatives”), not the words people use. So search also spans ~54,000 O*NET alternate titles — type “personal banker” or “UX designer” and it resolves to the right occupation, showing the title that matched. When an exact phrase misses, we fall back to ranking by how many of your words appear, so improvised job titles still land somewhere sensible.
How we check ourselves
A tall stack of modeled assumptions is only worth trusting if the total matches reality, so we benchmark the engine against the MIT Living Wage Calculator. For a single adult in California our housing lands within ~1%, food within ~5%. That check also caught two biases we then fixed: transportation was overstated by assuming everyone carries a new-car loan (corrected to a blended payment, since many own outright), and medical was understated by counting only premiums (we added out-of-pocket). Where we still sit above a benchmark it's usually deliberate — our rent is market-rate where theirs uses below-market HUD figures, and our taxes include the full payroll bite.
What this is and isn't
This is a comparison tool built from official public data, designed to make the real, after-everything cost of a career legible — not a personal budget or financial advice. Modeled lifestyle figures are national estimates you can override with your own numbers. Treat the ranking as a strong directional signal, not a quote.