What “accurate” means
We optimize for paycheck withholding (what a typical employer remits each pay period), not a full year-end tax return. That is the same goal ADP and PaycheckCity free tools target.
United States engine (2026)
- Federal income tax: IRS Publication 15-T percentage method, including Form W-4 Step 2 (multiple jobs), Step 3 (dependents credit), and Step 4 (other income, deductions, extra withholding).
- Bonus / supplemental wages: optional IRS flat supplemental rate (22%, or 37% above the annual high threshold).
- FICA: Social Security 6.2% up to the SSA wage base; Medicare 1.45% plus Additional Medicare where applicable. Traditional 401(k) reduces FIT only; Section 125-style benefits reduce FIT and FICA.
- State income tax: 2026 rate tables for all 50 states, cross-checked against Tax Foundation’s Jan 1, 2026 compilation and corrected where mid-2026 legislation is retroactive (e.g. Georgia HB 463 → 4.99%; Utah SB 60 → 4.45%; South Carolina H.4216 → 1.99%/5.21% + SCIAD). Standard deductions included where the state publishes them. California uses FTB Schedule X/Y/Z, the California standard deduction, and Form 540 personal exemption credits with the AGI Limitation Worksheet phase-out. New York uses NY brackets plus the NY standard deduction; NYC resident tax uses the progressive IT-201 city schedule when a NYC ZIP is entered. Personal exemption credits outside California are not fully modeled (rates ± standard deduction only).
- California SDI: EDD employee rate (1.3% for 2026, no wage cap). We estimate Form 540–style tax after credits, not the full EDD DE 4 wage-bracket withholding worksheets.
- Local tax: ZIP lookup is visible in the main calculator. New York pages preload ZIP 10001 for NYC resident tax. Maryland: county/city local income tax is mandatory for every resident (Comptroller 2026 rates 2.25%–3.30%); enter a MD ZIP for your county or we default to 3.20%. Not fully modeled (enter custom %): Pennsylvania EIT across ~2,500 municipalities/school districts (Philly / Pittsburgh samples only); Ohio municipal / RITA (three city samples); Kentucky county occupational taxes (Louisville sample only). Indiana county locals are not mapped.
International engines
UK (Income Tax + NI), Canada (federal + provincial + CPP/EI), Australia, and Tier-1 European countries use country-specific simplified official tables. Expand advanced W-4-style detail is US-first because that is where ADP/PaycheckCity compete hardest.
Sources we follow
- IRS Publication 15-T — Federal income tax withholding (Percentage Method)
- SSA Contribution and Benefit Base — Social Security wage base for 2026
- IRS Topic 751 — Medicare tax rates and Additional Medicare Tax
- State Departments of Revenue withholding guides — e.g. California FTB, New York DTF, Maryland Comptroller
- Tax Foundation — 2026 State Income Tax Rates and Brackets (compiled from state statutes/forms as of Jan 1, 2026)
- Published city/county earned-income or local income tax rates (NYC, Maryland counties, Philadelphia, Ohio municipals, etc.)
- HMRC (UK) / CRA (Canada) / national revenue tables for other countries
Keeping rates and page copy in sync
Worked examples and many FAQ take-home figures are generated from the same calculator engine at build time, so a tax-table fix updates those numbers automatically. Static rate claims in editorial notes (for example prior-year “down from” figures) are guarded by an automated content-drift test in CI — when we change STATE_TAX or related constants, that test fails if prose still cites a stale percentage or deduction. After mid-year legislation, we update the engine first, then re-run npm test before publishing.
Why results can still differ from your payslip
- Employer-specific benefits, garnishments, or rounding
- State withholding formulas (e.g. CA DE 4 tables) that differ slightly from Form 540 liability after exemption credits
- Local taxes not in our ZIP map (enter a custom %)
- Mid-year law changes before our annual update
Use the calculator with advanced W-4 and ZIP fields for the closest estimate. This is not tax advice.