U.S. Federal Tax Brackets

How federal tax brackets actually work

Federal tax brackets are marginal, not flat. If you earn $100,000 as a single filer in 2026, you don't pay 22% on all of it. You pay 10% on the first $11,925, then 12% on the next portion, and so on. Only the last dollar earned is taxed at your "tax bracket" rate.

This is why the difference between the headline marginal rate and the effective rate (total tax ÷ total income) is large. In 1960, the top marginal rate was 91%, but the effective rate paid by the top 1% was about 42%. Today, a 22%-bracket filer typically has an effective rate closer to 13-15%.

The number of brackets and where they kick in moves with legislation. The Tax Reform Act of 1986 collapsed the bracket count from 14 to 2. The Bush cuts (2001) re-expanded it to 6. The Tax Cuts and Jobs Act (2017) re-set it to 7 - the current structure. See the full history chart →

Marginal vs effective rate: the math most people get wrong

Take a 2026 single filer earning $200,000 of taxable income. The headline marginal rate is 32% - that's the bracket that contains the last dollar earned, which kicks in at $197,300 for single filers. But that filer doesn't pay 32% of $200,000 ($64,000). They pay 10% on the first $11,925, 12% on income from $11,925 to $48,475, 22% from $48,475 to $103,350, 24% from $103,350 to $197,300, and 32% only on the $2,700 between $197,300 and $200,000. Add it up and the total federal income tax is about $40,200 - an effective rate of 20.1%, twelve points below the headline marginal rate.

The marginal rate matters for incremental-dollar decisions: should I take the $30,000 bonus, exercise the NSOs, do the Roth conversion? Each marginal dollar is taxed at the bracket that dollar falls into, so the 32% comparison is what counts. The effective rate matters for budgeting: how much federal income tax will I owe in total this year? It's also the number worth quoting when comparing total tax burdens between filers, states, or eras. The pre-1986 90% top rate sounds dramatic, but the effective rate the top 1% paid was about 42% - because so few dollars actually reached that 90% bracket. Confusing the two numbers makes nearly every "is this worth it" tax decision go sideways.

Why bracket boundaries inflate every year (and why 2023 was unusual)

Since 1985, the IRS has indexed bracket thresholds to inflation each fall, publishing the next year's schedule via Revenue Procedure in October or November. The mechanism uses the Chained Consumer Price Index for All Urban Consumers (C-CPI-U), which since 2018 has replaced the prior CPI-U as the indexing basis under the Tax Cuts and Jobs Act. The chained version captures consumer substitution behavior and tends to grow about 0.25 percentage points slower per year than the unchained CPI-U, which means bracket boundaries (and the standard deduction) creep up slightly slower than they otherwise would. Over a decade, that gap moves real-dollar tax burdens noticeably higher even when nominal rates don't change.

2023 was the outlier. The 7.1% inflation adjustment was the largest single-year bracket move in TCJA history, lifting every threshold by roughly that amount and pushing the standard deduction up by $900 (single) and $1,800 (married filing jointly). That's why a household earning the same nominal income in 2023 as 2022 paid noticeably less federal tax - a quiet inflation-adjustment windfall that didn't get the press the headline inflation rate did. The 2024 and 2025 adjustments returned to the normal 2-3% range. Without bracket indexing, every inflationary year would silently raise effective rates as nominal-dollar earners drift into higher brackets, a phenomenon economists call "bracket creep." Indexing prevents that - but only at the chained-CPI rate.

Filing status changes the picture more than most filers expect

The four filing statuses (single, married filing jointly, head of household, and married filing separately) use entirely different bracket schedules. A married couple filing jointly has roughly double the bracket widths of a single filer for the lower brackets, which means two earners can pool income before hitting higher rates. But the doubling is imperfect at the top: the 37% bracket starts at $626,350 for a single filer and $751,600 for MFJ, not $1.25 million. This is the marriage penalty at the highest brackets; two single filers each earning $400,000 owe less combined federal income tax than the same household filing jointly.

Head of household sits between single and MFJ in most brackets, with a standard deduction of $24,000 for 2026 versus $16,000 single. Qualifying requires being unmarried, paying more than half the cost of keeping up a home, and having a qualifying child or dependent for more than half the year. Married filing separately uses brackets exactly half the width of MFJ, so it almost always produces higher total tax than MFJ unless one spouse has unusually large medical expenses, miscellaneous deductions hitting AGI floors, or income-driven student loan repayment plans where filing separately materially reduces the payment. The choice isn't usually a strategic optimization (it's locked in by your situation on December 31), but knowing which brackets you'll use shapes every year-end decision from RSU exercises to IRA conversions.

What "tax bracket arbitrage" looks like in practice

High earners with control over the timing of income can move dollars between tax years to land them in lower brackets. The classic case is the consultant or freelancer choosing whether to invoice in December or January - December income lands in the current year's brackets, January in the next year's. If next year will bring a sabbatical, parental leave, or a year between jobs, deferring January income at the 32% bracket into a year where it'd land at 22% is a 10-percentage-point real saving. Same logic in reverse: pulling a Roth conversion or NSO exercise forward into a low-income year to fill the lower brackets without bumping into the next one.

Retirement adds another layer. Once W-2 income stops, brackets refill with whatever you draw from accounts. A retiree with a $1.5M Traditional IRA and $200K of taxable-account capital gains has a five-to-ten-year window in their early 60s to do partial Roth conversions in the 12% or 22% brackets, before required minimum distributions and Social Security force their income upward into higher brackets. Missing that window is one of the most common and most expensive tax mistakes in retirement planning. The Roth vs Traditional calculator runs the apples-to-apples comparison with after-tax balances; the retirement calculator stress-tests whether the drawdown plan survives sequence-of-returns risk.