• Contributions land at the end of each period. Money added in March doesn't earn a full year of growth; it earns the remainder of the year. This is the conservative convention.
• The annual return is converted to a periodic rate. A 7% annual return becomes the bi-weekly rate that compounds to exactly 7% over a year, not 7% divided by 26. The difference compounds noticeably over decades.
• Contribution growth applies once a year, on the anniversary, not continuously.
• A partial final period is dropped rather than counted whole.
• Fees come off the return before anything else is calculated.
• Inflation adjustment divides by (1 + inflation) raised to the number of years, converting future dollars into what they'd buy today.
• The withdrawal figure is the first year of retirement. The model stops at your retirement date; it does not simulate spending the money down.

The Basic tab answers five questions and nothing more: your age, when you want to stop working, what you've saved, what you add, and roughly how it's invested. It's the right starting point if retirement planning is new to you.

What makes it simple is that it works in real terms. The investment choices are stated as growth after inflation, so every number you see is already in today's dollars, with no separate inflation setting and no mental translation of what $2 million means in forty years.

That simplicity costs you two things. It assumes you nudge your contribution up a little each year to keep pace with inflation; if you set $500 a month and never change it, real progress will be slower than shown. And it ignores tax and fees entirely, so the income figure is before both.

The investment choices map to these rates of growth after inflation:

Very conservative · 2.0%

Mostly cash and bonds. Steady, and barely ahead of rising prices.

Conservative · 3.0%

Bond heavy, with a slice of stocks.

Balanced · 4.5%

A mix of stocks and bonds. A common middle-of-the-road choice.

Growth · 5.75%

Mostly stocks. Bigger swings, more expected over long periods.

Aggressive · 7.0%

Nearly all stocks. Roughly what US stocks have returned after inflation over the very long run: an optimistic ceiling, not a floor.

The shaded band on the chart shows the same plan running 1.5 points better or worse, because no one earns an identical return every year. When you want tax, fees, volatility or changing assumptions, Open these numbers in Advanced carries your answers to the Advanced tab, splitting the real rate back into a return and 3% inflation (Advanced's own default), with contributions set to rise alongside it. Basic's own contribution already steps up once a year behind the scenes, the same way Advanced's does, so the inflation-adjusted total should land on the Basic figure almost exactly. Advanced does apply a 10% tax rate that Basic ignores, so the income figure will drop.

The Advanced tab holds one set of assumptions for the whole period, but goes further than Basic in almost every direction: return and inflation as two separate figures instead of one blended real rate, an effective tax rate, fees, volatility for Monte Carlo, and two routes to solve backwards from a target. It's also where the historical comparison band, the Monte Carlo fan, and a Coast FIRE date once you've set a target all live.

Starting value

What the account is worth today. Zero is fine.

Contribution & period

What you add each time, and how often. Match it to how you actually save, whether per paycheck or per month. The Convert frequency link under the period translates between schedules.

Contribution growth

How much the contribution amount itself increases each year, as a percentage. It is a rate in its own right, not something added on top of the inflation figure. Match it to your inflation rate to keep contributions flat in today's dollars; match it to your expected raises to keep them a steady share of your income. Typical: 0–5%.

Time period

Years until you start drawing on the money.

Rate of return

Expected average annual return before inflation. For context, broad US stock indices have averaged roughly 9–10% a year over the very long run, bonds far less, and a mixed portfolio somewhere between. Many planners use 6–8% for a stock-heavy portfolio to stay conservative. Lower it as you shift toward bonds. The ± button lets you enter a negative return to stress-test a downturn.

Glide path optional

Holds your rate of return steady, then blends it down in a straight line to an end rate over however many final years you choose: the common practice of shifting toward bonds as retirement nears. It shapes the projection, the chart, and both routes to a target.

Inflation

How fast prices rise. Long-run US inflation has averaged roughly 2–3%. The US Federal Reserve targets 2%. Most people use 2.5–3%.

Volatility Monte Carlo only

How much returns bounce around year to year. Broad stock funds have historically run near 15–18%; a balanced stock-and-bond portfolio nearer 8–12%; bond-heavy lower still. Higher volatility widens the fan without changing the average.

Fees optional

Expense ratios plus any advisory fee, subtracted from your return. Index funds often charge under 0.10%; actively managed funds frequently 0.5–1%; advisors commonly around 1%. Leave at 0 to ignore. The Milestones panel shows what fees cost you over the full period, which is usually larger than expected.

Withdrawal rate

The share of the portfolio you take in the first year of retirement. The widely cited starting point is 4%, from research suggesting that rate survived historical 30-year retirements. It is a rule of thumb, not a guarantee; some argue for 3–3.5% given today's conditions, others for more flexibility year to year.

Effective tax rate

The share of withdrawals lost to tax: your average rate, not your top bracket. It depends heavily on account type: withdrawals from a traditional 401(k) or IRA are generally taxed as income, Roth withdrawals generally are not, and taxable brokerage accounts are usually taxed at capital-gains rates. Many retirees land somewhere in the 10–15% range, but this varies enormously. Use 0 for an all-Roth plan.

Split by account type optional

Swaps the single starting value, contribution and tax rate for a balance and contribution in each of traditional, Roth and taxable brokerage accounts, plus an employer match and your filing status and state. Every account grows at the same return, so the projection is unchanged; what changes is the tax. The first year's withdrawal is taken from each account in proportion to its balance and taxed with the 2026 rules the Income Tax tool uses: traditional as ordinary income, only the growth in the brokerage account at capital-gain rates, and Roth not at all. Contributions aren't capped at the annual limits, since backdoor and mega backdoor Roth strategies can go past them; enter what you actually put in. Switching back to one total carries the totals and the calculated rate across, so the answer doesn't jump.

Target

What you're aiming for, in today's dollars, either an after-tax income or a portfolio balance, your choice. It drives the two solve routes, the Coast FIRE figure, and the Monte Carlo success rate. On the Stages tab the solve changes the final stage only; see that section.

Once you've got a projection, here's what the results panel is showing:

Future value

The balance at the end, in the dollars of that future year.

Inflation adjusted

The same balance expressed in today's spending power. This is the more meaningful of the two, and it is the smaller number at any inflation rate above zero.

After-tax income, per year

Withdrawal rate applied to the inflation-adjusted balance, less tax. Roughly what the portfolio could pay you in its first year of retirement, in money you can compare to your salary today.

Amount invested vs. growth

What you contributed versus what the market added. Over long periods growth typically overtakes contributions by a wide margin.

Crossover

The year your portfolio's growth first exceeds what you put in that year. After it, the account is doing more of the work than you are.

Work backwards from a target

Two routes to the same goal: raise the contribution, or extend the timeline. Either button applies its change to your inputs. On the Stages tab both routes act on the final stage. Both follow a glide if you have one set, and the timeline route re-anchors it: land on 26 years with a 5-year glide and the ramp runs years 22 through 26, not 26 through 30.

Coast FIRE

How far you are from the point where contributions could stop entirely and growth alone would still reach your target on schedule.

Stages chains several periods together instead of holding one set of assumptions for the whole timeline. Use it for a raise partway through your career, a stretch of lighter contributions, or shifting toward a more conservative mix as retirement nears, anything that changes partway through the plan. Each stage has its own length, contribution, growth rate and return, and builds on whatever balance the stage before it left behind.

Every input and result works the same way it does on the Advanced tab; see that section for what each one means. A few things are specific to Stages:

Inflation adjusted stage 2 onward

Restates the contribution you typed into that stage's future dollars. Type $3,000 for a stage starting in year 5 at 2.5% inflation and it's modeled as $3,394, the amount that feels like $3,000 by then. The stage's own growth rate compounds from there. Turn it off to use the number exactly as typed.

Glide path final stage only

Available on the last stage only, since that's the one closest to retirement; earlier stages don't get the option. Works the same way as on the Advanced tab: holds steady, then blends down to an end rate over however many final years you choose within that stage.

Target solving

The solve, and "Work backwards from a target" on the results side, change the final stage only. Everything before it is treated as settled, which is what makes the answer meaningful when stages carry different assumptions.

Split by account type optional

Works like the Advanced version, spread across stages. Starting balances, the employer match and your filing status and state are set once for the whole run; each stage card then says where that stage's contribution goes: a share to traditional, a share to Roth, and the rest to the taxable brokerage account. That lets a plan go Roth early in a career and traditional later, and the tax on your retirement income reflects the mix you end up with. The employer match works as on Advanced, from your salary, the match rate and the share of salary it applies to; your salary is assumed to grow at each stage's contribution growth rate.

The Guide tab walks through your finances one question at a time, in the order a planner would: income, take-home pay, spending, emergency fund, debt, housing, college, then retirement. When you don't know an answer, the step opens the tool on this site that finds it, filled in with what you've told the guide, and a panel docked over the tool lists what to do there and ticks items off as you go. Back to guide brings the result with you. Your answers are kept in this browser only, and the facts the household bar also holds (ages, state, income, savings, spending) are written through to it.

As you answer, a readiness score out of 100 builds from five areas: the retirement outlook (40 points), savings rate (20), emergency fund (15), debt (15) and monthly cash flow (10). It appears once two areas are answered, and ends in an ordered list of next moves, each with the tool that does it.

Retirement outlook

Your savings are projected to your retirement age the same way the Basic tab does it: a steady return after inflation for the mix you pick, with contributions (yours and your employer's) raised with inflation each year so they stay the same in today's dollars. That balance is then run through every historical retirement since 1926 with the Drawdown Simulator's engine, spending a fixed amount that rises with inflation until you, or the younger of you and a spouse, reach 95. The outlook is the share of those retirements where the money lasted, and the guide also solves for what you'd need saved by retirement to last in 90% of them.

Social Security and pensions

Estimated from your income and the years you'll have worked by retirement, counted from 22, since Social Security averages your best 35 years and missing years count as zeros. A lower-earning spouse gets at least half the higher earner's benefit. It starts at 67, or at retirement if that's later, unless you choose a claiming age from 62 to 70; a figure from your Social Security statement replaces the estimate and is scaled for the age you claim. A pension or other steady income can be added with its own start age, fixed or rising with inflation.

Adjust your plan

Compares your plan with what it needs and offers ways to change it, each solved against a target: plans that lasted in 90%, 95% or every historical retirement. Ahead of target, it finds how much earlier you could retire, the age you could stop contributing and coast, how much less you could save, or how much more you could spend. Behind, it finds how much more to save, how much later to retire, or how much less to spend. Balance it combines the changes you allow, moving each the same share of the way to its own answer, and finds the smallest combination that reaches the target (or the most the plan can afford, when ahead). Try your own numbers takes any mix. Each option is drawn against your current plan on a chart and a before/after table, and Apply to my plan makes it your plan, with Undo.

Drawing it down

Runs all six of the Drawdown Simulator's withdrawal strategies on your plan and shows, for each, how often the money lasted, a typical year's spending and the leanest single year on record. A minimum yearly spending sets a floor the flexible strategies never go below while money remains, which shows what that floor costs in safety. The score itself always tests the fixed approach, the cautious case. From here, a seven-part tour of the Drawdown Simulator covers your result, a bad starting year, each strategy, your stock mix, Social Security timing, one-time costs and extra income, and stress tests; a strategy, mix, claiming age or minimum you change there comes back into your plan.

Taxes not included

Retirement spending is before income tax. Withdrawals from traditional accounts, a pension and part of Social Security are taxable, so the guide offers a trip into the Income Tax tool's Retirement income mode, filled in with your plan's first year, and can add the estimate to your spending. Home equity isn't counted.

Save, share and print

With the Guide tab open, Save/Delete saves your answers and plan under a name, like any other tool's scenario, and picking it from the list later loads the whole plan back, household bar included. From the results, or the Share button while on the Guide tab, print the finished plan on one page or copy a link that opens it in the guide. The link carries your answers (not your budget or debt lists), so share it only with people you'd show your finances to. Someone opening it who has guide answers of their own is asked before theirs are replaced.

Every other tool in this app answers "how much will I have?" This one answers the harder question: once you stop contributing and start spending, will it last? It takes a portfolio balance, a withdrawal strategy, and tests it against real market history and randomized simulation to produce a success rate: the share of tested retirements where the money didn't run out.

Historical mode runs your plan starting from every single year since 1926: retire in 1929 right before the Depression, retire in 1966 into a lost decade, retire in 2000 before two crashes in one decade. Each is a genuine sequence of stock returns, bond returns, and inflation that actually happened, not a statistical approximation of one. This matters because the order returns arrive in changes everything: a bad decade in your first few years of retirement does far more damage than the same bad decade arriving later, even though the average return over the full period is identical. This is called sequence-of-returns risk, and it's the single biggest reason a plan that looks safe on a spreadsheet can fail in reality.

Monte Carlo mode draws years at random from that same 100-year record, thousands of times, to show the range of outcomes a purely random ordering could produce. It's a useful complement, but it misses the way real bad years cluster together; the historical mode is the more honest test of a specific plan.

Click any row in "How each starting year fared" to drill into that exact period: what was withdrawn each year, how the balance moved, and how much of that came from Social Security versus the portfolio itself.

Withdrawal strategy

The rule for how much to take out each year. Fixed amount sets a dollar figure and raises it with inflation every year, no matter what markets do: the classic 4% rule. Percentage of portfolio takes the same share of whatever the account is worth, so it can never run dry but income swings with the market. Guyton-Klinger Guardrails follows inflation but cuts or raises spending if the withdrawal rate drifts too far from target. Floor & ceiling aims at a percentage of the balance but limits how much spending can change from one year to the next. Yale Endowment blends last year's spending with a fresh percentage of the current balance, smoothing swings while still tracking the market. Variable percentage withdrawal (VPW), from the Bogleheads community, takes the payment that would spend the current balance down to a future value (usually $0) over the years left at an expected real return, so the share rises as the horizon shortens. How the strategies compare, under the picker, covers each in depth with its pros and cons.

Other income & future expenses

Beyond Social Security, add any other income (a pension, rental property, part-time work) or a planned future cost like a car or long-term care. Each can start in a specific year, last once, for a set number of years, or the rest of retirement, and can grow with inflation or stay fixed. Income offsets what the portfolio needs to provide, the same way Social Security does, with any surplus invested rather than wasted; expenses add directly to that year's spending regardless of the strategy chosen above.

Minimum spending

An optional hard floor, in today's dollars, that spending never falls below, even in a year the strategy above would otherwise call for less. Available on every strategy except fixed amount, which by design never drops in real terms anyway.

Maximum spending

The mirror of the minimum: a hard cap, in today's dollars, that the strategy's spending never goes above, however much a good year would allow. A future expense you add is on top of it, so that year can go past the cap by the expense. If the minimum is set above the maximum, the maximum wins. Available on the same strategies as the minimum.

Stock / bond mix

How the portfolio is split during retirement. A higher stock allocation has historically grown faster but swings harder in a downturn, which matters more once you're withdrawing from it.

Social Security

Optional income that offsets what the portfolio needs to provide. Enter your income for a rough estimate using the real Social Security benefit formula, or paste your own number from a statement at ssa.gov, which is always more accurate since it reflects your actual earnings history rather than an assumed steady career.

Success rate

The share of tested periods, historical years or Monte Carlo runs, where the portfolio lasted the full length of retirement without hitting zero. Higher isn't automatically better on its own: a 100% success rate paired with very low spending might mean money is being left behind that could have funded a better retirement.

What it leaves out: the 1926–2025 record is US markets only, so it says nothing about how other countries' markets have behaved, and the historical test has at most 100 starting years (71 for a 30-year retirement), not an infinite set of possible futures. Required minimum distributions and taxes on withdrawals aren't modeled. Above all: a strategy surviving every year since 1926 is strong evidence, not a guarantee: markets have never been obligated to repeat their history, and the next 30 years don't have to look like any of the last 100.

Retire before 59½ and most of your savings sits behind a 10% additional tax. This tool plans the years in between. You enter your balances by account type, your spending and your state, and it builds a separate plan for every way the tax code lets you reach that money early, runs each one year by year with the real 2026 tax, penalty and ACA rules, and tests each against every market since 1926.

Brokerage, then Roth contributions

The simplest bridge: sell taxable investments, where only the growth is taxed and often at 0%, then take back your own Roth contributions, which are never taxed or penalized.

Roth conversion ladder

Convert a year's spending from traditional to Roth every year, paying income tax at today's low early-retirement rates, and spend each conversion five years later, penalty-free. The first five years still need another source.

72(t) payments

Substantially equal periodic payments from an IRA carry no penalty at any age, but they're rigid: set by IRS formula and locked in until the later of five years or 59½. The plan splits off an IRA just big enough for the payment.

Rule of 55

Leave an employer in or after the year you turn 55 and that employer's 401(k) pays out penalty-free.

Pay the 10% penalty

Shown for comparison: what it costs to ignore every exception.

The blended plan

Mixes the routes: fills low tax brackets with conversions (or rule of 55 withdrawals), lives off the brokerage and Roth contributions while the ladder matures, and adds 72(t) payments only if the plan would otherwise run short. It tries each conversion level and keeps the one that holds up in the most historical markets, then the cheapest, counting lost ACA subsidies as a real cost.

Holds up in

The share of historical start years in which a plan gets to 59½ without running short or dipping into penalized money it didn't intend to. Monte Carlo draws random years from the same record instead.

At 59½

The tool stops at 59½ and shows what each plan leaves in traditional, Roth and brokerage accounts. From there, one button starts the Drawdown Simulator with that total, and another loads a year of withdrawals into Income Tax's retirement mode.

What it does not do. Everything is one household with one age, and both spouses' accounts are treated as one person's. Payroll tax is counted on part-time pay, but not the earned income credit or Roth contributions from that pay. The 72(t) annuitization method, HSAs, the SECURE 2.0 emergency and disability exceptions and state-specific exclusions that depend on age are left out, and the ACA premium uses your state's average benchmark plan unless you enter your own. Brackets and thresholds are held fixed in real terms.

Every projection in this app starts with a rate of return you have to supply, and most people pick one out of the air. This tool is where that number comes from. Choose a stock and bond mix, choose a stretch of history, and it reports what that mix did: compound return, the same return after inflation, volatility, the best and worst years by name, and the deepest fall from a previous high.

The rolling-returns table is the part worth sitting with. It shows every overlapping 1, 5, 10, 20 and 30-year window in the range, annualized. The median tells you what was typical; the worst column tells you what someone who started at the wrong moment actually earned, which is the figure a plan should survive.

Inflation gets its own panel rather than being buried in the real-return figure, because the spread is the point: the record runs from about −10% in 1932 to +18% in 1946 in single years, and the stretches that did the damage to retirements were sustained ones, not spikes. The ten-year line is drawn over the annual one for exactly that reason.

Use these figures in Advanced carries the return, the volatility and the average inflation over together. They belong together: a nominal return earned through a high-inflation stretch is only meaningful next to the inflation that came with it, and pairing that return with a different inflation assumption quietly changes the answer.

The data. Monthly, January 1926 to December 2025. Stock prices, dividend yields and 10-year Treasury yields come from Robert Shiller's dataset (Yale); CPI-U comes from the US Bureau of Labor Statistics. Monthly stock total return is the price change plus that month's share of the trailing dividend; monthly bond return is one month of coupon plus the price move implied by the change in yield. The calendar-year figures the Drawdown Simulator and this tool use are compounded up from those same monthly numbers, so there is one dataset aggregated two ways rather than two datasets that can disagree.

Shiller's prices are monthly averages of daily closes, not month-end closes — his long-standing convention, and the one most long-run retirement research is built on. It slightly damps single-year extremes against a December-to-December series (2008 reads about −39% here rather than −37%) while leaving long-run compounding alone: stocks still run near 10% a year over the period.

The assumptions. Stock returns are S&P 500 total return with dividends reinvested, and bond returns include the coupon, not just the price move — which is why stocks here compound near 10% a year rather than the 6–7% price-only figure people often have in mind. Annual rebalancing, no fees, no taxes, no trading costs, so weigh in your own fund's expense ratio separately. The growth chart runs from a fixed $10,000, since that changes the dollar amounts but nothing else about the mix's behavior. The bond series is 10-year Treasuries, not a broad bond fund. Past returns describe what happened; they are not a forecast, and the further back the data goes the less the world it describes resembles this one.

Two modes, switched at the top of the input panel. Normal income estimates what you keep from a salary after federal income tax, state income tax and FICA. Retirement income answers a different question: you are no longer earning a paycheck, you are drawing one down, and what a withdrawal costs depends entirely on which account it comes out of.

The numbers are tax year 2026. Federal brackets, the standard deduction and FICA limits come from the IRS inflation adjustments (Rev. Proc. 2025-32). State brackets, standard deductions and personal exemptions are as of 1 January 2026. Standard deduction is $16,100 single and $32,200 married filing jointly. Social Security is 6.2% on the first $184,500, Medicare 1.45% on everything, plus a 0.9% surtax above $200,000 single or $250,000 joint.

Normal income

Gross income

Salary before anything is taken out.

Pre-tax deductions

Traditional 401(k), HSA, and health premiums. These reduce income tax but not Social Security and Medicare, which is why the FICA figure doesn't move at all when you raise them.

Standard vs itemized

You take whichever is larger. Most people take the standard deduction; itemizing wins mainly with a big mortgage, large charitable giving, or high state and local taxes.

Effective vs marginal rate

Marginal is the rate on your next dollar. Effective is what you actually paid across all your income, and it's always lower. A single filer on $100,000 sits in the 22% bracket but pays about 13% of income in federal tax.

What Normal income mode leaves out: tax credits (child tax credit, earned income credit), local and city income taxes, the Alternative Minimum Tax, capital gains and investment income, self-employment tax, head-of-household and married-filing-separately status, state credits and phase-outs, and the extra deductions available at 65 and over. Most states phase out deductions at higher incomes in ways this doesn't model; Illinois is the one exception, since its exemption cliff at $250,000 / $500,000 of income is applied. Treat the result as a solid estimate, not a tax return.

Retirement income

A salary is one kind of income taxed one way. A year of retirement spending is usually four or five kinds of income taxed four or five different ways, and the total bill depends less on how much you withdraw than on which accounts you withdraw it from. Instead of one gross income figure, you enter each source separately; the tool totals them for reference and then taxes each one on its own terms. There is no FICA, because none of it is wages, and no net pay vs. take-home switch, because you are no longer contributing to anything.

Traditional 401(k) / IRA withdrawal

You deducted it going in, so all of it comes out as ordinary income at the same 10–37% brackets a salary would face. Required minimum distributions belong here too.

Roth withdrawal

Zero tax, federal and state. It also stays out of the provisional income figure that decides how much of your Social Security is taxable, and out of the MAGI that drives the net investment income tax and the senior deduction phase-out. That second effect is invisible on a tax return but worth real money.

Taxable brokerage withdrawal, and the gain percentage

Only the growth is taxable; the rest is your own basis coming back untouched, which is why the gain percentage matters as much as the withdrawal itself. Sell $40,000 from a position that is 30% gain and only $12,000 hits the return. Everything here is assumed to be long-term, held over a year, so it gets the preferential 0/15/20% rates rather than ordinary ones.

Social Security benefits

Somewhere between none and 85% of the benefit becomes taxable, worked out under the actual IRC §86 formula rather than assumed. What drives it is provisional income, everything else on the return plus half your benefits, against thresholds of $25,000 and $34,000 single, or $32,000 and $44,000 joint. Those four numbers were set in 1983 and 1993 and have never been indexed, which is why a rising share of retirees crosses them every year.

Other ordinary income

Pensions, annuity payments, interest, non-qualified dividends, rental income. Ordinary rates, no FICA.

Age 65 or older

Two separate deductions stack on top of the regular standard deduction. The long-standing age add-on is $2,050 for a single filer or $1,650 per qualifying spouse, and needs the standard deduction. The newer senior deduction from the 2025 tax act (OBBBA §70103) is $6,000 per qualifying person, runs only through 2028, is available to itemizers too, and shrinks by 6 cents per dollar of income above $75,000 single or $150,000 joint. A single filer 65 and over can reach $24,150 of deduction against the $16,100 a younger filer gets.

Why gains stack. Long-term gains do not get their own run at the brackets. Ordinary income fills the brackets first and then acts as the floor the gain sits on, so what the gain costs depends on what else you withdrew that year. In 2026 the 0% band runs to $49,450 of total taxable income single and $98,900 joint, 15% to $545,500 and $613,700, and 20% above. The stacking chart draws exactly this: your ordinary income as the floor, the gain on top of it, and the band boundaries it crosses. Where there is unused room in the 0% band the tool tells you how much, because that headroom is the whole basis of gain harvesting.

Why the marginal rate isn't the bracket. On a salary, the rate on your next dollar is just the bracket you are in. In retirement it usually isn't. Another $1,000 from a traditional account can pull several hundred dollars of Social Security into the tax base alongside it, and can shove gain out of the 0% band into the 15% band, so the real cost of that $1,000 runs well above the nominal rate through a wide band of middle incomes. Rather than report the bracket and call it the marginal rate, this runs the entire calculation a second time with $1,000 more of ordinary income and reports what actually changed. A 12% bracket showing a 22% marginal rate is not a bug; it is the effect worth planning around.

The 3.8% surtax. Above $200,000 of MAGI single or $250,000 joint, the net investment income tax applies to the lesser of your investment income and the amount you are over the line, on top of the capital gain rate, not instead of it. Those thresholds are statutory and have never been indexed either.

Pension / annuity, and why it has its own field. Pension income is split out from Other ordinary income because state law splits it out. A dozen states exempt government pensions in full while taxing private ones at full rates, and most states that give a retirement exclusion give it to pensions and retirement-account withdrawals but not to interest or rent. Federally the distinction does not exist; it is all ordinary income, so it changes only the state figure, but it can change it by thousands.

States, in full. Every state that taxes income is modeled on its own retirement rules, not just its brackets: whether Social Security is in the base, what pension and retirement-account income is excluded and up to what ceiling, what extra deduction, exemption or credit arrives at 65, and how long-term gain is treated. The rules are as of 1 January 2026 and include the phase-ins that finish this year. Select a state and the per-source table explains in a line what it does.

Social Security. Eight states still include some benefits in taxable income for 2026: Colorado, Connecticut, Minnesota, Montana, New Mexico, Rhode Island, Utah and Vermont. West Virginia finished phasing its tax out effective this year, and Missouri, Kansas and Nebraska dropped off earlier. Six of the eight are income-tested and the test is applied here: Connecticut and Rhode Island cut off at fixed AGI thresholds, Minnesota and Vermont phase out over a band, New Mexico is a cliff, and Colorado exempts benefits outright at 65. Utah taxes benefits and then hands back a credit for the tax on them, withdrawn above $54,000 single or $90,000 joint; that credit is computed here. Montana alone taxes the federal taxable amount flat, with no relief.

Where retirement income is barely taxed at all. Illinois, Mississippi and Pennsylvania take qualified retirement income out of the base entirely (pensions, annuities, 401(k) and IRA withdrawals alike), so in those three only the gain on a brokerage sale and genuinely other income are left. Iowa has done the same since 2023 for anyone 55 or older. Alabama and Hawaii exempt pension income in full but still tax the retirement account. Michigan's deduction is fully restored for 2026, the last step of its 2023 phase-in, at roughly $68,000 single and $136,000 joint.

Where the exclusion has a ceiling or an income test. Kentucky gives $31,110 a person at any age; Georgia $65,000 a person at 65, against unearned income of any kind; New York $20,000 a person on top of a full exemption for government pensions; New Jersey up to $100,000 joint but on a hard income test that steps to nothing by $200,000; Connecticut a full exemption below $75,000 single or $100,000 joint, phased out over the next $25,000. Maine and Maryland reduce their exclusions dollar for dollar by the Social Security you receive, which for a large benefit can consume them outright. South Carolina, Virginia, West Virginia and Montana give a flat deduction at 65 against income of any kind rather than a retirement exclusion; Virginia's is withdrawn dollar for dollar above $50,000 single or $75,000 joint.

State capital gain treatment. Most states tax long-term gain at their ordinary rates, and that is the default here. The exceptions are modeled: Arkansas excludes 50%, South Carolina 44%, New Mexico and North Dakota 40%, Wisconsin 30%, Vermont a flat $5,000. Hawaii caps the rate on gain at 7.25% and Montana taxes it at reduced rates of 3.0% and 4.1%, both as alternative computations that can never cost more than ordinary treatment.

What the state figure still leaves out. Age thresholds below 65 are treated as met only when you mark someone 65 or older, so a 62-year-old in Georgia or New Jersey, or a 60-year-old in Delaware, is shown a higher state bill than they would actually pay. Per-person exclusions are applied per qualifying person without checking which spouse the income belongs to. Occupational carve-outs (military, police, fire, railroad, federal Civil Service) and rules that turn on a birth year rather than an age are not modeled, and they are generous where they apply. Neither are local and city income taxes, which matter most in Maryland, New York City, Ohio and the Portland area. Also absent everywhere: short-term gains and non-qualified dividends, which would be taxed as ordinary income; qualified charitable distributions; IRMAA Medicare premium surcharges, which behave like a tax cliff just above these thresholds; the Alternative Minimum Tax; and state credits that phase out on income in ways too intricate to generalize. Treat the federal number as solid and the state number as a good estimate rather than a return.

A traditional 401(k) or IRA is a loan from the IRS, not a gift. The tax was deferred, not forgiven, and required minimum distributions are the year the bill comes due: from age 73, or 75 if you were born in 1960 or later, a slice of the balance has to come out every year whether you need the money or not. The slice is set by the IRS Uniform Lifetime Table and it widens with age, from about 3.8% at 73 to over 15% in your nineties.

That is the problem a Roth conversion is trying to get ahead of. Moving money from traditional to Roth means paying tax on it now, voluntarily, at a rate you choose, in order to avoid paying tax on a larger balance later at a rate somebody else chooses. The tool runs your household twice, once converting on the schedule you set and once not touching anything, and puts the two side by side.

Fill to the top of a bracket

The usual approach. Each year it converts exactly enough to reach the top of the bracket you pick and not a dollar more, after Social Security, RMDs and everything else has been counted. Because it solves the whole return rather than a simple subtraction, it picks up the Social Security tax torpedo and capital-gain stacking on the way.

Fill to an IRMAA threshold

The same idea against a different line. Medicare's income-related surcharge is a cliff, not a phase-in: one dollar over and the whole tier applies, so for some households the binding constraint is the IRMAA bracket, not the tax bracket.

Where the tax comes from

The largest single lever here. Paying conversion tax out of a taxable account moves the full conversion into the Roth. Withholding it from the conversion means a smaller amount actually lands, and every future year compounds on the smaller number. Switch between the two and watch the after-tax figure move.

The survivor's bracket

Sometimes called the widow's penalty. When the first spouse dies, the survivor files single the following year: roughly the same income, but single brackets and a single standard deduction. It is one of the strongest arguments for converting while both are alive, and the tool flags the year it happens.

Lifetime tax, present value

Every dollar of federal, state and IRMAA discounted back to today. A conversion always costs money now to save money later, so comparing raw totals across thirty years flatters it; discounting is the honest version.

After-tax net worth

Roth and brokerage at face value, traditional discounted by the rate you expect to be paid on it, whether by you later or by whoever inherits it. Comparing pre-tax balances across a Roth and a traditional account compares two different currencies.

Break-even

The first year the converting plan's after-tax net worth passes the do-nothing plan and stays there. Early years always look worse. If break-even lands past your horizon, the conversion is a bet on your heirs, not on you.

What it does not do. Returns are a single real rate, not a sequence, so a market crash in the middle of a conversion window changes the answer and is not modeled here. Estate tax, state estate tax, QCDs, the ACA premium credit for anyone retiring before 65, and the inherited-IRA ten-year rule are all absent. Brackets and thresholds are held fixed in real terms, so a change in the law, the most likely reason any of this turns out differently, is outside the model by construction.

Works out the full monthly cost of owning, not just the loan payment, and shows how the balance falls over time.

Principal and interest use the standard amortization formula. Every payment is the same size, but the split shifts: early on almost all of it is interest, and only near the end does most of it go to principal. The chart makes that crossover visible, and it's usually later than people expect.

Down payment

Enter a percentage or a dollar amount; the other updates itself.

Interest rate

Pre-filled with 6.71%, the Freddie Mac national average for a 30-year fixed loan as of 3 September 2026. Rates move weekly, and your own depends on credit score, term, points and lender, so replace it with a current quote.

PMI

Private mortgage insurance, charged when you put down less than 20%. Typically 0.3–1.5% of the loan a year. Modeled here as ending once the balance falls to 80% of the purchase price, which is when you can request cancellation; federal law only forces automatic removal later, at 78%, if you never ask.

Property tax and insurance

Pre-filled with rough national averages. Both vary enormously by location: property tax alone ranges from well under 0.5% to over 2% of value, so replace them with local figures if you have them.

The Already have this loan? toggle below the main inputs opens extra payments, a recast, and a refinance comparison — skip it entirely for a quick estimate on a home you haven't bought yet.

Extra payments

A recurring amount, a one-time amount, or both, applied straight to principal. By default the required payment stays what it was — you're just finishing early and paying less interest along the way.

Recast

After a one-time payment, choose to lower the monthly payment instead of shortening the loan: the balance re-amortizes over whatever's left of the original term, at the same rate. The payoff date doesn't move, but every payment after that point is smaller. Lenders often charge $150–500 for this, which isn't included.

Refinance comparison

Assumes you refinance today, on the loan as entered, into a new rate and term with the same extra-payment plan carried over — so the comparison isolates the refinance itself. Shows the new payment, how long closing costs take to earn back in lower payments, and which loan costs less in total once those costs are counted. Closing costs are assumed paid out of pocket rather than rolled into the new balance.

What it leaves out: closing costs rolled into the loan, adjustable rates, cash-out refinancing, the mortgage interest deduction, and maintenance beyond a flat percentage. A common rule of thumb is to budget roughly 1% of the home's value a year for upkeep.

Works backward from a target: pick a school type or type in your own annual cost, and it solves for the monthly savings needed to cover it, accounting for tuition rising faster than general prices.

School type presets

Rough all-in figures (tuition, room and board, and fees) for a public in-state school, a private non-profit, and an elite or Ivy League school. Overwrite the cost field with your own number if you have a specific school in mind.

Tuition inflation

College costs have historically risen faster than the inflation used elsewhere in this app, often 3–5% a year versus the 2–3% typical of general prices. The default reflects that gap.

What it leaves out: financial aid, scholarships, 529 plan tax advantages, and the possibility of a shorter or longer program than four years unless you change that input yourself.

The question isn't just "which payment is bigger"; it's which path leaves you with more net worth after some number of years. This tool runs both scenarios month by month: the buyer builds equity as the mortgage is paid down and the home appreciates; the renter invests whatever they didn't spend, starting with the down payment and closing costs they never paid, and continuing every month renting is cheaper than owning.

Break-even point

The year buying's net worth first overtakes renting's. Before that point, renting and investing the difference wins; after it, owning does. This shifts a lot with the interest rate, how long you stay, and how fast rent rises where you live.

Renter invests

The amount the renter has invested from day one (the down payment and closing costs they didn't spend), which then compounds for the entire comparison. This opportunity cost is often the most underrated part of the rent-versus-buy decision.

Home appreciation & investment return

Two separate assumptions that drive most of the result. Home appreciation has historically run below broad stock market returns over long periods, which is part of why the comparison isn't as one-sided as it might first appear.

What it leaves out: the mortgage interest deduction, moving costs beyond a percentage-based selling cost, rent control, and the non-financial value of owning (stability, the ability to renovate, not having a landlord), which matters to a lot of people and isn't something a spreadsheet can price in.

A simple worksheet for laying out where your money goes and seeing what's left. It starts with your income after taxes, lists the usual expenses grouped by category, and totals everything into a yearly and monthly figure.

Each line has its own amount and a per-month or per-year switch, because some costs are naturally one and some the other: rent is monthly, travel and car repairs are easier to think of as a yearly lump. Everything is converted to a common basis before totaling, so mixing the two is fine.

Income after taxes

Type it in, or use Copy from Income Tax to pull your net pay (gross minus taxes) straight from that tool.

The preset rows

A starting point, not a rulebook. Leave anything at zero to ignore it.

Add custom item

Adds a blank row you can name yourself, for anything the presets miss.

+ Retirement contribution & + College savings

Pull a monthly figure straight from another tool instead of retyping it. If more than one retirement mode (Basic, Advanced, Stages) has a contribution set, you'll be asked which one to use, with each option's monthly amount shown so the choice is never a guess.

Left over

Income minus everything budgeted, shown per year and per month, green when positive and red when you've allocated more than you make. The percentage is how much of your income is left, or how far over you are.

Like the other tools, a budget can be saved, and Save/Delete act on the budget while you're on it. It keeps its own separate list.

Two well-known ways to clear several debts at once, and they differ in exactly one respect: which debt gets the extra money. Everything else, how much you pay each month, the minimums, the rates, is identical.

Avalanche

Attack the highest interest rate first. This is arithmetically optimal: no other ordering of the same dollars pays less interest. If the only thing you care about is cost, the argument is over.

Snowball

Attack the smallest balance first, regardless of rate. It costs more, sometimes a lot more, but it clears an entire debt sooner, and there is decent evidence that people stick with it better. The tool prices that directly: it shows how much earlier the first debt disappears, and what those months cost you in interest. That is the real decision, and it is a personal one rather than a mathematical one.

The rollover

The engine behind both. Your monthly payment is held fixed at the sum of the original minimums plus anything extra. When a debt is cleared, its minimum does not go back into your pocket; it rolls onto the next debt in line. That is why the last debt gets paid off so much faster than the first, and why both strategies crush paying minimums alone.

Minimums only

The third row of the comparison, and the one worth looking at. It pays each minimum with no extra and no rollover, which is what happens by default if you do nothing. On typical card balances it is usually decades and tens of thousands of dollars worse.

Underwater minimums

If a minimum payment is smaller than one month of interest, the balance grows no matter how long you pay. The tool flags this rather than quietly projecting a payoff date sixty years out.

What it does not do. Real credit card minimums are a percentage of the balance, so they shrink as you pay down; this uses the fixed figure you enter, which makes the projection slightly optimistic unless you keep paying the original amount (which is what you should do anyway). Rates are fixed, so promotional 0% periods, variable-rate cards and balance transfers are not modeled, and neither is new borrowing. It also ignores the case for skipping all of this and grabbing an employer match first: a 50% match beats paying off a 24% card.

Healthcare is one of the largest and most variable costs in early retirement. This tool covers two phases: the ACA bridge between retirement and Medicare, and Medicare itself from age 65 on.

Pre-65: ACA bridge. Most early retirees buy coverage on the ACA marketplace until Medicare starts. Premiums are set by age and state; the tool uses the official HHS age-rating multipliers to scale a state-level benchmark to your retirement age. The premium tax credit reduces what you actually pay, and its size depends on where your income falls relative to the federal poverty line.

Retirement MAGI

Modified Adjusted Gross Income (MAGI) is the income figure the ACA and Medicare both use. In retirement this is typically taxable withdrawals, pension income, taxable Social Security, and capital gains added together. Use Copy from Income Tax to pull it automatically when the Income Tax tool is in Retirement income mode.

2026 rules, and the enhanced credits

The standard rules cap your premium at a sliding percentage of income, from 2.10% to 9.96% in 2026, and end subsidies above 400% FPL (the "subsidy cliff"). Enhanced rules, in place from 2021 through 2025, removed the cliff and capped contributions at 8.5% at every income. They expired at the end of 2025, so 2026 coverage runs on the standard rules; the tool also shows what the enhanced rules would give, in case Congress restores them.

The 400% FPL cliff

Under standard rules, crossing this line eliminates the entire subsidy in one dollar. The tool warns you when you're close and quantifies the exact monthly cost, which matters when weighing Roth conversions, capital gains realizations, or IRA withdrawal timing against ACA premium exposure.

State and age adjustment

The benchmark Silver premium is KFF's 2026 average for each state for a 40-year-old, scaled to your retirement age via the HHS multiplier table. Enter your own quote from healthcare.gov in the sidebar for a precise number.

Post-65: Medicare. Medicare Part B (outpatient) and Part D (prescriptions) both carry income-related surcharges known as IRMAA, based on your income from two years prior. Most enrollees add Medigap coverage to cap out-of-pocket exposure.

IRMAA

Income-Related Monthly Adjustment Amount. If your MAGI two years ago exceeded a threshold ($109,000 single / $218,000 married in 2026), Part B and Part D cost more, in six tiers up to 3.4 times the $202.90 standard Part B premium. The tool shows your tier, the dollar surcharge, and the exact savings from keeping income below the next tier down, which matters for withdrawal sequencing and Roth conversion decisions in the years before Medicare starts.

Medigap

Supplement plans like Plan G cover most of what Original Medicare leaves unpaid. Premiums vary widely by state, insurer, and age at enrollment; the range shown is a rough national estimate for a 65-year-old. Medicare Advantage (Part C) plans have lower premiums but narrower networks and different cost-sharing, and are not modeled here.

Monte Carlo asks what could happen if returns are drawn from a distribution. Historical asks a different question: what would this plan have done in the past that actually happened? It takes your contribution schedule and runs it through every overlapping window of the real record, January 1926 to December 2025, compounding month by month. A 30-year plan gets 841 runs — one starting in each month from January 1926 to January 1996.

The monthly step is the point. If the market falls 5% in March and you contribute that month, you buy at March's lower level and get the full benefit of whatever April does. Running a year as a single step averages that away. It also means a quarterly contributor buys in March, June, September and December rather than in twelve equal slices, and that starting in July 1974 is a different plan from starting in January 1974.

Two inputs stop being used in this mode, and it is worth being clear about which:

• Your rate of return is ignored. Returns come from the stock mix instead: that share in the S&P 500, the rest in 10-year Treasuries, rebalanced once a year. Fees still come off every year.
• Your inflation setting is ignored. Each window is converted to today's dollars using the inflation that window actually had, so a run through the 1970s is judged against 1970s prices.

A glide works differently here too. There is no rate to glide, so it walks the stock mix down instead, from your starting share to the ending share over the final years of the plan.

The bands read the same way the Monte Carlo fan does, except each one is a real stretch of history rather than a simulated draw. The line labeled worst is not a bad scenario someone invented; it is a period people lived through while saving. Worth knowing: the worst outcome for a saver is rarely the crash, because a crash early buys shares cheaply for decades afterwards. A long flat stretch just before retirement does more damage.

What it can't do. A century is a single overlapping sample, heavily US-biased, and the windows share almost all their months with each other, so 841 runs are emphatically not 841 independent trials — consecutive windows differ by one month at each end. A plan longer than 100 years has no window to run in and the mode says so instead of guessing.

FIRE (Financial Independence, Retire Early) means accumulating a portfolio large enough that investment returns can fund your spending indefinitely. The classic benchmark is 25× your annual expenses, which corresponds to a 4% annual withdrawal rate — though you can enter any target and withdrawal rate here.

FIRE mode finds the year your inflation-adjusted portfolio first reaches your target. The number shown is in today's dollars, so it stays meaningful even decades out.

Coast FIRE mode finds the year your portfolio is large enough that, with no further contributions, it will compound to your target by your planned retirement age. You still need earned income to live on — but the saving race is over. The "if you kept saving" stat shows how much larger your retirement portfolio would be had you never coasted, and the gap is the optional extra you're leaving on the table in exchange for financial freedom from saving.

Rate band shades the range between a return ± your chosen band, giving a fast read on how sensitive your FIRE date is to return assumptions. Historical mode runs the plan against every starting year since 1926, showing the p10/p25/p50/p75/p90 spread of real outcomes.

Success rate slider. The slider picks a percentile of historical starting years. At 50% you see the median outcome — half of historical periods hit your target sooner, half later. At 75% you see the date by which 75% of periods succeeded, which requires a later (more conservative) target date. Move it toward 99% to stress-test against nearly every historical period; toward 1% for the aggressive best-case. The success rate is purely historical — it says nothing about future markets.

The plain projection assumes you earn the same return every single year. Real markets don't work that way, and the order of good and bad years matters: a poor stretch early does more damage than the same stretch late, even at an identical average.

Switching the chart to Monte Carlo runs your plan hundreds or thousands of times, drawing each period's return at random from a distribution centered on your expected return, with the spread set by volatility. The result is a fan of outcomes rather than a single line.

• The median is the middle outcome: half of runs did better.
• The 10th to 90th percentile band covers 80% of runs. One run in ten lands below the bottom edge.
• The dashed line is the no-volatility projection. The median usually sits below it. That gap is real and is called volatility drag.
• The success rate is the share of runs that reached your target.

Results are reproducible: the same inputs give the same fan every time. Re-roll deliberately draws a fresh set.

Save

Names the current setup and adds it to the dropdown. Saving under the name of the scenario you already have loaded updates it; saving under a different existing name asks first before overwriting it. Scenarios are stored in your browser, on the device you're using; they don't sync between your phone and computer, and clearing site data removes them.

The dropdown

Each mode and tool keeps its own saved list (Basic, Advanced, Stages and every tool in the Tools tab), and the dropdown shows whichever list belongs to the one you're on. It remembers which scenario you loaded or saved as you move between tabs, and marks it "(edited)" once your current numbers no longer match what's saved. "Unsaved" means the current numbers don't match anything saved.

Compare

Opens a separate screen that puts two or three saved retirement scenarios side by side: their balances on one chart, then every result and every input in tables, with differing inputs highlighted. It reads saved scenarios only and changes nothing, so set your inputs and save them first. Each slot picks a mode (Basic, Advanced or Stages) and then a saved scenario from that mode's own list.

Delete

Removes the selected scenario. It doesn't change the numbers on screen, only the saved copy.

Reset

Returns just the tool or mode you're currently viewing to its defaults: Basic, Advanced and Stages each reset separately, and so does each tool in the Tools tab. Saved scenarios are untouched.

Share

One button, four options. Copy link encodes every input into the address itself, so anyone opening it sees exactly your numbers; nothing is uploaded, the whole scenario travels inside the URL, and it's also the most reliable way to move a setup between your own devices. On a phone it's Share link and opens your share sheet, so a text shows the site's preview card rather than a long address. Summary builds a clean one-page version and opens your print dialog, where you can choose Save as PDF; on iPhone use Share then Print. Save image card and Copy image card generate a shareable square image with your headline numbers, ready to post or send, built entirely in your browser, nothing leaves your device to create it.

• The retirement projections stop at retirement itself: "success" there means reaching a number, not that it lasts. The Drawdown Simulator is the tool that answers whether the money actually holds up once you start spending it.
• Tax in Basic, and in Advanced and Stages unless you split by account type, is a single flat rate. Split by account type works it out with the Income Tax tool's 2026 rules, but for the first year of retirement only, and without Social Security or other income.
• Contributions are assumed to continue uninterrupted. Job loss, career breaks and emergencies aren't modeled.
• Contribution limits and employer matching are only modeled when you split by account type. Social Security is modeled in the Drawdown Simulator, Income Tax and Roth Conversion tools, not in the retirement projections.
• Monte Carlo draws each period independently. Real markets show streaks and mean reversion that random draws don't capture; the Drawdown Simulator's historical mode exists specifically to get around this limitation for retirement spending.
• Every output is only as good as the assumptions typed in. Small changes to the return compound into very large differences over decades, which is exactly why the comparison band and Monte Carlo modes exist.