The full method behind the planner — how income is funded each year, how every tax, premium and benefit is modeled, how the readiness verdict is reached, and what is deliberately left out.
← Back to the modelThe target-income waterfall. You set a desired annual income after federal & state tax, Medicare, out-of-pocket medical and pre-65 health premiums (today's dollars, grown by inflation). Each year the model funds it from a priority waterfall: other income and Social Security first, then a retirement-account withdrawal solved by bisection to hit the target, then personal savings for any remainder. Within the accounts, withdrawals follow their own order — pre-tax (Traditional, taxed as ordinary income) drains first, then any HSA surplus, then taxable personal savings, and finally Roth, the last-resort filler so it keeps compounding tax-free. The Health Savings Account (HSA) gets special handling: it grows tax-free and is reserved first to cover your projected future Medicare premiums and out-of-pocket medical (both of which it can pay tax-free), with the reserve sized by present-valuing that remaining cost stream at your growth rate; only the surplus above the reserve is drained as ordinary income, so the account isn't spent down before the bills it exists to cover arrive. In Plain rule mode the model instead withdraws a fixed percentage of the starting balance, grown by inflation.
Taxes on your withdrawals. Personal savings splits into already-taxed principal (spendable tax-free) and unrealized gains, realized proportionally on each withdrawal and taxed at long-term capital-gains rates (0/15/20%), stacked on top of ordinary income. The Net Investment Income Tax (NIIT) — an extra 3.8% on investment income above $200k of MAGI (Modified Adjusted Gross Income) for a single filer, or $250k married filing jointly — is included, as is taxable Social Security via the provisional-income formula. The OBBBA senior bonus deduction — an extra $6,000 per filer aged 65+ for tax years 2025–2028, phasing out at 6% of MAGI above $75k single / $150k joint — is applied while it's in force. Filing status is single, or married filing jointly while both spouses are alive (set automatically by whether a spouse age is entered); federal brackets and the standard deduction (plus the 65+ addition) are inflation-indexed.
State income tax — coverage. The model covers all 50 states and DC: the nine no-income-tax states ($0), the fifteen flat-rate states, and every graduated-rate state with its full 2025 brackets (rates and thresholds from the Tax Foundation's 2025 tables, including California's 1% surcharge over $1M and the Massachusetts 4% surtax over ~$1.08M). Each state applies its own 2025 standard deduction or personal exemption (CPI-indexed), and Social Security is exempt in every state (only nine tax it, mostly via income limits). A custom flat rate is available for any edge case.
State income tax — retirement income & capital gains. Each state's retirement-income rules are applied: Illinois, Pennsylvania, Mississippi, Iowa and Michigan fully exempt IRA/401(k) and pension withdrawals; New Jersey and Connecticut fully exempt them below an income limit; and many others (Georgia, Colorado, Kentucky, Louisiana, New York, Delaware, Virginia, South Carolina, Maryland, Oklahoma, Rhode Island, West Virginia, Wisconsin, Maine, Arkansas) exempt a capped amount. Capital gains are taxed as ordinary income except where a state excludes part of them — Missouri (fully exempt), South Carolina (44% excluded), Arkansas (50%), Wisconsin (30%), and North Dakota and New Mexico (40%). Local and county income taxes (e.g. Maryland, Ohio, New York City, Kentucky), tax-benefit recapture, and credit-based deductions aren't modeled, so state figures are close estimates rather than line-by-line returns.
Bracket & IRMAA smoothing. An optional control caps pre-tax withdrawals at the lower of the first IRMAA cliff and the top of the 24% bracket, funding the rest from taxable savings then Roth — pre-tax is still drained first, just held under the cliff where savings and Roth can cover the gap.
Roth conversions. An optional "fill the bracket" conversion does the opposite in low-income years: after funding the year's target, it converts pre-tax → Roth up to the top of a chosen bracket (12%, 22% or 24%), paying the conversion tax from taxable savings so spending stays on target. This deliberately pays tax now at a known low rate to shrink future Required Minimum Distributions and build a Roth balance that later withdraws tax-free and never counts toward MAGI or IRMAA. Conversions are paused during an active 72(t) SEPP period (extra distributions would bust the plan) and resume once it ends; the amount converted each year appears in the Income & tax ledger.
Tax-strategy optimization. Rather than defaulting both controls off, the model runs all eight combinations of bracket/IRMAA smoothing (off/on) and Roth-conversion bracket (off, 12%, 22%, 24%) and selects the one with the smallest total tax footprint in today's dollars — lifetime federal & state income tax, the IRMAA surcharge portion of Medicare, and the income/estate tax your heirs would owe at the final death, each year deflated back to present value. Because a Roth conversion costs tax now but shrinks future RMD-driven income, IRMAA and the heirs' bill, only the full footprint reveals which combination is genuinely cheaper; a shortfall guard is applied first, so a strategy is never chosen if it would make the plan deplete earlier. The winning combination is used as the default and named in the verdict banner with its dollar advantage over doing nothing; override either control under Advanced and the banner shows the trade-off.
Planning horizon & survivor scenario. The planning age defaults to your life expectancy from the SSA period life table for your age and sex — editable; since mean life expectancy implies roughly a 50% chance of living longer, planning to 90+ is a common safety margin. For a couple, each spouse is projected to live to their own plan-to age and the projection runs to whichever spouse lives longer. When the first spouse passes, the survivor switches to single filing (single brackets, standard deduction, IRMAA and NIIT thresholds), keeps only the larger of the two Social Security benefits (the smaller check stops), and incurs Medicare premiums and out-of-pocket costs for one person. The shared portfolio keeps funding the same target for the survivor, and those years are marked with a ▸ in the Income & tax ledger.
Social Security. A career earnings record (from a career-start age through retirement — years worked are inferred, not entered) is built to track the national wage index: the salary you enter is read as your standing relative to the average wage and held across the career, so each year's earnings scale with SSA's Average Wage Index (AWI) and are capped at the wage base. The earnings are then indexed to today's wage level and the highest 35 are averaged into the AIME (Average Indexed Monthly Earnings), and the PIA (Primary Insurance Amount — the benefit at the full retirement age of 67) applies the 2026 bend-point formula: 90% of the first $1,286, 32% up to $7,749, and 15% above. A retirement benefit also requires being fully insured — 40 credits, about 10 years of covered work; a record with fewer working years produces no own benefit ($0 PIA). A spouse with a short record can still draw a spousal benefit (up to 50% of the insured spouse's PIA) and a survivor benefit on the other's record, and entering a benefit override bypasses the check if you in fact qualify from work not captured here. Indexing to the wage trend rather than back-projecting a nominal salary keeps the benefit in today's dollars and independent of the inflation assumption — changing the inflation rate no longer moves the projected benefit. A separate real wage growth assumption (default 1.1% above inflation, the AWI trend; editable under Advanced › Market & horizon, set 0 to disable) then lifts the benefit by (1 + real growth) compounded over the years until your age-62 eligibility: a worker retiring decades from now earns a higher real benefit, because the wage base and bend points they will face rise in real terms over their career, while someone already past 62 sees a small reduction. You can type any year's actual wages from your SSA statement to replace the estimate entirely. Claiming adjustments relative to FRA are −5⁄9% per month for the first 36 months early and −5⁄12% beyond (age 62 ≈ 70% of PIA), or +8%/yr delayed credit to age 70 (≈124%). The estimate is benchmarked against SSA's published 2026 maximum-benefit figures. The earnings records are editable under the Social Security input section — type any year's actual wages from your SSA statement for a far more accurate benefit, or reset to the modeled default. For a couple, both you and your spouse keep your own earnings records (salary, career-start age and an optional benefit override each), and each benefit is adjusted on that person's own claiming age. Spousal benefits follow the post-2015 deemed-filing rules, applied symmetrically with no setting to choose. Each spouse receives their own claim-adjusted benefit from their own claiming age — and an early claim is a permanent reduction that is never erased. Once both spouses have filed, whichever earned less also receives the spousal excess: 50% of the higher earner's PIA minus their own PIA (if positive), stacked on top of their reduced own benefit. So a spouse who claimed at 62 does not later jump to a fresh 50% when the higher earner files — they keep their reduced own benefit plus the excess. Only the excess is reduced for an early start (25⁄36 of 1% per month before the recipient's FRA of 67, then 5⁄12 of 1% per month), measured from when it first becomes payable — the later of the recipient's claim age and their age when the earner files — and it earns no delayed-retirement credit. (Restricted spousal-only applications, available only to those born before 1954, are not modeled; nor are SSA-44 appeals.) The Social Security summary tab shows each person's benefit and the household's lifetime benefits claimed, in both nominal and today's dollars. If one spouse dies, the survivor keeps the larger of the two benefits and the smaller one stops — so household Social Security steps down to a single check.
Optimal claiming strategy. The claiming decision deserves its own model, separate from the drawdown. Rather than summing benefits to a single plan-to age (which is only the drawdown horizon and moves with the preset), the model maximizes the expected present value of lifetime benefits — the academic best practice — by weighting each future year's benefit by the probability of being alive to collect it. It searches every claiming age from 62 to 70 (for both spouses jointly, when married) and picks the combination with the highest survival-weighted household total in today's dollars. Mortality model. Each spouse gets their own survival curve by sex (one male, one female): a Gompertz law, μ(x) = (1/b)·e(x−m)/b, with the modal age m and dispersion b calibrated to the SSA period life table (it reproduces male life expectancy at 62/65/70 of 19.7/17.5/14.0 years and female 22.6/20.2/16.3 within 0.02 of a year). Conditional survival from a current age a0 over t years is tpa₀ = exp( e(a₀−m)/b(1 − et/b) ). Joint states. For a couple each year is split into both alive (probability pyou·pspouse, paying each own benefit plus any spousal excess), only one alive (the survivor keeps the larger of the two records), weighted by those probabilities and summed over all future years. Two consequences fall out automatically: a later, larger check keeps its full delayed-credit premium because COLA equals inflation; and delaying the higher earner pays off twice — once for that earner's own longevity and again for every year the longer-living survivor (often the wife) collects the larger benefit — so the higher earner is usually nudged toward 70 while the lower earner may claim at 62. Because it's survival-weighted rather than tied to a single endpoint, the recommended claiming ages don't change when you switch presets. The chosen strategy and its dollar value appear in the verdict banner; overriding either age under Advanced shows what your choice costs versus the optimum. The Social Security summary tab shows the full claiming-age comparison — every combination (a 9×9 matrix for a couple) with its expected survival-weighted total, the optimal cell highlighted, your current selection outlined, and a hover breakdown of each strategy. Real discount rate. Future benefits are also discounted at a real rate you can set (default 2.5%, ≈ the 20-year TIPS yield). The rationale, following the standard approach: Social Security is an inflation-protected, government-backed stream, so the right discount rate is the return on the closest comparable asset — inflation-protected Treasuries — which is also the safe asset you'd spend down to fund delaying. A higher discount rate values later (delayed) checks less, so it shades the optimal toward claiming earlier; at 0% the model maximizes pure expected real benefits. The discount rate affects only this Social Security valuation, not the drawdown projection.
Medicare, out-of-pocket & health before 65. From 65, the model applies the Part B ($202.90/mo) and Part D ($34.50/mo) base premiums plus full IRMAA surcharges (doubled for a couple both on Medicare), grown to nominal dollars at a Medicare-specific premium trend (default 5.5% — Part B has risen ~4.5%/yr over 20 years and ~5.8% recently per CMS, below the broader commercial trend). IRMAA is the income-related surcharge for higher earners: the tier is set by your MAGI (Modified Adjusted Gross Income) from two years prior, compared against the 2026 brackets ($109k single / $218k joint and up) which are themselves inflation-indexed forward, so ordinary inflation alone doesn't push you into higher tiers. In the first two retirement years the lookback reaches back into working years, so the model uses the household's combined salary (projected to that calendar year) as the MAGI — matching the common experience of owing IRMAA right after retiring on a high final salary. (In reality a work-stoppage appeal on Form SSA-44 can often get that reduced; appeals aren't modeled, so this is the no-appeal baseline.) Hover any Medicare cell in the ledger for the exact tier and per-person breakdown. On top of premiums, the model adds out-of-pocket medical (cost-sharing, dental/vision/hearing and uncovered care) at every retirement age — defaulting to $5,000/person before 65 (marketplace deductibles run higher; average ~$3,800, KFF 2026) and $4,000/person on Medicare (KFF reports ~$6,460/yr total including premiums), both editable and grown at the healthcare trend. A larger Roth balance helps twice: Roth withdrawals don't count toward MAGI, so they neither add tax nor push you over an IRMAA cliff. If you hold an HSA, the model pays Medicare premiums and out-of-pocket medical from it tax-free first — including before 65, since an HSA covers qualified out-of-pocket costs at any age. Retiring before 65 also triggers an estimated unsubsidized ACA premium scaled to age (≈$1,448/mo for a 60+ adult, national average) at the healthcare trend (default 6.5%, vs ~7.5% recent individual-market trend); HSAs can't pay ACA premiums tax-free, so only those are funded from cash flow.
Early-withdrawal rules. The IRS 10% early-withdrawal penalty shapes how retirement accounts can be tapped before 59½, so the model adapts to your retirement age. At 55–59½ the Rule of 55 lets you draw your former employer's 401(k) penalty-free, so withdrawals stay flexible to hit the target (rows shaded amber, §) — note that IRAs don't qualify and would need to be rolled into the 401(k) first. Retire under 55 and penalty-free access requires 72(t) "substantially equal periodic payments" (SEPP): the model fixes the annual draw at an amortization-method amount (single-life expectancy, ~5%) until 59½ and routes any extra need to personal savings (rows shaded red, §). At 59½, normal flexible withdrawals resume.
Required Minimum Distributions. Under SECURE 2.0, pre-tax (Traditional) accounts must distribute a minimum each year from your required-beginning age — 73 if born 1951–1959, 75 if born 1960 or later (72 if born 1950 or earlier); Roth accounts are exempt during the owner's lifetime. The model treats the RMD as a withdrawal floor: it divides the prior year-end balance by the IRS Uniform Lifetime Table divisor (26.5 at 73, 24.6 at 75, 20.2 at 80, 12.2 at 90…) and forces the larger amount if it exceeds your planned draw (marked ‡). Because forced distributions are taxable, any after-tax surplus is reinvested into personal savings.
The readiness verdict — retiring now. The model solves for your maximum sustainable income: the level that exactly lasts to your planning horizon (for a couple, the later of the two plan-to ages). Green if that meets or beats your target; yellow if it's within 10% (or savings deplete only in the final tenth of the horizon); red below that. This is the steady-return grade — it assumes the same return every year. The volatility stress test (below) then re-grades the headline banner, so the colour, badge and title you see reflect the more pessimistic of the two views.
The readiness verdict — retiring later. If your retirement age is in the future, the model first grows today's balances to that age at your expected return with no new contributions, then asks by bisection how large a nest egg would make your target exactly sustainable — the nest egg needed. The difference is the gap, and it's closed with a level annual saving (a standard future-value-of-an-annuity calculation at your return over the years remaining). The savings-plan table then fills tax-advantaged retirement accounts first — 401(k) to the 2026 limit, then IRA (both doubled for a couple) — and routes any remainder to a taxable brokerage as personal savings. The verdict is green when the required saving fits inside those retirement-account limits, yellow when it also needs taxable saving on top, and red when it would exceed roughly 60% of your salary — a sign the target, timeline, or balances need rethinking.
Volatility stress test. The deterministic projection assumes the same return every year, which hides sequence-of-returns risk — an early run of bad markets can sink a plan that the identical average return would have saved. To measure that, the model re-runs the full drawdown 200 times with annual returns drawn from a lognormal distribution calibrated so the mean growth factor equals your expected return, with your chosen volatility (roughly 9% for a 30/70 portfolio, 12% for 60/40, 17% for all-stock; the assumption preset sets it to match its return — 9% / 12% / 16% — and it's editable under Advanced › Market & horizon, set 0 to disable). A fixed random seed makes results repeatable, and the 200 paths run in small asynchronous batches just after the main projection so typing stays responsive. Every tax, Social Security, Medicare/IRMAA and RMD rule applies inside every path. By default the model assumes a realistic spending-flexibility guardrail: in any year the market falls, the retiree trims spending by up to a set amount (default 20%, editable under Advanced › Market & horizon; set 0 to require the full target every year) and resumes the full target once markets recover. A path succeeds if it funds that spending in every year without ever depleting — so a down-year path kept alive by the trim counts as a success, which is how most retirees actually behave. Because that flexibility is the norm, it drives the headline success rate; the stress box also reports the stricter rate with spending held rigid at the full target, so you can see how many points of resilience the flexibility buys. It reports that success rate, the median and decile ending balances in today's dollars, and the typical depletion age in failing paths. How it sets the overall decision. The success rate is graded — ≥80% green, 65–79% yellow, below 65% red — and the headline verdict becomes the worse of the steady-return grade and this stress grade. So a plan that funds the target at the expected return but survives only a minority of volatile sequences is no longer shown green: its banner is pulled down to yellow (with a "mind the market sequence" caveat) or red ("vulnerable to a bad market sequence"), and the stress box states plainly whether the test confirmed or downgraded the steady-return call. A plan can never be graded up by the stress test — only held or pulled down. (Setting volatility to 0 disables the test and leaves the steady-return verdict untouched.) For a future retirement date the test assumes you arrive at retirement with the planned nest egg (the savings plan followed) and applies volatility to the drawdown years. (A freshly-built savings plan typically passes only about half the paths — see the FAQ for why, and the next section for how to add a cushion.)
How to pass the stress test. When the stress success rate lands below the 80% green line, the model doesn't just flag the problem — it reverse-engineers the smallest change that would lift you back to 80%+, by re-running the very same 200 seeded market paths inside a bisection search. It quotes three single-lever fixes plus one balanced combination. (1) A cash reserve. The model finds how large an out-of-market buffer (high-yield savings, money market, short T-bills) you'd hold and spend from during down-market years instead of selling — the classic defense against sequence risk, since drawing from cash in a downturn avoids locking in losses on depressed assets. In the simulation this reserve earns 0% real (it explicitly does not get market returns), isn't replenished, and is drawn first in any year the market falls. Crucially, the buffer is capped at about seven years of spending: beyond that, “cash” stops being a bridge over a downturn and just becomes a low-return slice of the portfolio (its drag then a real cost). So if even a seven-year buffer can't reach 80%, the model says a cash reserve alone won't fix this plan — the shortfall is structural (a long horizon and/or thin real return), and the answer is to spend less, work longer, or hold a genuinely lower-volatility portfolio, not to hoard cash. (Research is mixed on cash buffers generally — replacing bonds with cash helps worst-case longevity more than replacing stocks does.) (2) A spending trim. The highest target income that still clears 80% — equivalently, how much to cut. This is a permanent cut to the base target, applied every year on top of the down-year flexibility already baked into the headline — a larger lifestyle change than the guardrail. If you'd rather lean harder on flexibility instead of a permanent cut, raise the spending-flexibility dial under Advanced. (3) Working longer. The fewest extra years until retirement that pass, capturing both added compounding and a shorter drawdown (it ignores the extra saving you'd also do, so it understates the benefit). (4) A balanced combination. Each single-lever fix is the full dose of one move; the combination applies the same fraction of all three at once and solves for the smallest fraction that still reaches 80% — typically only ~30–45% of each, so no single lever is taken to an extreme (a modest cash buffer, a small trim, and a year or two longer, together). The panel also notes delaying Social Security toward 70 as a further, un-quantified lever — it lifts your guaranteed, inflation-linked income floor, which is market-proof by construction.
Estate & heirs. Whatever remains at the final death (the last surviving spouse) passes to heirs, and the model estimates two layers of tax. Federal estate tax applies to the gross estate — all accounts at face value — above the exemption, which the One Big Beautiful Bill Act set at $15 million per person from 2026 (a couple gets roughly twice that through portability of the first spouse's unused exclusion), inflation-indexed to the year of death; anything above is taxed at the top 40% rate, and it's flagged in the verdict banner up top when it bites. Then heirs face income tax as they take the money: an inherited Traditional IRA/401(k) must be emptied within 10 years and is taxed as ordinary income; a non-spouse's inherited HSA is immediately taxable in full. Both are estimated at an assumed 22% heir bracket. An inherited Roth comes out tax-free, and the taxable brokerage gets a stepped-up cost basis at death, wiping out the built-in capital-gains tax. The Accounts tab shows the full breakdown. (State estate/inheritance taxes and the income-in-respect-of-a-decedent deduction for estate tax paid are not modeled.)
Assumption presets. One control at the top sets the economic and longevity assumptions together: Conservative (4.5% return, 3.5% inflation, 7.5% / 6.5% healthcare and Medicare-premium trends, higher out-of-pocket, and a planning age at roughly the 90th percentile of survival — the age only ~10% of people your age and sex live beyond), Average (6% return, 2.5% inflation, 6.5% / 5.5% trends, mean life expectancy — the standard defaults), and Lenient (7.5% return, 2% inflation, lower trends and out-of-pocket, mean life expectancy). Each preset also sets the return volatility to match the portfolio its return implies — 9% Conservative, 12% Average, 16% Lenient (a bond-heavy mix swings less than an equity-heavy one) — so the stress test isn't run at a single uniform volatility regardless of how the money is invested. Conservative is a stress test; lenient is optimistic. A preset simply fills the relevant fields — you can still override any of them individually afterward. The 90th-percentile survival ages come from the CDC/NCHS 2022 U.S. period life table (the number-surviving column).
What isn't modeled. Tax credits, the 25% missed-RMD excise penalty, the joint-life RMD table for a much-younger spouse, Medicare Advantage, the Part B and Part D late-enrollment penalties (the Part B surcharge is a permanent 10% of the base premium for each full 12 months you delayed enrolling) and IRMAA appeals (Form SSA-44), and future tax-law changes (other than the senior deduction's scheduled 2028 sunset) are all out of scope. Retirement timing is a single household event — the one "retirement age" you set applies to both spouses (the spouse is taken to retire in the same calendar year, at whatever age they are then), so a staggered or phased retirement, where one spouse keeps working while the other has retired, isn't modeled directly (the FAQ shows a workaround). On the state side, local and county income taxes, tax-benefit recapture, and credit-based deductions are excluded; income-tested retirement exclusions (e.g. New Jersey, Connecticut) are approximated as simple income cliffs; and a handful of states' capital-gains and senior-specific breaks are simplified — so state figures are solid estimates rather than line-by-line returns.
Two portfolios can earn the exact same average return and still end up worlds apart — purely because of the order the good and bad years arrive in. That’s sequence-of-returns risk, and it only bites once you’re withdrawing.
Here are two retirees who both start with $1,000,000, both draw $70,000 a year, and both live through the same ten yearly returns — five years of +25% and five of −15%. The only difference is the order they come in:
| Scenario | Start | After 5 years | End (year 10) |
|---|---|---|---|
| Bull years first (the five +25% years, then the five −15%) | $1.00M | $2.48M | $0.84M — survives |
| Bear years first (the five −15% years, then the five +25%) | $1.00M | $0.18M | $0 — runs dry |
The bear-first retiree gets the identical five +25% years — just later — but there’s too little left for them to rescue the plan, and it hits zero in year 10. The bull-first retiree rides the downturn from a height and still ends with $840,000.
Both averaged 5% a year (about 3.1% compounded), and with no withdrawals both would have ended at the identical $1,354,081 — proof the order alone, not the returns, caused the gap. Drawing a fixed income turns an early downturn into permanent damage: you sell more shares while prices are low, leaving fewer to rebound. The volatility stress test exists to catch exactly this — it re-runs your plan across 200 different orderings and reports how many survive.
The big banner verdict assumes the same return every year. The volatility stress test then re-runs the full plan across 200 randomized market sequences and grades the success rate — ≥80% green, 65–79% yellow, below 65% red. The headline becomes the worse of the two views, so a plan that funds the target at a steady return but survives only a minority of bad sequences is pulled down to caution or red. It’s never graded up by the stress test — only held or lowered. Set volatility to 0 to turn the test off and see the steady-return grade alone.
For a future retirement, the savings plan is sized to make your target income exactly sustainable at the steady expected return — a zero-margin plan. With spending held rigid, a zero-margin plan is essentially a coin flip: roughly half of randomized sequences do a little better than average and survive, half do a little worse and fall short, so the strict success rate hovers near ~50%. The default 20% down-year flexibility adds a cushion and lifts the headline into the low-to-mid 60s — better, but still short of the 80% green line. That isn’t a bug — it’s the signal to build in more margin (save a bit more, plan to spend a bit less, flex a bit more in down years, or hold a cash buffer). The “how to pass the stress test” panel quantifies exactly how much of each it would take to reach 80%.
When stress success is below 80%, the model reverse-engineers the smallest change that would lift it back to 80%, re-running the very same 200 market paths inside a bisection search. It quotes three single-lever fixes — the cash reserve to hold and spend in down years, the spending trim, and the extra years of work — each the full dose of one move, plus a balanced combination that applies a moderate fraction of all three at once so nothing is taken to an extreme.
Two things to know about the cash figure. It’s modeled as money held outside the market at 0% real (no growth), so the model fully accounts for cash not earning returns — but the amount is quoted as a buffer held on top of your portfolio, which is why it doesn’t lower the portfolio’s return in the math. And it’s capped at about seven years of spending: if even that can’t reach 80%, the tool tells you a cash buffer alone won’t fix the plan, because the shortfall is structural rather than a short market dip. See the full method for the mechanics and the research behind it.
Almost everything you set and read is in today’s (real) dollars — your target income, the sustainable-income figure, lifetime Social Security and tax totals, and the ending-balance medians. Internally the model inflates each year to nominal dollars to apply real-world brackets, premiums and COLAs correctly, then deflates the results back to today’s purchasing power so the numbers stay comparable to what money is worth now. The year-by-year ledger shows nominal dollars for the year in question.
An own retirement benefit requires being fully insured — 40 credits, about 10 years of covered work. If the earnings record entered has fewer working years, the model includes $0 for that person’s own benefit. A married person who isn’t insured can still draw a spousal or survivor benefit on an insured spouse’s record. If you actually qualify from earlier work that isn’t captured, lower the career-start age or enter a benefit override in the Social Security section.
By design. The optimal claiming recommendation is computed from a survival-probability model (a male and a female mortality curve calibrated to the SSA life table), not from your plan-to age — so picking Conservative, Average, or Lenient (which set return, inflation and the drawdown horizon) leaves it unchanged. That’s the right behavior: when to claim Social Security should depend on your longevity odds and earnings records, not on how cautiously you’ve set the rest of the plan. Your sex (and your spouse’s) does move it — a longer-lived profile favors delaying. The plan-to age still governs everything else (whether the money lasts, taxes, the drawdown), just not this decision.
IRMAA — the Medicare high-income surcharge — is set by your MAGI from two years prior. In your first couple of Medicare years that lookback reaches back into your working years, so the model uses your final salary as that year’s MAGI — matching the common surprise of owing IRMAA right after retiring on a high income. In reality a work-stoppage appeal (Form SSA-44) can often get this reduced; appeals aren’t modeled, so the figure is the no-appeal baseline.
Not directly — the single “retire at” age is one household event applied to both spouses, so a staggered or phased retirement isn’t modeled on its own. As a workaround, set the retirement age to when the last paycheck stops, and enter the still-working spouse’s pay in the meantime as other income so the household cash flow stays right.
The savings plan fills tax-advantaged accounts first — 401(k) to the 2026 limit, then IRA (both doubled for a couple) — and routes any remainder to a taxable brokerage. A yellow “needs taxable too” verdict simply means the required saving exceeds those limits and you’d invest the extra in an ordinary brokerage account. A red verdict means it would take more than ~60% of your salary — a sign the target, timeline, or starting balances need rethinking.
No. The planner is a single static page that runs entirely in your browser — nothing you type or import is sent to any server, logged, or stored, and a refresh clears it. The only network requests are for the chart library and web fonts, which carry none of your data. Use Export scenario if you want to save your inputs to a file yourself.
Sources: IRS Notice 2025-67 (2026 contribution limits & tax brackets) and the OBBBA-adjusted 2025 standard deduction, bonus senior deduction, and $15M estate-tax exemption (40% top rate); CMS 2026 Medicare Part A & B premiums and IRMAA tiers; SSA — wage base, Average Wage Index, 2026 bend points and maximum-benefit figures, the Uniform Lifetime & single-life tables, and the 2022 period life table (2025 Trustees Report) for life expectancy, with CDC/NCHS 2022 U.S. life-table survivorship for the conservative preset's 90th-percentile longevity; IRS 72(t) guidance; Tax Foundation, State Individual Income Tax Rates and Brackets, 2025 (state rates, brackets & standard deductions for all 50 states and DC), with California FTB 2025 brackets and state departments of revenue for retirement-income exclusions and capital-gains treatment; KFF / MoneyGeek / PwC (2026 ACA premiums & medical cost trend). A hypothetical, educational model — not financial advice.