The receipts behind the math.

Alabama lets a household deduct its federal income tax. We size that deduction from your federal tax before a few credits (education, Earned Income, additional Child Tax Credit) that the state's worksheet subtracts. For most retirees those are zero; for a lower-income or working household with children the deduction can be slightly too large, understating Alabama tax a little.

A Coverdell balance that isn't used up must be paid out once the child turns 30, with the earnings taxed plus a 10% penalty. We don't yet model that age-30 payout, so a Coverdell you never draw on keeps growing tax-free past 30 in your plan — making it look slightly better than the law allows. Most Coverdells are spent on college well before 30, so this only matters for a balance left unused.

The 10% penalty on pre-59.5 retirement withdrawals is applied to the money you draw to spend and to the amounts drawn to pay that year's federal and state tax. Only a small residual from the one-pass tax-on-tax estimate is left out, so the penalty (and cost) can be very slightly understated.

When you don't give us the acquisition and sale dates, long-term and other holding-period tests for equity compensation are measured in whole calendar years, so a holding just under a year that crosses New Year's can be treated as long-term and taxed at the lower rate. Give us the acquisition and sale dates to classify the holding to the day.

For a married couple we assume the first spouse's unused estate-tax exemption carries over to the survivor (the portability election). If that election isn't actually filed the real shield is smaller, so estate tax can be understated. Binds only on estates above one exemption (about $15M).

Once your lifetime gifts use up your remaining gift-and-estate tax exemption, the gift tax owed on the excess is paid by selling from your taxable investments. We don't add the capital-gains tax on that sale, so your projected wealth in those years can run slightly high.

A fallback path taxes a capital gain at a flat 15% instead of the full 0/15/20% brackets plus the 3.8% investment tax. In real plans the full bracket path runs instead; this flat leg only appears in isolated internal calculations.

When HSA funds are treated as non-qualified, we apply the income tax and 20% penalty both to the amount you draw to spend and to the HSA dollars pulled to pay your federal and state tax bill. A small residual (HSA dollars covering the one-pass tax-on-tax estimate's own draw or a one-off outflow, a rare deep-deficit path) can still escape, so tax can be very slightly understated.

A life-insurance-trust benefit is modeled as fully outside the taxable estate, assuming a newly issued policy. If an existing policy was moved into the trust within three years of death, the law pulls it back into the estate, so estate tax can be understated in that case.

Maryland's 2% surtax on large capital gains applies once your federal income clears $350,000. When a business-sale or rental gain is the main thing that would carry you over that $350,000 line, we don't always count it toward that test — even though it still counts toward the gains the surtax is measured on — so in that case the surtax can be missed and your tax understated.

For a married couple filing separately in a community-property state, the state tax on out-of-state municipal-bond interest can be left out of the separate-return calculation, which understates the tax and overstates your wealth. This affects only that narrow filing situation.

A post-retirement stock-compensation sale is now taxed correctly against your concurrent retirement withdrawals when stock compensation is your only such income. In two cases the capital-gain rate can still be set too low (undertaxed): when you also have rental, self-employment, or business income that year, or when you file married-separately.

For a home sale (or a jointly-owned inherited-asset sale) we now apply the 3.8% investment tax when it is your household's only investment income. Two gaps remain: when another investment already carries that tax, this sale's gain isn't added to it, and each gain is bracketed on its own rather than stacked with a year's other sales, so a multi-sale year can be undertaxed.

When you don't say otherwise, we assume you actively participate in your rental — the typical case for a direct owner — which grants the up-to-$25,000 special allowance to offset other income. A purely passive investor (a limited partner, or under 10% ownership) gets none, so this can understate tax in that case. Telling us your participation removes the assumption.

A real-estate professional can elect to treat all rentals as one activity, and we assume you make that election. If you don't, each rental is tested on its own: one you don't run actively enough is passive, so its loss is deferred instead of used now and its income stays in the 3.8% investment tax — so this can make your plan look slightly better until you tell us you didn't elect.

Roth earnings drawn before 59.5 to pay federal and state tax are now taxed and penalized; only a small residual from the one-pass tax-on-tax estimate can still escape in a deep-deficit year (once every other account is exhausted), so tax can be very slightly understated in that rare case.

When you sell your primary home, the exclusion ($250k single / $500k married) applies only if you qualify — a home bought during your plan and sold within two years no longer gets it. We still assume a home owned before your plan met the two-of-five-year ownership/use test, and that you had no earlier qualifying sale unless you tell us, so a recently bought or rented home could be understated.

If you have foreign earned income that US law lets an expatriate exclude, it still counts toward the income figure behind health-insurance subsidies, Social Security taxation and Medicare high-income surcharges. We do not model foreign earned income yet, so for such a household that figure is understated — overstating the subsidy and understating those taxes and your wealth.

For the few state tax figures a state's own law leaves frozen (not inflation-indexed), we hold them steady in real terms rather than letting them erode. That credits slightly wider brackets or deductions than reality, so state tax can be understated over time.

In Maine, the standard deduction phases out as income rises, over a fixed dollar income range that Maine does not inflation-adjust (only the deduction amount is adjusted). Held steady in today's dollars, that range lets a later-year Maine household in the phase-out band keep slightly more deduction than the law allows -- a small understatement of Maine tax that grows over a long horizon.

The income-contingent student-loan repayment schedule has only single and married columns, so a married-filing-separately borrower is calibrated on the single column, which produces a slightly lower payment.

For a trust you fund before retirement that pays income out to you, our detailed year-by-year (deterministic) view doesn't yet apply your tax on those payouts during your working years — so on that view your wealth can look somewhat high. Your main projection, the one your plan is scored on, taxes those payouts correctly in every phase.

For a trust you fund during your lifetime, we compute its yearly income tax and carry a down-year investment loss forward to offset a later gain. One simplification: we treat the trust's earnings as dividends and long-term gains, not as interest or other ordinary income, so if it holds bonds or similar its tax can be slightly understated.

When you inherit from a parent's estate, we apply federal estate tax on it before the inheritance reaches you, plus the parent's state estate tax once you tell us the state they lived in. We still leave off state death tax for a few states whose inheritance tax we don't yet model, or when the parent's state hasn't been given, so the inheritance can be overstated in those cases.

Your working-years state income tax follows the state you live in each year, which is exact when your work moves with you or you work remotely. If you instead commute to a job in a different state than where you live, and that state's income tax is higher than your home state's, your state tax can be understated until you can tell us where you work.

You enter one annual health-insurance premium as this year's cost, and we now age it up the Federal age-rating curve as your household gets older, so a pre-65 bridge gets more expensive over time. We still use that entry as the benchmark that sizes your ACA subsidy, and real premiums also vary by county and plan tier, so it can run high or low for your area and plan.

We estimate the banker, legal and escrow costs of selling a business as a flat percentage of the sale price, which can be above or below your actual transaction costs.

Whether a business-sale gain counts as net investment income is decided from a single materially-participated flag rather than the full active-versus-passive asset test, which can push the 3.8% investment tax too high or too low.

In the nine community-property states, married-filing-separately returns keep passive-loss and Social-Security-taxation figures pooled at the household level rather than split per return, which can run either way depending on your situation.

For a jointly-filing couple whose accounts are not individually labeled, the 10% early-withdrawal penalty is gated on the primary spouse's age only, while the retirement pool it applies to is shared, so a couple with one spouse over 59.5 and one under can be penalized a bit too much or too little. A separately-filing couple with labeled accounts is penalized per spouse.

When you don't tell us how much of a 529 balance is your original contributions, we assume the whole balance is contributions. A seasoned account usually holds embedded gains, so the tax on a non-qualified withdrawal can be understated. Telling us the contributed amount removes this.

We treat a 529 contribution as pure cash out, with no state income-tax deduction or credit. About 30 states plus DC offer one, so for a contributing resident of those states we may slightly overstate state tax.

Lifetime gifts to family heirs leave your taxable estate but stay in the family, so we add them back into your projected inheritance, grown at the same return your own investments earn. We credit them at full value without modeling the heirs' own capital-gains tax on the gifted assets or a return that differs from yours, so the inheritance shown for the gift can run either high or low.

For a married-filing-separately household we count one Medicare enrollee unless the plan splits ownership between spouses, which can understate a 65+ second spouse's cost. The separate-filer ACA household-size default (1) now scales an out-of-pocket estimate whose right coverage-count is an open question, so it can run high or low. The separate ACA subsidy denial is handled correctly.

We now model retirement health costs beyond premiums — deductibles and coinsurance up to the out-of-pocket max, Medicare Advantage cost-sharing, a Medigap/supplement premium, drug out-of-pocket, and dental/vision/hearing — using national averages and a default Medicare Advantage coverage choice. Your real cost can be higher or lower; you can enter your own figures to replace the averages.

When the beneficiaries you've named don't add up to your whole estate, the leftover share is taxed at a single assumed heir tax rate, because the remaining heirs' own incomes aren't known. That assumed rate can be higher or lower than their real rate, so the tax on this leftover portion can run either way.

We model the cost of selling a home as a flat ~8% of the price (commission plus closing), not the itemized, market- and state-specific costs, so net proceeds can be a little high or low. This stays an approximation even if you override the percentage.

A completed emergency-savings goal is parked in a holding account earning a 0% real return. When true savings rates beat inflation this understates growth; when they trail inflation it overstates it.

International stocks have no data before 1970, so for earlier simulated years they are estimated from US stocks of the same year plus a market-specific wobble. From 1970 on, actual paired international returns are used, so US and international move together realistically. Only the estimated pre-1970 stretch can make the outcome range slightly wider or narrower.

Medicare's high-income surcharge is legally based on your income from two years earlier, and we apply that lag. For the first year or two of a plan where you reach Medicare right away, that earlier year predates your plan: tell us your income from then and we use it, otherwise we fall back to current-year income for those years, which can run the surcharge a little high or low.

Rent from a jointly inherited property is taxed at your ordinary and flat state rates only, with no depreciation or business-income deduction (which tends to overstate tax) but also skipping the 3.8% investment tax (which understates it). The net depends on your income.

After retirement, the tax rate on your rental, self-employment, or business income is set using a smooth estimate of your portfolio withdrawals that year rather than the exact amount, so its rate can be slightly too low in a year you draw more than estimated (undertaxed) or slightly too high in a year you draw less. The estimate keeps the dashboard and the range-of-outcomes projection consistent.

The tax saving from the $3,000 capital-loss deduction is valued at a single flat marginal rate rather than the exact bracket effect in the year it's used, which can overstate or understate the saving. The $3,000 cap itself is inflation-adjusted correctly.

Long-term-care costs use your state's median as the finest detail, which overstates lower-cost rural areas and understates expensive metros within a state.

The AGE at which long-term care begins is modeled with an age-rising hazard (care concentrated in late life), but each episode's LENGTH is drawn from a fixed range regardless of onset age. Since real episodes tend to be shorter when care starts very late, this can overstate the most-common late-life episodes and understate early-onset ones. Your lifetime chance of needing care is held fixed.

Missouri lets a household deduct part of its federal income tax -- capped ($5,000, or $10,000 jointly) and phased to zero above $125,000 of Missouri income. We model this, including how Missouri's capital-gains exclusion can lift you into a more generous tier. A few smaller worksheet details stay approximated, but the whole effect is small -- at most about $235/$470 a year.

If you move between states and haven't told us which state issued each of your municipal bonds, the taxable share is estimated from one in-state percentage for both the before- and after-move periods, rather than re-figured from each bond's issuing state — right for the state you left but possibly wrong for the one you moved to. Telling us the breakdown by issuing state makes it exact.

The tax on selling a property is applied at your home state's capital-gains rate unless you tell us which state the property is in. For a property in your own state that is exact; for one in a higher-tax state it can understate the tax, and in a lower-tax state overstate it. Entering the property's location removes this.

When you don't specify a land value, we assume land is 25% of the price (so 75% is depreciable building) — a national-typical split. In high-land markets this over-depreciates; in low-land markets it under-depreciates. Giving a land value removes this.

When you don't enter a rental's property tax, we model it at about 1.1% of the purchase price — a national-average rate — instead of leaving it at zero. Real property-tax rates vary widely by location (roughly 0.3% to over 2%), so for a high-tax area this can understate the expense and for a low-tax area overstate it. Entering the actual property tax removes this.

The cost of selling a rental is modeled as a flat percentage of the price, which can be above or below your actual commission and closing costs.

When you don't specify a vacancy rate, we assume about 7% of gross rent is lost to vacancy and credit loss — the national-typical rate from the U.S. Census rental vacancy survey. A property with lower true vacancy is slightly understated; one with higher true vacancy is slightly overstated. Entering your own vacancy rate removes this.

A fallback required-minimum-distribution start age (73) is used only when a birth year is missing. It can run two years early, pushing distributions sooner but also closing the Roth-conversion window sooner, opposite effects. In real plans a birth year is always present, so this fallback doesn't run.

When you don't tell us how much of your Roth is contributions versus growth, we estimate that split from your age and the annual contribution limits. The pre-59.5 income tax and 10% penalty on an early earnings withdrawal are modeled from the estimate, so the tax can be a little high or low if your real history differs. Give us your Roth contribution basis to remove this opening-split estimate.

A self-employed defined-benefit or cash-balance plan's deductible contribution is sized with an assumed 5% growth rate and a standard mortality table, standing in for the rates and tables your plan's actuary would use — so it can run either way versus your actual plan. The funded benefit is correctly capped at your earned income, not the headline dollar maximum.

With your birth month we apply the Social Security earnings test month by month in the year you reach full retirement age — the higher exemption covers only the months before that birthday, none after. Without one we fall back to a whole-year estimate that can over- or under-state the benefits you keep depending on your earnings and how long you live. Add your birth month to model it exactly.

When we estimate your Social Security benefit, we approximate a 35-year earnings history with your current income, then apply the SSA formula. This overstates the benefit for someone at a late-career peak and understates it for someone now below their peak. It applies only when the benefit was estimated, not when you entered your own figure.

A state capital loss is given no state benefit, with no $3,000 ordinary offset and no carryforward, while most states follow the federal treatment, so state tax can be slightly overstated.

For states without a modeled bracket table, we apply the top marginal rate to the whole state base rather than graduated brackets. That overstates tax for a graduated state at typical incomes, but understates it where local (city or county) income taxes apply. New York is no longer in this group -- its graduated brackets and high-earner benefit recapture are now modeled.

For the two years after a spouse's death, a survivor keeping a home for a child under 19 now gets married-level brackets and standard deduction in retirement instead of single filing. Edges: not applied to working-year gains, a 19-23 student child, or married-filing-separately (may overstate tax); a young non-child dependent could wrongly trigger it (may understate tax).

A credit-shelter (bypass) trust funded at the first spouse's death passes to your heirs outside the surviving spouse's estate; we now add it back into your projected inheritance. Because it gets no fresh cost-basis step-up, we estimate your heirs' capital-gains tax on its built-in gain at a single long-term rate from the household's bracket, so the inheritance shown can run slightly high or low.

We compute the state tax a lifetime-funded trust owes on what it keeps, based on the state it's a resident of — which you tell us, and which defaults to where it's sited (nothing for a no-trust-tax state like Nevada or Delaware). Which state counts as its residence can be a judgment call, and some state-specific trust rules aren't modeled yet, so the figure can run high or low.

In Alabama we leave out a few smaller items that would lower your state tax -- the per-dependent exemption, a repayment of excess health-insurance subsidy, the federal-tax deduction on a trust's own return, and, in a year you move into or out of Alabama, the federal-tax deduction on the Alabama part-year portion. Each omission can only over-state Alabama tax, so it is on the conservative side.

In Arkansas, we apply the state's tax schedule without subtracting its $29-per-person personal tax credit, so a modeled Arkansas household's state tax is slightly overstated -- about $29 per household member each year. This is deliberately conservative: leaving out a credit can only over-state tax, never understate it.

When you sell a business in a year you also own a rental, we model how the two interact in both directions. The one small exception: if that rental runs an operating LOSS in the sale year and you're a real-estate professional, we don't let the loss lower the tax rate on the sale's capital gain, so the sale can be taxed a touch too high (never too low).

A charitable gift left at death lowers your estate tax and passes to charity instead of your heirs, so we reduce both. We assume it's funded evenly from all your accounts rather than first from pre-tax retirement accounts — which a charity, unlike your heirs, owes no income tax on — so the inheritance shown for your heirs is estimated on the low, conservative side.

The education tax credit is now applied in your working years too, to the cent. Two small choices stay cautious (safe-direction): the credit's cash is added to your savings just after that year's growth rather than compounding within it, and the credit that can offset tax is measured a touch conservatively. Both only ever understate your wealth, never flatter it.

The education tax credit is computed exactly to the law, and we claim it only when it actually helps you — never paying the matching tax for a credit your income has phased out. A couple of choices stay cautious (safe-direction): by default we apply it to the largest eligible tuition and offset it only against ordinary income tax. Entering your out-of-pocket expense removes the first.

A business loss is allowed to offset your wage and other income with no annual limit. The law caps how much business loss can offset non-business income in a year and carries the excess forward against later income, so in a large-loss year your tax can be understated. Building that cap and the carryforward is a planned next step.

When you name your estate's beneficiaries, each heir's share of your inherited pre-tax retirement balances (traditional 401(k)/IRA and HSA) is taxed at that heir's own income-tax rate. We tax the full amount as if received in one year rather than spread over the ten years the law allows, which can reach higher brackets — so the tax is estimated on the high (conservative, safe) side.

In the stock-option alternative-minimum-tax calculation, your retirement-years AMT now keeps the itemized deductions AMT allows (home-mortgage interest, charity, and medical) and adds back only state and local taxes. Your working years still add back the whole standard deduction — exact for a standard-deduction filer, but it can overstate AMT income for someone who would itemize then.

Maryland's 2% capital-gains surtax applies once your federal income clears $350,000. We let a self-employment loss reduce that income (capped by the federal excess-business-loss limit), and carry the disallowed part forward so it also lowers a later profitable year, not over-applying the surtax when it brings you back under. Any remaining effect only overstates your tax, never understates it.

For a Maryland resident we model the mandatory county income tax at a representative 3.20% — the highest rate the law allows. Real county rates run about 2.25% to 3.20%, so for a lower-rate county we slightly overstate the tax (never understate it) until you enter your own county's rate.

For qualified small-business stock, when you don't give us the sale date we round the holding period down to whole years, which can understate the exclusion tier you qualify for and so overstate the tax, a cautious direction. Give us the sale date to compute your exclusion tier to the day.

When you move to a new state without giving the month, the move year is taxed at whichever state's tax is higher — the safe side, so it may run high; giving the month switches it to an exact split. Washington's excise, Maryland's surtax, a state deduction for federal estate tax on an inherited IRA, and a state 529 tax break all work this way — the move year gets the smaller break.

Rental depreciation now takes a half-year in the purchase year (the unbiased average of the tax rules' part-year convention, since the plan carries no purchase month), fixing the former buy-year over-depreciation. A remaining sale-year timing gap under-depreciates in the year of sale, which can slightly over-state taxes (never under-state wealth).

IRS rules let you group properties into one activity so their hours count together toward the material-participation tests; we don't model that grouping, so each property is tested on its own hours. If grouping would have let you clear the bar, we may treat a property as passive — keeping its income in the 3.8% investment tax and deferring its loss. This runs cautious, never in your favor.

When you enter the hours you work on a rental, we apply the IRS 500-hour material-participation test. The other six ways to qualify (comparisons with others' hours, prior years of involvement, and a facts-and-circumstances judgment) aren't modeled, so a rental you're active in another way may be treated as passive, keeping its income in the 3.8% investment tax and lowering its business deduction.

When you inherit a pre-tax retirement account from an estate that owes federal estate tax, we deduct that estate tax against your federal income tax on the account, and against your STATE income tax when your state allows the deduction (most itemizing states do). A few states that allow it only through a credit or a dollar cap are treated conservatively (no state reduction).

State death taxes are now modeled for every state that levies one, including the heir-relationship inheritance tax in PA, NJ, KY, NE and Maryland's inheritance layer. A couple of narrow edges are handled in the safe direction only (a young child's exemption, a child's spouse), so for those heirs the inheritance tax can be slightly over-stated, never understated.

In the bracket-modeled states, a capital gain is taxed at the state's top marginal rate rather than folded into the graduated brackets, which tends to overstate the state tax on the gain.

We model Michigan's capped retirement deduction and Georgia's and Colorado's per-person, age-based exclusions exactly — now including a genuine two-pension couple, where each spouse's own pension counts against their own cap. New York isn't yet modeled (its government-pension exemption needs a pension-source detail we don't collect), so it's taxed in full — conservative.

We model how Connecticut, Minnesota, New Mexico, Rhode Island, Utah and Vermont — and Colorado at 55-64 — exempt Social Security below an income threshold, now including Minnesota's better-of-two methods and Utah's real credit phase-out. Two conservative simplifications remain: an age test read from the primary spouse (Rhode Island, Colorado) and some unconfirmed MFS thresholds.

Washington allows a deduction when a qualified family-owned small business is sold (RCW 82.87.070). The engine taxes the full business-exit gain without that deduction, so a household selling a qualifying family business can be OVER-taxed by the Washington excise (a conservative, safe-direction error).

Washington's capital-gains excise applies to long-term gains only. Because the accumulation turnover legs carry no holding-period, all such gains are treated as long-term and taxed, so a household with short-term turnover gains can be slightly OVER-taxed by the Washington excise (a conservative, safe-direction error).

Your Coverdell ESA's funding is modeled to the §530 rules — the $2,000 contribution limit, the income phase-out, the age-18 contribution cutoff, and tax-free growth — but because a Coverdell never appears on a federal tax return, it isn't cross-checked against an outside calculator. (The rule that an unused balance be paid out by age 30 is not yet modeled; see that disclosure.)

We compute the 2026 SALT and mortgage-interest caps exactly to the new law and adjust them for inflation; this is not an approximation. Because the law is too recent for our external tax reference to cover, we pin the values to the published statute rather than validating them against an outside source.

On a small-business-stock sale we now determine, by state and by the year of sale, whether your state taxes the federally-excluded gain — a handful decouple (California, Pennsylvania and others; Hawaii taxes half), most do not — and model it exactly to each state law. Because our tax reference is federal-only, this state result is pinned to the published statutes rather than validated.

We decide real-estate-professional status by adding your hours across all your rental properties and applying the IRS tests — over 750 hours in real estate and more than half your total work time. Because the calculators we check against treat each property's active-or-passive status as an input rather than deriving it from hours, this is disclosed rather than independently verified.

Give the month and the move year splits exactly like a part-year resident's return — each state's tax prorated by the share of the year you lived there, so it's not an approximation. A state deduction for federal estate tax on an inherited IRA and a state 529 tax break split the same way. A two-state return isn't covered by our federal-only reference and assumes even income, so it's disclosed.

We compute the new senior deduction exactly to the 2025 law ($6,000 per eligible person, phased down above $75k/$150k of income, inflation-adjusted); this is not an approximation. Because the law is too recent for our external tax reference to cover, we pin it to the published IRS guidance rather than validating it against an outside source.

Washington's capital-gains excise standard deduction is inflation-indexed; we apply the most recent published figure ($278,000 for 2025) as a real-dollar constant, which is the standard treatment for an indexed threshold. The 2026 figure was unpublished as of September 2026, so the exact new number can't be verified yet, and state tax isn't checked against our federal-only reference.