Methodology

What the model computes, in the order it computes it. Enough to reimplement it, disagree with it, or find a mistake in it.

What is compared

An option is a way of covering you for the whole horizon. For one person the options are the three plans, and, if you switch it on, the HDHP for some years followed by a copay plan. For a couple they are every way two people can be covered from the same three plans: each on your own plan in any combination, one plan covering both at its employee + spouse rate, and the two all-HSA arrangements followed by a switch. When the two of you need the same care, “you on POS, your spouse on Kaiser” and the other way round cost exactly the same, so they are listed once.

If your spouse's employer offers its own plans, each of you chooses from your own employer's three: nine pairs, plus any of the six plans covering both at its employee + spouse rate, plus the three all-HSA arrangements followed by a switch — up to eighteen options. Nobody can join the other employer's plan alone, only as the employee's spouse. Both menus can hold a Kaiser plan, so every plan is then named after its employer — “Emory's Kaiser covering both”, “Walmart's Kaiser covering both” — from the two names you enter.

Every option is ranked. The best four are charted and given columns in the tables; for one person that is all of them.

The one rule

Every option is priced against the same care, and every option's household spends the same take-home pay on health each year. Whichever option costs least in a year leaves money over, and that money is invested in an ordinary taxable account. An HDHP household's own HSA contributions are part of what it spends, and the HSA it builds is part of what it owns.

The option that leaves the most is the one that cost least, with every dollar dated. The model checks that lifetime spending matches across every household and refuses to report a result if it ever diverges.

What is reported: net cost

Wealth is how options are compared, not a number to read on its own. It depends on the budget, and the budget is set by the costliest option, so anything that makes that option costlier — needing more care, above all — raises every household's wealth, the healthy plans' included. Wealth from two scenarios cannot be compared. So the page reports net cost instead:

The budget cancels out. Net cost ranks the options exactly as wealth does, with exactly the same gaps, and moves only when something about that option changes. Below zero, the HSA is worth more than everything spent on health.

“The most any option could cost” means its premiums, the full IRS contribution if it is on an HDHP, and all of its care paid from pocket, whatever the household actually does. Any shared budget gives the same gaps between options and the same net costs.

One year is the familiar number

With one year and nothing growing, an option's net cost is exactly its yearly net cost — the number most plan comparisons stop at:

An integrity check holds the two together. Money left in the HSA at the end of the year is still yours, so it is never counted as a cost.

A year of care

A year is a count of uses of eight services — primary care, specialists, urgent care, the ER, labs and imaging, generic and brand prescriptions, and hospital stays — each at a price. The price is the allowed amount: what the insurer has negotiated, which is what you pay before the deductible. It is not the provider's list price. Preventive care is left out because the ACA makes it free on every plan, high-deductible plans included.

The same year is priced on every plan. “How much care” multiplies every count at once, which is what the solver and the sensitivity grid move. A spouse's year is yours multiplied by “your spouse's care”.

Getting older

Most people need more care as they age. “Care growth with age” raises every count by that much each year, on top of prices rising, so the year of care at 50 is heavier than at 35. Because a heavier year is not just a bigger bill — it crosses deductibles and maximums — every year is priced on its own rather than scaled. The 2.5% default is an assumption, not a measurement: CMS's 2020 figures put spending per person at about $9,150 for working-age adults and $22,356 at 65 and over, which is the direction, not the rate, for any one person.

Switching plans part of the way through

The switching option is the HDHP for the years you choose, then the plan you choose for the rest of the horizon. From the switch:

• Contributions stop, yours and the employer's. The law allows them only while an HDHP covers you.
• The HSA does not close. It stays invested at the same return, and can pay for qualified care on the new plan, tax-free — if “pay care from the HSA” is on, it does, until it runs out.
• Receipts keep counting: care paid from your pocket on the new plan can still be reimbursed from the HSA later.

The option is the same as staying on the HDHP until the switch, so it can only beat staying put if the copay plan is cheaper by then, which is what care growing with age is there to test.

Couples

Each person's care is priced on the policy that covers them. A plan covering both pools their care against its family deductible and family maximum, both treated as aggregate — Emory's are — and charges its employee + spouse premium. Two people each on their own plan pay two employee-only premiums and meet two separate deductibles.

On HDHPs, each self-only plan brings its own HSA, employer deposit and self-only contribution; one family HDHP brings one HSA with the family amounts. The two accounts of a couple are pooled in the model, because an HSA can pay a spouse's qualified care whatever plan the spouse is on, and both grow at the same return.

Your spouse's plans are yours by default, as two people at the same employer would have. Set “your spouse's plans” to their own employer's and they choose from a second menu of the same shape — an HSA plan and two that are not — with its own premiums, deductibles, maximums, copays and HSA deposit. Their HSA plan brings its own self-only or family limit, exactly as yours does, and “contribute the IRS maximum” applies to both. On a switch, you move to your copay plan and your spouse to the one of theirs you pick.

How each plan prices that year

For each service with count n and price p:

• If the plan has a copay c for it, you pay n × min(c, p). A copay never costs more than the service itself.
• Otherwise n × p joins the amount that goes through the deductible. On the HSA-qualified plan, every service goes this way: that is what makes it HSA-qualified.

Plans often charge more on the ER and hospital stays than on other care, so each plan carries two coinsurance rates. Real bills meet the deductible in the order care happens; the model shares the deductible between the two kinds of care in proportion to what each costs, which keeps the bill continuous and never lets more care cost less.

Copays are assumed to apply from the first visit, and hospital copays are per admission. Some plans charge copays only after the deductible, which this model does not separate.

A health reimbursement account

Your spouse's third plan can carry an HRA: money the employer puts in each year to pay your share of the bill before you do. It cannot be invested and is forfeited on leaving the plan, so it is not an account here but the employer paying part of the bill. The rules are Walmart's:

The HRA pays everything but prescriptions. Counting the prescriptions first, rather than last, keeps more care from ever costing less once the maximum binds. What is left rolls over, but never past the deductible, and the allocation rises with medical costs like every other term of the plan. A switch onto or off the plan starts the balance again from nothing.

Tax

Employer plans almost always take premiums and HSA contributions out of pay before tax. A pre-tax dollar then skips your marginal income tax — federal plus state — and the 7.65% Social Security and Medicare tax. Above the Social Security wage base ($184,500 in 2026) only the 1.45% Medicare part is saved, which you can set. Switch payroll off and premiums are paid after tax, while the HSA contribution is deducted on your return and saves income tax only.

Premiums get the same saving on every plan, so it does not favour any of them: it shrinks the difference between premiums by your tax rate, which is why a $120 a month premium gap is worth less than $1,440 a year.

How much goes into the HSA

By default, the most the IRS allows. The limit covers employer and employee money together, so your own contribution is the limit less the employer deposit — $4,500 less Emory's $300, or $4,200, for one person in 2027; $9,000 less $600, or $8,400, on a family HDHP. Two people each on their own HDHP each fill a self-only limit; on Walmart's, that is $4,500 less $350, or $4,150.

The limit is indexed to inflation, so it is grown at the HSA contribution growth rate, as is the deposit. From 55 the $1,000 catch-up is added; it is set in the statute and not indexed. A couple on one family HDHP can make two catch-ups — the second goes into the spouse's own HSA — and both of you are taken to be the same age. Untick “contribute the IRS maximum” to enter an amount of your own instead.

Where the money goes over the years

Each year, in this order:

1. Premiums for the year are paid, grown by premium growth.
2. The employer deposit and your contribution go into the HSA, grown by HSA contribution growth, which roughly tracks the IRS limit.
3. The year's care is priced at that year's prices. Deductibles, maximums and copays rise with medical costs, because insurers reset them in line with prices; an integrity check prices a later year directly to hold that true.
4. If the HSA pays for care, the year's bill comes out of it, as far as the balance goes. Anything paid from your own pocket instead becomes a receipt: a qualified expense you can reimburse yourself for, tax-free, at any later date.
5. Whichever household spent least invests the difference to the year's budget.
6. Both accounts grow for the year. The HSA earns the market return less its fees, untaxed. The taxable account earns the same return less a yearly tax drag on dividends and distributions.

Two ways to use an HSA

Pay care from the HSA is what most people do, and it is the default. Pay from your pocket and keep the receipts leaves the whole HSA compounding, and because a receipt can be reimbursed years later it gives up nothing but the cash flow. How much that is worth depends on your numbers, which is why the scenario table always shows the other strategy alongside yours.

What the HSA is worth at the end

An HSA is not cash until it is spent, so it is valued for what it would actually buy:

• Up to your unreimbursed receipts, it can come out tax-free at any time: face value.
• Of the rest, the share you eventually spend on medical care is also tax-free. Medicare premiums count, and most retirees spend far more on care than a typical HSA holds, so the default is 100%.
• Anything else comes out as ordinary income, plus a 20% penalty before 65.

The taxable account is charged capital-gains tax on its gain as if it were sold at the horizon. That is the conservative end: never selling would charge less.

The horizon stops at 65

Medicare usually replaces employer coverage at 65 and ends HSA contributions, so the model will not price a run of years past it. The HSA itself carries on, and the value above is what it is worth when you get there.

The one-year crossover

The second chart scales your year by a factor from zero upward, multiplying every count while keeping the mix of services, and plots each plan's net cost for one year against the total care at full price. It runs far enough for every plan to reach its maximum, up to $90,000 of care. Crossings are found by scanning 600 steps for a change of sign in the difference between two plans, then bisecting. Two plans can cross twice: a copay plan often overtakes the HDHP in the middle of the range, and the HDHP wins back the far end once both hit their maximums.

One answer is never the whole answer

The scenario table

Nine runs of years, each changing one thing: a light run (a quarter of your care every year), yours, a heavy run (twice yours), a hospital stay in year one, every year at the maximum, markets three points weaker and three points stronger, the HSA left in cash, and the other way of using the HSA. Plan terms, tax and the horizon are held exactly as entered. Every option is priced in every row; the columns show the charted ones, and the best of all of them is named.

The badge under the verdict says how much of the table agrees with the leading option, in the same bands every calculator here uses: Robust at 80% or more, Assumption-sensitive from 60%, Finely balanced below that or when two options split the table between them.

Where the line is

A grid of the winning option across how much care you need (none to four times yours, columns) and the investment return (four points either side of yours, rows). More care usually favours the copay plans. A higher return is often said to favour the HSA, but what it really does is magnify the lead of whichever household saves more each year: on Emory's plans that is the Kaiser household, so a stronger market widens its lead. The caption names the care level at which, at your own return, a different option takes over. Cells carry the option codes from the ranking table.

Calling a tie

Two options closer than the larger of $150 or 2% of what the leading option spends on health over the horizon are reported as a tie. Plan prices change every year and nobody's care matches an estimate, so a smaller gap is noise. The strength badge measures the margin on the same scale. It is the leader's own spending, not the budget every household is held to, because the budget is set by the costliest option, and one bad option — a POS plan covering a couple, say — would otherwise make every real gap look smaller.

What would have to be true?

For each of a dozen inputs — how much care, each premium, the HDHP deductible, both HSA amounts, the return, your tax rate, both growth rates, care growth with age and the horizon, plus your spouse's care, the employee + spouse premiums, your spouse's own employer's premiums and the years before switching when those are on — the solver holds everything else fixed and finds the value at which the runner-up overtakes the leader, by bisection. Rows are sorted closest first. A separate ranking nudges each input by a plausible amount of error and reports which moves the answer most; joined with the thresholds, that names the one number worth checking with your benefits office before you choose.

Sampling the years instead of guessing them

Nobody gets sick exactly on average, and a year with two ER visits and a year with none do not average out to a year with one, because every deductible and maximum bends the cost. So the uncertainty analysis samples each year of each run:

• Each service count is drawn from a Poisson distribution around your expected count.
• Each price is multiplied by a mean-preserving lognormal factor, exp(σZ − σ²/2), so the average price is still yours.
• An unplanned hospital stay is added in exactly round(chance × runs) of the runs in every year, shuffled, so a rare event is never over- or under-sampled by luck.
• The market's return that year is drawn around yours. The HSA and the taxable account see the same year.

The generator is seeded, so the same scenario always gives the same table and a shared link reproduces it. Every sampled run goes through the same ledger as the calculator itself, and runs closer than the tie margin count as ties, not wins.

Checks that must pass

On every build, and published on the model integrity page:

1. What each plan charges matches a separate implementation written from the plan rules, at a quarter, one, two and a half and ten times your care.
2. No plan charges more than its out-of-pocket maximum, and a five-million-dollar year charges exactly the maximum.
3. More care never makes a plan cheaper.
4. Every household spends the same, and each taxable account rebuilds from its deposits.
5. The HSA balance rebuilds from its deposits, withdrawals and growth.
6. One year at a zero return: each option's net cost is exactly its yearly net cost.
7. A later year priced directly matches the inflated bill.
8. A couple on two employers' plans is priced on the right policies, and an HRA's balance rebuilds year by year from its rules.

The IRS limits

From the revenue procedures themselves: Rev. Proc. 2025-19 (2026) and Rev. Proc. 2026-24 (2027). The catch-up and the 20% additional tax are in IRS Publication 969, and the $184,500 Social Security wage base in Publication 15. The warnings check plan year 2027, the year chosen at this autumn's open enrolment. Entries that break these rules raise a warning but are still calculated, because plans change and you may know something the table does not.

The default plans

The three plans the calculator opens with are Emory University's for 2027, from its benefits guide: employee-only coverage, the tier 1 network, a full-time employee in the $60,000–$100,000 salary band. Replace them with yours. A few of Emory's rules have no slot in this model and were entered as their equivalent at the default prices:

• The POS plan charges 10% on diagnostic labs with no deductible: entered as a $45 copay, 10% of the $450 lab price.
• POS prescriptions are coinsurance with a minimum and maximum per fill and no deductible: generic 10% ($10–$25), preferred brand 20% ($30–$75). At the default prices that is $10 and $75.
• After the deductible, the HSA plan charges 20% on preferred brands and caps each fill; the model charges its ordinary 10%.
• Urgent care copays are not in the 2027 guide. The POS figure is the 2026 Summary of Benefits' $25; Kaiser's is assumed to equal its $25 primary-care copay.
• The $425 wellness incentive is left out: it is offered on all three plans, so it does not separate them.

Your spouse's employer

When your spouse's plans are their own employer's, the calculator opens with Walmart's for 2027 at a Georgia work location: the Premier PPO, the Saver HSA and the Kaiser of Georgia Low Option HMO, network benefits, from the medical plan comparison on Walmart's enrolment site for that location. The PPO and HSA plan terms match the 2027 Associate Benefits Book (pages 92, 141 and 151), which does not cover the HMO. Premiums are the 2027 tobacco-free rates, per biweekly pay period on the enrolment site and entered a month at a time (× 26 / 12). Which plans Walmart offers depends on the work location, so replace them with what your spouse's enrolment screen shows. Where Walmart's rules have no slot:

• On the PPO and the Saver HSA the ER is the deductible, then a $300 copay and nothing more. It is charged the ordinary 25% after the deductible, as hospital stays are, which overstates a visit once the deductible is met. On Kaiser it really is 25%.
• The Saver HSA's employer money is a dollar-for-dollar match up to $350 ($700 covering both), not a deposit. Any contribution of $350 or more collects all of it, so it is entered as a deposit.
• Prescriptions are copays with no deductible on the PPO (generics up to $4, covered brands up to $50, for 30 days) and on Kaiser ($10 and $50). On the Saver HSA they go through the deductible and then cost the same copays; the model charges its 25% instead. The $0 tier is left out, because which drugs are on it is the formulary's business, not a price.
• Elsewhere, Walmart offers a Contribution HRA Plan instead of an HMO: $112.50 every two weeks ($380.30 covering both), a $1,750 / $3,500 deductible, no office-visit copays, and a $250 / $500 HRA. Enter it in the HMO's column with its HRA.

What is not modelled

• Out-of-network care, balance billing, and any difference in which doctors each network includes. For an HMO such as Kaiser, the closed network is the real trade, and it is not a number.
• Flexible spending accounts paired with the copay plans.
• Embedded per-person maximums inside a family plan; the family deductible and maximum are treated as aggregate. Children, and the employee + children tier.
• Copays that apply only after the deductible, and HDHPs that cover some chronic-condition medicines before it.
• An HSA-qualified HMO. Some employers offer a Kaiser plan that qualifies; model it as the HDHP.
• Switching more than once, or to the HDHP from a copay plan.
• Retirement accounts. The cheaper plans' savings go into an ordinary taxable account, which is the neutral alternative; a 401(k) would narrow the HSA's advantage on the growth, not on the way in.
• Anything that is not money: which doctors you can see, how long appointments take, and how much a bill you did not expect would hurt.