Why retirement healthcare is two problems, not one
Retirement medical spending has a sharp discontinuity in it at age 65, and any projection that ignores it will be wrong in both directions at once. Before 65 you are buying individual coverage on the open market — an ACA marketplace plan, COBRA continuation, or a retiree plan if your employer still offers one — at the oldest and most expensive age band the rating rules allow. Under the Affordable Care Act an insurer may charge a 64-year-old at most three times what it charges a 21-year-old for the same plan, and carriers price close to that ceiling. Those are the most expensive health insurance years of your life.
At 65 the cost structure changes completely. Part A is premium-free if you or a spouse have 40 quarters of covered employment. Part B carries a monthly premium set annually by CMS, with income-related adjustments layered on top for higher earners. You then choose between a Medigap supplement plus a standalone Part D drug plan, or a Medicare Advantage plan that bundles them. The total is normally well below a pre-65 marketplace premium, but it is not zero and it never stops.
The honest way to model this is as two annuities of different sizes running back to back, both growing faster than the general price level. This calculator does exactly that, keeps the phases separate in the output so you can see which one dominates, and discounts the whole stream so you end up with a savings target rather than an alarming headline.
To model the Medicare-side detail rather than a single annual figure, work the income surcharge with the Medicare IRMAA surcharge calculator, the drug benefit with the Medicare Part D phases calculator, and the supplement decision with the Medigap vs Advantage calculator.
The formula, and why the two rates matter more than the totals
Each year's cost is the appropriate base figure grown by compound healthcare inflation: Ct × (1+g)t, where t counts years since you retire and Ct switches from the pre-Medicare figure to the Medicare figure at the eligibility age. Sum those and you have the nominal lifetime total — the actual number of dollars that will pass through your accounts.
That nominal total is a poor planning target, because dollars spent in year 28 are not dollars sitting in your account today. Dividing each year by (1+d)t, where d is the return you expect on the money you set aside, converts the stream into a single lump sum. The two adjustments collapse into one ratio, ((1+g)/(1+d))t, and that ratio drives the answer.
Three regimes fall out of it. When your discount rate exceeds healthcare inflation the ratio is below 1, distant years shrink, and the present value comes in below the nominal total. When healthcare inflation exceeds your discount rate the ratio is above 1, so each later year contributes more than its cost at today's prices, and a 30-year horizon becomes punishing. When the two are equal the ratio is exactly 1: every year contributes its today's-dollar cost and the present value is simply the annual figure multiplied by the number of years. That last case is the cleanest way to check the tool — set both rates the same and confirm the arithmetic in your head.
Costs are charged at the start of each year, so year 0 is neither inflated nor discounted and the first year's cost equals the figure you typed.
Worked example: a couple retiring at 62 and planning to 92
Both stop working at 62, both reach Medicare at 65, and they plan to age 92. Their pre-Medicare cost is $11,000 each per year including out-of-pocket spending; their expected Medicare-era cost is $7,200 each per year. They assume 5% healthcare inflation and a 4% after-tax return.
- Set the horizon. 92 − 62 = 30 years funded, of which 65 − 62 = 3 are bridge years and 27 are Medicare years.
- Scale for two people. Bridge cost is $11,000 × 2 = $22,000 a year at today's prices; Medicare cost is $7,200 × 2 = $14,400.
- Grow the bridge years. $22,000 × (1 + 1.05 + 1.052) = $22,000 × 3.1525 = $69,355 nominal across ages 62, 63 and 64.
- Grow the Medicare years. Year 3 (age 65) costs $14,400 × 1.053 = $16,670. Each following year is 5% larger, ending at $14,400 × 1.0529 = $59,272 in the final year. Across all 27 years that sums to $911,323 nominal.
- Discount the stream. With g = 5% and d = 4%, the ratio 1.05 ÷ 1.04 = 1.009615 sits just above 1, so later years carry slightly more weight than their cost at today's prices. The present value lands at $521,025 against a nominal total of $980,678.
Read that lump sum as the amount which, invested at 4% after tax and drawn against these bills, funds the whole 30 years. It is 53% of the nominal total. The three bridge years contribute $66,637 of it — 12.8% of the present value from one tenth of the horizon — because they are spent first and therefore discounted least. That concentration is the most useful thing on this page.
How to read the result
Start with the bridge-years total against the Medicare-years total. If you retire at 62 or earlier, a small number of years carries a disproportionate share of the present value. That is also the part you can most easily change — by working to 65, by managing modified adjusted gross income so a premium tax credit survives, or by staying on a spouse's employer plan.
Next, set the lump sum against the rest of your portfolio. Healthcare is not an extra expense sitting on top of a retirement plan; for most households it is one of the largest line items inside it, and it is the one that grows fastest. If this number is a large fraction of projected assets, the plan is fragile in a specific way: the liability is indexed to medical inflation, which does not fall when markets do.
Treat the inflation assumption as the main sensitivity rather than the cost figures. Moving from 4% to 6% over 30 years shifts the present value far more than a $1,000 error in the annual estimate does. Run it at both and plan against the higher figure.
Two exclusions matter. This projection does not include long-term care, which is a separate and far more skewed risk — use the long-term care cost projection calculator for that. It also ignores the tax character of the account you draw from; a dollar of Medicare premium paid from a traditional IRA costs more than a dollar paid from an HSA, which is why the HSA triple tax advantage calculator belongs beside this one.
What $1,000 a year of retirement medical cost is worth today
| Years funded | g = 4%, d = 6% | g = 5%, d = 5% | g = 5%, d = 4% | g = 6%, d = 4% |
|---|---|---|---|---|
| 10 | $9,192 | $10,000 | $10,444 | $10,911 |
| 15 | $13,172 | $15,000 | $16,053 | $17,198 |
| 20 | $16,790 | $20,000 | $21,937 | $24,112 |
| 25 | $20,080 | $25,000 | $28,109 | $31,718 |
| 30 | $23,070 | $30,000 | $34,584 | $40,083 |
| 35 | $25,789 | $35,000 | $41,376 | $49,284 |
Each cell is $1,000 × Σ(t=0..N−1) ((1+g)/(1+d))^t, the same expression the calculator evaluates. The g = d column is exactly $1,000 × N, which is the check to run first.
The pre-65 bridge is where early retirement plans break
Retiring before 65 without employer coverage means buying an individual plan at the top of the ACA age curve. The premium tax credit can cut that sharply, but it is calculated from household modified adjusted gross income, so a large Roth conversion, a realised capital gain or a lump-sum pension election in those specific years can withdraw the subsidy and add five figures to the cost. Model the bridge years with the MAGI you actually expect in each one — the ACA premium tax credit calculator handles that arithmetic. COBRA is a genuine bridge but runs 18 months in most cases, which does not span a retirement at 62.
Assumptions and limits you should know about
- One inflation rate for everything. Premiums, deductibles and drug costs have not historically moved together. A single rate is a planning simplification, not a forecast.
- No long-term care. Custodial care is excluded because its distribution is different in kind: most people spend nothing and a minority spend enormously. Averaging it into an annual figure hides the risk you are insuring against.
- No IRMAA modelling. Income-related monthly adjustment amounts add to Part B and Part D premiums above set income thresholds and are recalculated each year from a two-year-old tax return. If your income sits near a bracket, add the surcharge to the Medicare figure yourself.
- Costs are per person and identical. A couple with different ages, health status or Medicare start dates should run the calculator once per person and add the results.
- No mortality weighting. Every year to your plan-to age is funded in full. An actuarial present value weighted by survival probability would be lower, but planning to a median outcome underfunds half of them.
Where this sits among the other numbers in a retirement plan
The single-figure retiree healthcare estimates published each year by large asset managers are lifetime totals for someone retiring at 65 today, usually excluding long-term care and often excluding dental and vision. They are a useful reference point and a poor plan, because they encode somebody else's retirement age, health status, income band and inflation assumption. The figure this calculator produces is worth more to you precisely because you supplied those inputs.
Treat the result as a liability sitting alongside your other retirement liabilities rather than as a savings goal in its own right. Most people fund it from the same portfolio that funds everything else, so the useful question is what fraction of the portfolio it consumes. The exception is a health savings account, the only vehicle in the US tax code that is deductible going in, untaxed while invested, and untaxed coming out when spent on qualified medical care — including Medicare premiums, though not Medigap premiums. If you have access to one, retirement medical cost is its best use.
Re-run the projection every few years. The two inputs that move most are your plan-to age, which rises as you do, and your Medicare-era cost, which you will eventually be able to replace with a real figure instead of an estimate. Households also carrying a disability or income-replacement need should look at the Social Security benefit taxation calculator, because taxable benefit income feeds the same MAGI that drives IRMAA.