What Long-Term Care Costs: Why the Average Misleads

Most writing about this leads with a national average annual cost. That number is close to useless for making a decision, for three reasons. Prices vary enormously by region, and the average blends situations that have almost nothing in common. And the thing that decides whether this breaks a plan is not the annual price at all. It is the duration.

So this page does the arithmetic in the order that actually matters — how long, then how much, then what it does to the rest of the plan — and shows you how to run it with your own numbers instead of a national one.

First, how long

These are the federal government’s own figures, from LongTermCare.gov, which is run by the Administration for Community Living at HHS. Worth knowing before you lean on them: that page was last modified in February 2020, so treat the durations as structurally sound and the underlying survey as some years old.

• Someone turning 65 today has almost a 70% chance of needing some type of long-term care services and supports in their remaining years.
• Women need care longer — 3.7 years, against 2.2 years for men.
• One-third of today’s 65-year-olds may never need long-term care at all. About 20% will need it for longer than five years.
• Averaged across everyone who uses any services at all: about three years.
• More people receive care at home, and for longer, than in facilities.

Read those together and the shape of the risk appears, and it is not the shape most people assume.

This is not a high-probability, moderate-cost event. It is a moderate-probability event with an enormous tail. A third of people spend nothing. The average is around three years. And one in five is in it for more than five. Planning to the average is planning to a number that will not describe you — you will land well short of it or well past it.

The number that actually decides it

Long-term care rarely breaks a plan because the annual cost is high. It breaks a plan because a long case removes money from the portfolio in the years when the portfolio has the least time left to recover, and it does so on a schedule nobody chooses.

That is the same mechanism as sequence-of-returns risk, arriving through a different door. And it usually arrives late in life, which is exactly when a withdrawal cannot be deferred and when there are fewest remaining years to earn it back.

So the question to run is not “can I afford the annual cost.” It is:

If this ran for five years starting at 82, what does the plan look like at the end — and what is my spouse left with?

Why the national average is the wrong input

Two reasons, and the second is the one people miss.

Prices are local, not national. Care is labor, and labor is priced regionally. A national average blends metropolitan California with rural markets where the same hours cost a fraction as much. Nobody buys care at the national average.

The averages blend different products. “Long-term care” covers unpaid family help at home, paid help at home for a few hours a day, assisted living, and skilled nursing — and those differ by multiples. Worse, most real cases move through them: a few hours a week of help, then daily help, then assisted living, then memory care. Averaging across the path hides the escalation, and the escalation is the whole financial story.

Running it on your own numbers

Four inputs, and you can get all of them in an afternoon.

1. Your local price for paid help at home, per hour. Call two or three agencies in your area and ask. This is the number that governs the early years, and it is the one people underestimate most, because forty hours a week of help at home is a full-time wage.
2. Your local price for assisted living, per month, and separately for memory care, which costs more. Ask the same way. Facilities publish rates and will quote them.
3. How many years to model. Model two: the average, around three years, and the tail, five or more. If the plan survives the tail, you have your answer. If it only survives the average, you have a different one.
4. Inflation on care, separately from general inflation. Care is wages, and wages in this sector have not tracked general inflation. Whatever assumption your plan uses elsewhere, this line deserves its own.

Then run it as a withdrawal, not as an expense. Take it out of the portfolio, in those specific years, and look at what remains — not at whether the annual figure fits the annual budget.

If it helps to have the numbers in one place, I keep them all in the Ultimate Retirement Guide.

The part most people leave out, and it is the biggest one

Run it for one spouse, and then look at what the other one has.

Almost every serious long-term care case in a couple is asymmetric. One person needs care. The other does not, and then outlives them. Three things happen at once, and they compound:

• The portfolio is smaller, by however much the care cost.
• Household income falls. One Social Security benefit stops — the household keeps the larger of the two, not both.
• The survivor files as single. The same income now runs through single brackets and single Medicare surcharge thresholds, so the tax bill on what is left goes up rather than down. That is the widow’s tax trap, and a long care episode is the most common way a household walks into it having already spent the buffer.

A plan that survives the care cost and leaves the survivor short has not survived it. That is the calculation, and it is the one almost nobody runs.

What is deductible, and what that is worth

The tax code treats this as medical care. Under 26 U.S.C. §213(d)(1)(C), “medical care” includes qualified long-term care services, and unreimbursed medical expenses are deductible to the extent they exceed a percentage floor of adjusted gross income.

Two things follow that are genuinely useful:

• In a heavy care year the deduction can be large, because the expense is large relative to the floor. A year with substantial care costs is a year to look hard at, not to assume is like every other year.
• Premiums for a qualified long-term care insurance contract are treated as medical expenses too, but only up to an age-banded cap set by 26 U.S.C. §213(d)(10). The cap rises with age and is indexed annually, so check the current year’s figure rather than an old one — and note it is a cap on what counts, not a credit.

You have to itemize to use any of it, which for many retirees means the deduction is worth nothing in ordinary years and a lot in a care year. That asymmetry is worth planning around and is easy to miss.

What do people get wrong about this?

• “I’ll plan for the average.” A third of people need none and one in five needs more than five years. The average describes almost nobody.
• “It’s a nursing home cost.” Most care is at home, and paid help at home for many hours a week is comparable to a facility.
• “The house covers it.” Perhaps — but if one spouse is still living in it, it is not available, which is precisely the case where the money is needed.
• “Our health is good, so this is unlikely.” The longest and costliest cases are cognitive rather than physical, and current physical health says little about them.
• “We can decide when it happens.” Some of it, yes. But the options that require underwriting stop being available once there is a diagnosis, so that particular door closes before the need arrives.
• “It’s one big number.” It is an escalating sequence, and modeling it as a flat annual cost understates the later years, which are the expensive ones.

What this page deliberately does not do

It does not tell you whether to insure, self-fund, or use a hybrid product, and it does not name a carrier or a policy.

That is not squeamishness — it is the line this site holds. The risk and the arithmetic are facts, and you are entitled to all of them. Which side of the decision to land on depends on your balance sheet, your family, your health and your preferences, and anyone answering that from a web page is guessing about you. Run the arithmetic above, and the decision gets much easier to have properly — with an advisor, and ideally before it is urgent.

Where to take it next

If you have not read the coverage half, start with does Medicare cover long-term care — the answer is no, and the reason is a single clause in the statute.

The survivor consequence is the widow’s tax trap, and the mechanism that makes a badly-timed withdrawal expensive is sequence-of-returns risk. For everything else healthcare costs in retirement, start at Healthcare in Retirement.