Protecting Your Retirement Plan from the Cost of Long-Term Care

Most Americans carry insurance against events that will probably never happen to them. They insure automobiles against collisions that, for the typical driver, resolve in days and cost thousands. They insure homes against fires that, in the overwhelming majority of cases, never occur. They buy travel policies, umbrella liability, and identity theft protection — coverage for events whose annual probability sits well below one percent.

Then they leave uninsured the one event that is more likely to happen than all of those combined: the need for extended care.

This is not ignorance of the risk. It is a persistent underestimate of it — a mix of optimism, avoidance, and the widely held assumption that Medicare, personal savings, or family members will absorb the cost when the time comes. None of those assumptions survive contact with the published research, and the research on this subject is better than most people realize.

The Frame

The Comparison That Reframes the Decision

Set long-term care beside the risks a household already insures without hesitation. The comparison is not rhetorical — the frequency data are published by the insurance industry’s own statistical organizations.

Property and auto frequencies: Insurance Information Institute / ISO and Highway Loss Data Institute claim-frequency data. Care figures: ASPE / Urban Institute (2022, revised) and CareScout 2025 California medians.

A one-in-430 annual chance of fire compounds to roughly a seven percent chance across thirty years of homeownership. Nearly every one of those homeowners is insured. The care event, at better than even odds, is the one most of them face bare.

The Evidence

What the Independent Research Actually Shows

The conclusions on this page are not mine alone. They track what the federal agencies, the actuarial literature, and the congressional research offices have published — and where those sources disagree, I would rather show you the disagreement than pick the number that sells best.

The probability: two federal figures, honestly stated

You will see 70 percent quoted almost everywhere in this industry, including on pages I have written. It comes from an HHS analysis (Johnson, 2019) finding that 70 percent of adults who survive to 65 develop severe care needs before they die — counting care delivered by unpaid family members as well as paid providers.

The more recent HHS / Urban Institute projection (Johnson & Dey, 2022, revised) puts the figure at 56 percent for people turning 65 between 2021 and 2025 — measured at the stricter threshold that actually triggers benefits under a tax-qualified policy: help with at least two activities of daily living expected to last 90 days, or substantial supervision due to severe cognitive impairment.

Both are HHS figures. They measure different things. The 56 percent figure is the conservative one and the one that matters for insurance planning, because it is drawn at the benefit trigger. I use it throughout this page. The argument does not need the larger number.

The duration and cost, from the same source

The same projection carries the duration and cost figures, broken out by sex. Reading them together is what keeps a planning conversation honest: the probability, the length, and the price all come from one place.

ASPE Research Brief, Long-Term Services and Supports for Older Americans: Risks and Financing, 2022 (revised August 2022), Tables 1, 3, 5 and 6. Costs in 2020 dollars, national, and include only paid care.

Two cautions about that cost column, both of which the source states plainly. It is a national average across everyone, including the 44 percent who never develop a significant need — among those who do use paid care, the projected average is $245,400. And it is in 2020 dollars at national prices. California prices are materially higher, as a later section shows.

The same brief values the unpaid care that families provide at $204,000 per person among those receiving it — more than the value of all paid care. That is the part of the bill that does not arrive as an invoice. It arrives as a daughter reducing her hours.

A second duration source, and why the numbers move

You will also see a shorter set of duration figures in circulation — about three years on average, 3.7 years for women against 2.2 for men. Those come from the Administration for Community Living and are measured differently. They are not wrong; they answer a different question, and the gap between men and women is wider on that basis than on the federal projection used above.

It matters because the gap is often reported as a headline. In the ACL figures, women need care roughly 70 percent longer than men do. On the ASPE projection, the difference is closer to 45 percent. Either way, the direction is the same and the planning point is unchanged: a couple’s exposure is not symmetrical, and planning that treats two spouses as a single averaged life understates one side of the household. I use the ASPE series on this page so that every duration, probability, and cost figure comes from one consistent projection rather than three sources stitched together.

What the dementia statistic actually says

One widely quoted line holds that one in three seniors develops Alzheimer’s or another dementia. That is a misstatement of the Alzheimer’s Association figure, which reports that one in three seniors dies with Alzheimer’s or another dementia — a different measure, and a larger number than the incidence.

The figure that matters for planning is duration, not incidence. Survival after a dementia diagnosis commonly runs four to eight years, and a minority of cases run considerably longer. That range, not a decade quoted as though it were typical, is what a benefit period must be sized against.

The Payer Mix

Who Actually Pays for Long-Term Care

This is the question the whole subject turns on, and it has a published answer. The Congressional Research Service tracks the national payer mix each year.

Congressional Research Service, Who Pays for Long-Term Services and Supports? (2025), using 2023 National Health Expenditure data. Figures exclude unpaid family care, which researchers estimate would roughly double total spending if valued at market wages.

Read that table twice. The largest payer is a means-tested welfare program. The second largest is the family checkbook. Private insurance — and that 8.7 percent includes ordinary health insurance, not just long-term care policies — is a distant third. In the HHS projection, private insurance covers about five percent of an individual’s lifetime care costs, and only 3.6 percent of older adults have any care cost paid by insurance at all.

That is the second half of this page’s title, quantified. Fewer than eight percent of Americans over 60 owned a stand-alone long-term care policy as of 2022, on HHS’s own count. The risk is close to universal. The coverage is not.

The Gap

The Perception Gap, Measured

It would be reasonable to assume that a risk this well documented is widely understood. It is not. The University of Michigan National Poll on Healthy Aging surveyed adults aged 50 and older and found a very wide gap between the actuarial picture and the public one.

Nearly half said planning felt too far off to act on, and roughly the same share said they did not know how to plan even if they wanted to. In a separate HHS-commissioned survey, only a quarter of Americans aged 40 to 70 could correctly identify Medicaid as the program that pays the most for long-term care in this country.

That last finding is the one I encounter most often in practice. People are not refusing to plan. They are planning against a picture of the system that is simply inaccurate — and by the time the picture corrects itself, the underwriting window has usually closed.

The Arithmetic

What Care Costs in California

National medians understate this state considerably. These are CareScout’s 2025 California statewide medians, published March 2026, with the three-year change alongside.

CareScout Cost of Care Survey 2025 (published March 2026), California statewide medians. Memory care is not separately surveyed for California; the figure shown is my planning estimate and should be treated as such. Bay Area costs commonly run above these statewide medians — that is my observation from practice, not a published survey result.

On the inflation question, I want to be more careful than this industry usually is. Care costs have clearly outpaced general inflation over the past several years — more than 20 percent on a California private room since 2022. But CareScout’s 2025 survey recorded the first real slowdown since the pandemic, with most settings rising between one and five percent, against 9.2 percent the year before. Anyone telling you care inflation reliably runs at two to three times general inflation is quoting a period, not a law. Plan for costs that rise faster than your other expenses; do not plan on a fixed multiple.

Projected forward at a deliberately conservative three percent, today’s California private-room median of $15,178 a month becomes roughly $20,400 a month in ten years and roughly $27,400 a month in twenty. A plan built on today’s cost of care is not a plan for the year care is likely to be needed.

The Test

Four Conditions for Carrying a Risk Yourself

A risk can reasonably be carried on your own balance sheet when four things are true: the probability is low, the cost is bounded, the timing is foreseeable, and the worst case can be absorbed without changing the plan. Long-term care satisfies none of the four.

The self-insurance strategy works when the risk is low-probability and the cost is bounded. Long-term care is high-probability and unbounded. That combination is precisely what insurance exists to address.

The Exposure

What Self-Funding Actually Costs

At the California private-room median, the total cost of a care event paid entirely from your own capital looks like this. These are today’s dollars, before any inflation, and a couple’s exposure is not a remote scenario — it is the ordinary case in which both spouses live long enough to need care.

Costs do not halve because you are married. One year of care for two people at the California private-room median is $364,272.

Self-funding depends not only on having sufficient assets today, but on those assets growing fast enough to keep pace with the cost of care. If care inflates at four percent annually and a portfolio earns five percent after tax, the real cushion is one percent a year. A sequence of below-average investment years, or a market drawdown at the wrong moment, eliminates that margin entirely.

The Timing Problem

Care Events Do Not Arrive on Schedule

Care needs arise unpredictably — following a stroke, a fall, a cognitive diagnosis, or a gradual decline that accelerates without warning. The financial demands begin immediately and do not pause for markets to recover. Sequence-of-returns risk, already the central hazard of a drawdown portfolio, is amplified when a care event forces large, unplanned withdrawals in a down year.

This is also the point at which the alternative closes. Coverage is medically underwritten; the practical window for obtaining it runs through a client’s fifties and early sixties, and it narrows from there. A household that decides at 65 to self-fund, then reconsiders at 73 after a health event or a closer look at the cost projections, often finds that the coverage it might have bought is no longer available at any price. The underwriting window, once closed, does not reopen.

The Self-Insurance Question

Two Objections, Examined

Among financially sophisticated families — those with meaningful retirement assets, professional incomes, and careful financial habits — the idea of self-funding long-term care risk has real surface appeal. If you have accumulated enough, why pay premiums when you could simply pay for care yourself? It is a reasonable question, and it deserves a rigorous answer, because the assumptions inside it are more fragile than they appear.

“I can self-fund the risk and pay for it myself.”

Long-term care is usually framed as an asset question. It behaves as an income question. Consider a married couple with $2 million of income-producing capital supporting their retirement. If an extended care event consumes a large share of that capital, the income it was producing falls with it — and the healthy spouse is left with a fraction of the income the retirement was built around, at the point in life when they are least able to rebuild it.

The household may have had enough on paper. What it did not have was a structure that kept the income intact while the care was paid for.

Illustrative. A $1.5 million care event against $2 million of income-producing capital would reduce that capital, and the income it generates, by roughly three-quarters. Figures of this magnitude assume an extended claim beginning years from now, with care-cost inflation compounding in the interim.

“Average care is only two to three years — why insure for life?”

Coverage is not held against the average. It is held against the catastrophe. Buying long-term care coverage sized to the average is like buying fire insurance that covers small kitchen fires — the average is what happens to most people. The tail is what financially damages a family. The distribution matters more than the mean: many people need modest help for a short period, and a meaningful minority need years of paid, high-intensity care. It is that minority, not the average, that determines whether a household’s plan survives.

The Household in Question

Why Affluent Families Insure Anyway

The households I work with can generally write the checks. That is exactly why the question deserves better than a reflex answer in either direction.

The HHS projections break out by income, and the pattern is instructive. Among people in the top income quintile at 65, only 5.7 percent will ever receive Medicaid-funded care — the safety net is not coming. Their average lifetime out-of-pocket care cost is $75,400, the highest of any quintile. And roughly one in ten of them will spend more than $250,000 out of pocket, in 2020 dollars, at national prices. In California, on the cost table above, that tail is heavier still.

So the affluent household faces a real but manageable central case and a genuinely damaging tail. That is the textbook shape of an insurable risk: not the loss you expect, but the loss you cannot absorb without dismantling something you spent a career building. The question is never whether you could pay. It is whether paying from portfolio — liquidating in whatever market happens to exist in the year care begins, with the tax consequences that follow — is a more efficient use of capital than transferring part of the exposure at a known cost.

For some households, the honest answer is that it is not, and they should keep the risk. I say so when I think so.

Four Things This Page Will Not Relitigate

Each of these has its own treatment elsewhere on this site. In summary:

The Alternative

A Design That Returns Value Even If Care Is Never Needed

The most reasonable objection to this coverage is the oldest one: that premiums paid for years may buy nothing at all. That objection had real force against the traditional products, and it has a structural answer in the modern ones.

Asset-based designs — also called hybrid or linked-benefit contracts — are structured so that if care is never needed, a death benefit passes to beneficiaries instead. Some designs also keep a guaranteed surrender value accessible during life, subject to surrender charges and the contract’s terms, and any amount drawn reduces the benefits that remain. The capital does something either way. What it does, and on what guarantees, varies enormously between carriers.

One point belongs in plain sight rather than in a footnote: on most of these designs the care benefit is paid by accelerating the death benefit. Care and legacy are funded from the same pool, and every dollar drawn for care reduces what heirs receive. The premium is not forfeited. But you are not given both in full.

What transferring the risk costs

The case against self-funding is not complete until it is priced. The relevant comparison is not premium against zero — it is premium against the exposure the premium removes. The figures below are from one current illustration for a healthy couple, both age 50 in California, on a single-premium whole life chassis with a qualified long-term care rider.

Put plainly: the single premium is roughly what seven months of combined care would cost out of pocket today. In exchange, the insurer assumes a lifetime obligation — whether the care event lasts two years or twenty-five. Self-funding is the better arithmetic only if you know, with certainty, that neither spouse will ever need more than about seven months of care. No one knows that.

One illustration, one carrier, one age and health class, as of the date shown — not a quotation available to you, and not a case result. Long-term care benefit payments reduce the policy’s death benefit and cash surrender value. Guarantees are subject to the issuing carrier’s claims-paying ability. Premium deductibility for a business entity depends on the entity type and the policy’s tax-qualified status and should be reviewed before it is relied on. Your own figures will differ.

That variation is the whole reason an independent review is worth running. Comparing the same client across multiple carriers routinely produces materially different structures and materially different guaranteed benefits for comparable outlay. That outcome is not luck. It is what an independent multi-carrier analysis produces, as against selecting a single insurer’s product and accepting its terms. I do not publish client case results here, because the design that fits one client’s age, health, entity structure, and objectives tells you almost nothing about what would fit yours.

What a funded plan preserves

The Second Opinion

If You Already Own a Policy

A review is not only for households starting from nothing. Coverage bought ten or twenty years ago was sized against the cost of care as it stood then, and several of its terms are worth re-testing against the figures on this page. Here is what an independent review examines:

• The benefit period and the daily or monthly benefit, measured against current local care costs rather than the costs at issue.
• Whether inflation protection was elected, and at what rate — the single largest source of gap in older contracts.
• Whether the premium is guaranteed or subject to future increases, and what a rate increase notice offers as alternatives.
• The elimination period, and whether home care counts toward satisfying it.
• Whether the riders elected years ago still match today’s objectives, and whether ownership and beneficiary designations — including business or trust ownership — are still correct.
• Whether an older contract should be kept as it is, supplemented, or exchanged. Older policies are sometimes better than anything available today, and that is a finding worth having in writing.

The Part That Costs Nothing

A Policy No One Knows About Is Not a Plan

Whatever a household decides here — to transfer part of the risk or to keep all of it — one step belongs with the decision and costs nothing. Tell your spouse and your adult children what exists, where the documents are kept, and who is to act. Care decisions are made under time pressure, usually by someone else, and usually at the worst week of a family’s year. A plan that lives only in your own head is not available to the people who will have to use it.

The same applies to the decision itself. The households that handle a long care event well are not, in my experience, the ones with the largest balance sheets. They are the ones that made a decision while there was still a decision available to make.