Long Term Care Insurance

Long-term care is the largest unhedged retirement expense most people fail to plan for. It is not medical care: Medicare, HMOs, and standard health insurance explicitly do not cover the custodial assistance an Alzheimer’s patient, a stroke survivor, or simply an 85-year-old with declining mobility actually needs. The cost lands either on a family member providing twenty-four-hour care — a load that breaks careers, marriages, and the caregiver’s own retirement plan — or on the household balance sheet, in six-figure annual outlays for years. The point of an LTC plan is to decide, while you are still healthy enough to choose, which of those outcomes you accept and how you fund it.

What actually triggers the benefit. A tax-qualified LTC policy pays when a licensed health care practitioner certifies you as a chronically ill individual under IRC §7702B, “Treatment of qualified long-term care insurance”. The statutory definition in IRC §7702B(c)(2) has two independent triggers, and most summaries — including most agents’ — give only the first:

1.

The ADL trigger. You are unable to perform, without substantial assistance from another individual, at least two of the six activity of daily livings listed below, and the condition is expected to last at least 90 days. The 90-day expectation is part of the statute, not fine print — a six-week recovery from surgery does not qualify no matter how dependent you are during it.

2.

The cognitive trigger. You require substantial supervision to protect you from threats to health and safety due to severe cognitive impairment — regardless of ADLs.

The second trigger is the one that matters most and is most often omitted. An early-to-mid-stage Alzheimer’s patient can frequently bathe, dress, eat, transfer, and toilet without assistance and is nonetheless exactly who this coverage exists for — because they cannot safely be left alone. If you read a policy summary that describes only the two-ADL test, that is a defect in the summary; if the policy covers only the two-ADL test, it is not a tax-qualified contract and you should not buy it. Certification must be renewed at least every 12 months.

The six activity of daily livings are:

Bathing

This includes washing oneself by sponge bath or in either a tub or shower, as well as the task of getting into or out of the tub or shower.

Dressing

This involves putting on and taking off all items of clothing and any necessary braces, fasteners, or artificial limbs. This does not include tasks related to bathing.

Eating

This refers to feeding oneself by getting food into the body from a receptacle (such as a plate, cup, or table) or by a feeding tube or intravenously. It does not include the preparation of meals.

Transferring

This means moving into or out of a bed, chair, or wheelchair. This does not include mobility outside of the home.

Toileting

This involves getting to and from the toilet, getting on and off the toilet, and performing associated personal hygiene. This does not include other activities that may be performed in a bathroom or lavatory.

Continence

This is the ability to maintain control of bowel and bladder function; or, when unable to maintain control, the ability to perform associated personal hygiene (including caring for a catheter or colostomy bag).

Collecting on the policy is a second skill, and someone else will need it. Everything above is about buying. The claim is a separate discipline, it runs for years, and by definition it is administered by someone other than the insured — which is why it belongs in the same conversation as the durable power of attorney and the named confidant of section “Late-Life Vulnerability: The Plan for Diminished Judgment”. Six things reliably surprise families.

The elimination period is usually counted in service days, not calendar days.

This is the costliest misunderstanding in the whole product. A “90-day elimination period” generally means 90 days on which paid, professional care was actually delivered. Days covered by a spouse, an adult child, or a neighbour count for nothing. If paid help starts at three days a week, ninety service days takes roughly seven months of paying out of pocket instead of three. Some contracts do use calendar days, and a few waive the period for home care — read which one you own before you need it, because it sets the size of the cash reserve the family must carry between diagnosis and first payment.

The carrier will run its own assessment regardless of your physician.

A neurologist’s certification on the attending-physician statement does not end the question; expect a contracted nurse assessor in the home, and expect that visit to be scheduled at the carrier’s convenience. Build the delay into the reserve above.

Invoices alone do not pay claims –- care notes do.

Carriers require contemporaneous documentation of what care was actually rendered (cueing meals, assistance dressing, overnight supervision), not merely hours billed. Home-care agencies do not always know this and will often send invoices only. Confirm in writing, with each agency, that care notes accompany every invoice, because days lacking them get discounted.

The paperwork must arrive from the provider, not from you.

Most carriers will not accept invoices submitted by the family. Ask each agency to copy you on every submission so you can verify what was actually sent.

Keep your own tally and check it monthly.

Miscounts run in the carrier’s favour more often than chance would suggest — the classic one is an overnight shift spanning two calendar dates being counted once. Reconciling invoices against payments takes an hour or two a month, indefinitely, and it is the family member three time zones away who ends up doing it.

On an indemnity policy, take the money yourself.

Agencies frequently offer to bill the carrier directly. On a per-diem contract that hands them the spread on every light-care day. Direct the benefit to the insured’s own account and pay the agency from it, so surplus days offset the twenty-four-hour ones. On a reimbursement contract this matters less, because there is no surplus to capture.

Note, families routinely cannot locate the contract when they need it — but the carrier has it, so call instead of searching the house, and expect the telephone queue to be more productive than email. And if the policy was placed by an agent who is still reachable, use them: an agent chasing an unresponsive claims department is free, and it is frequently the difference between a delayed claim and a lawyer.

The numbers worth holding in mind. Roughly 70% of people who reach age 65 will need some long-term care, but the median duration is around two years and the right tail is what matters: roughly 20% need care for more than five years, and a meaningful minority need it for a decade or longer. National medians run roughly: in-home aide care at $30–$35/hour, or $6,000–$7,000 per month at 44 hours a week and well over $20,000 for genuine around-the-clock coverage; assisted living near $6,000 per month, with memory-care wings 30–50% higher; skilled nursing near $9,000 per month semi-private and $11,000 private. California, the Northeast, and Alaska run 30–60% above those medians; parts of the South run below.

Take the national semi-private figure of $9,000 per month and inflate it at 4%, above headline CPI because care is a wage-driven service:

where $P_0$ is the current monthly cost and $g$ the annual escalation. Five years costs $540,000 flat and $585,000 escalated; ten years costs $1.08M flat and $1.30M escalated. Start the clock 15 years from now, when a 65-year-old reader actually needs it, and multiply by another $1.04^{15}\approx 1.8$. That is the unhedged liability your plan has to absorb, and it is why the self-insure threshold below is stated in millions, not hundreds of thousands.

The traditional-LTC product is damaged. Stand-alone LTC insurance — the kind that pays a daily benefit and nothing if you never need it — looked like a reasonable product in the 1990s and is not anymore. Insurers mispriced lapse rates (almost no one let their LTC policy lapse, contrary to actuarial assumptions) and assumed interest rates that the post-2008 environment never delivered. The result has been cumulative premium increases of 40–90% on existing policyholders, sometimes sequenced across multiple rounds with no contractual ceiling, and the exit of major carriers (Genworth, MetLife, John Hancock) from the new-policy market. A buyer of a standalone LTC policy today is buying both a benefit and a contingent obligation to absorb whatever premium hikes the carrier later determines it needs to remain solvent. Price that contingent obligation in, or do not buy.

The three real choices. The LTC plan is one of three structures, matched to balance-sheet size:

Self-insure.

At net worth meaningfully above the upper bound of plausible LTC outlay — roughly $3M+ outside the primary residence, more if you live in a high-cost state — the math usually favors funding the liability internally. Earmark an LTC sub-portfolio, keep it in conservative growth, and draw from it only if needed; if you never do, it flows to heirs intact. Jim Dahle’s essay on his father’s LTC experience walks through how this choice actually plays out for a high-income family.

Hybrid (asset-based) LTC.

A life-insurance policy or a single-premium annuity with an LTC rider. Pays an LTC benefit if you need care; pays a death benefit (life version) or annuity proceeds (annuity version) if you don’t. Premiums are locked at issue because the carrier is funding the benefit out of a single or fixed premium instead of betting on lapse assumptions — no surprise rate hikes. This has become the default recommendation in the $1M–$3M net-worth range, and for readers above that range who want certainty without committing to full self-insurance. The trade-off is a capped benefit pool (typically 2–4x the premium); coverage runs out where a standalone policy would not.

Traditional standalone LTC.

The narrow remaining use case is upper-middle-class balance sheets (roughly $500K–$1M) where there is enough to want to protect an inheritance but not enough to fund several years of care out of pocket. Even here, size the premium to absorb a 50% future increase without strain, and confirm the carrier holds an A or A+ AM Best rating with reserves to match.

Tax treatment, with the numbers. Premiums on a tax-qualified LTC policy count as a medical expense under IRC §213(d)(10), but only up to an age-based annual cap, and only to the extent total medical expenses clear the 7.5% AGI floor — which for this book’s reader usually means the deduction is worth nothing, because you are itemizing against a large standard deduction and 7.5% of a seven-figure AGI is a wall. The 2026 caps, per person:

Three practical takeaways follow: first, HSA balances can pay LTC premiums up to those same caps, tax-free and with no AGI floor — which converts a deduction most readers cannot use into a benefit every reader can. This is the cleanest use of an HSA in retirement and the reason to carry the balance into your seventies instead of spending it at 66. Second, a business owner can often do better than either: a C-corporation may deduct the full premium for an owner-employee without regard to the age caps, which is one of the few genuinely favorable owner-employee fringes left (chapter “The Business Owner’s Tax Architecture”). Third, benefits received under a tax-qualified policy are excluded from income under IRC §7702B(a)(2) and IRC §104(a)(3) — but indemnity (per-diem) policies are excluded only up to a statutory daily cap, $430 per day in 2026 under IRC §7702B(d), unless actual costs exceed it. Reimbursement policies that pay actual expenses have no such cap, which is a point in their favor for high-cost states.

Medicaid is a backstop, not a plan. Medicaid covers nursing-home care, and in many states a portion of in-home care, once the recipient’s countable assets have been spent down to the state threshold. For households without the assets to self-insure or fund private insurance, that is the catchment net. For households with the assets to self-fund or insure, relying on Medicaid is an actively bad outcome: it restricts your choice of facility, controls the level of care, and forces an asset spend-down before kicking in. Medicaid planning through irrevocable trusts is possible but constrained by the five-year look-back, which means assets must be moved out of your control five full years before care is needed. If Medicaid planning belongs in your plan at all, it belongs in the estate-planning conversation in your fifties, not your seventies.

California is the cautionary tale in building a plan on a policy: Medi-Cal eliminated its asset test entirely as of January 1, 2024 — and then AB 116 (2025) reinstated it effective January 1, 2026, at $130,000 for an individual and $65,000 for each additional household member. That is still far more generous than the $2,000 limit most states apply, but the whipsaw inside 24 months is the lesson. The income rules, the 30-month look-back for institutional care, and — most importantly — estate recovery against the probate estate all operate throughout, and the facility-choice and care-level objections above are unchanged. A Californian with the balance sheet this book assumes should still treat Medi-Cal as the outcome to avoid, not the plan to qualify for (section “Medicaid Asset Protection Trusts” tracks the statutory history).

Timing. Hybrid policies underwrite cleanly in your 50s and become meaningfully more expensive — or uninsurable — after a major health event. Self-insurance becomes feasible whenever your net worth crosses the threshold. Both are decisions to make while you are healthy and have time on your side; both become unavailable once you actually need care. The single failure mode worth naming is the reader in his sixties who put the decision off because nothing felt urgent, then receives a diagnosis that closes the underwriting window and discovers his balance sheet is too small to self-insure. Do not be that reader.