Long-Term and Social Care Economics

Why this matters in health economics

Populations are ageing almost everywhere, and with age comes dependency — the need for help with washing, dressing, eating, moving, and managing a household. This is long-term care (LTC): support, often for years, for people whose functional capacity has declined through age, disability, frailty, or chronic illness. Unlike an acute illness that is treated and resolved, dependency is a state to be sustained, and its cost accumulates quietly over a long tail. As the share of older people rises and birth rates fall, the dependency ratio shifts, and the question of who pays for care — and who provides it — becomes one of the defining fiscal problems of the century.

The stakes are large and concrete. LTC is expensive, its costs are catastrophic and unpredictable for the individual, and yet the private insurance market that should pool that risk has failed almost everywhere it has been tried. The public money at issue is growing fast, sits in more than one budget, and is fought over between them. And most care is not bought or sold at all: it is given, unpaid, by spouses, daughters, and sons, at a cost to their own earnings, health, and pensions that rarely appears in any account. A health economist who ignores this hidden economy is mismeasuring the true cost of an ageing society by a wide margin.

For a director, LTC is where the tidy boundaries of "the health system" break down. Decisions made in a hospital land in a care budget; a cut to home support fills a hospital ward; a family carer's collapse becomes an emergency admission. This chapter gives you the financing models, the valuation methods for unpaid care, and the boundary and integration problems you will have to manage — because in an ageing world, getting long-term care wrong is both a fiscal and a human failure.

Core concepts

Social care is the umbrella term, common in the United Kingdom and elsewhere, for personal and practical support with daily living — as distinct from medical treatment, though the two constantly overlap. Need is usually measured by limitations in the activities of daily living (ADLs) — bathing, dressing, toileting, transferring, continence, and feeding — and the instrumental activities of daily living (IADLs) such as shopping, cooking, managing money, and taking medicines. The number of ADL and IADL limitations a person has is the standard currency of eligibility and of costing: it grades dependency, drives the package of support, and predicts the cost.

Care is delivered along a spectrum of settings and intensities. At one end is home care — paid workers or family supporting someone in their own home. In the middle sit sheltered and supported housing and day services. At the other end is residential and nursing home care, where accommodation and round-the-clock support are combined. Dementia deserves special mention: it is the single largest driver of high-intensity, long-duration LTC need, it blends cognitive and physical dependency, and its costs fall heavily on families because so much of the care required is supervisory rather than clinical.

The defining economic feature of LTC is that it is a large, uncertain, correlated financial risk that markets insure badly. In principle, long-term care insurance should let people pool the risk of needing years of expensive care. In practice, the private market is thin and, in many countries, has all but collapsed. The reasons are the classic failures of insurance (see Chapter 1.3 — Market Failure): adverse selection, where those who expect to need care are keenest to buy, pushing up premiums and driving out the healthy; moral hazard, where insured people use more care; and long-horizon uncertainty, because insurers must guess decades ahead about future care costs, longevity, medical progress, and interest rates, and price that uncertainty into premiums people cannot afford. Add myopia and denial — people underestimate their own future frailty — and demand is weak on both sides. The market failure is not incidental; it is the reason the state is involved almost everywhere.

Because private insurance fails, societies fund LTC through one of a few broad models, and naming the system matters because they distribute cost very differently. The first is social LTC insurance, a form of social insurance in which mandatory contributions fund a defined benefit for everyone who reaches a dependency threshold. Germany introduced statutory LTC insurance (Pflegeversicherung) in 1995 as a distinct pillar of social insurance, paying graded benefits by assessed care level and deliberately allowing recipients to take a cash benefit and organize their own care. Japan introduced its own mandatory LTC insurance in 2000, funded by contributions from those aged 40 and over plus taxation, with benefits managed through local care-management. Both socialize the risk while sharing cost through co-payments.

The second model is means-tested, tax-funded safety-net provision, exemplified by England, where the state pays for social care only once a person's assets and income fall below defined thresholds; above them, people pay the full cost themselves. This concentrates public money on the poorest but exposes everyone else to potentially unlimited private cost — the "catastrophic cost" problem — and creates a hard cliff-edge around the means test. A third pattern, dominant across much of the low- and middle-income world, is implicit reliance on the family: little or no formal LTC financing, with dependency absorbed almost entirely by unpaid relatives, overwhelmingly women. Most systems are in fact hybrids of these three.

This points to the largest and most invisible part of the sector: informal care, the unpaid support provided by family and friends. A caregiver who is a spouse or adult child typically bears real economic costs — forgone earnings, reduced pension contributions, career interruption, and worse physical and mental health — none of which appears in a health-system account. Valuing this care work is essential to honest analysis, and there are two standard methods. The opportunity-cost method values a carer's time at what they give up — typically their forgone wage — capturing the loss to the carer and to the wider economy. The replacement-cost (or proxy-good) method values it at what it would cost to buy the same hours from a paid care worker in the market. The two can diverge sharply, and the choice can swing a societal-perspective evaluation, so it must be made explicitly (see Chapter 2.1 — Economic Evaluation).

Two further concepts complete the map. The health–social-care boundary is the administrative line between what a health budget pays for and what a social-care or welfare budget pays for — a line that patients and conditions ignore. Because the same person's needs straddle it, each budget has an incentive to define a need as the other's responsibility, producing cost-shifting: a decision that saves one budget by loading cost onto the other, with no gain, and often a loss, to the system as a whole. Integration is the attempt to overcome this by pooling budgets, aligning assessment, and organizing services around the person rather than the boundary. General equity in LTC — who deserves subsidy, and the fairness of asking families to pay — belongs to Chapter 3.4 — Equity; capability and wellbeing measures for older people, including the ICECAP-O instrument, belong to Chapter 3.5 — Capabilities; the distinctive economics of mental illness belongs to Chapter 3.9 — Mental Health Economics.

Best practices

1. Cost care over its whole trajectory, not as an episode. Dependency is a state sustained for months or years, so a snapshot cost understates the true burden badly. Model the expected duration of need, the likely progression through rising intensities of care, and the long tail of high-cost cases, ideally with a Markov-style model (see Chapter 2.2 — Modelling). A business case built on a single year of cost will misprice the commitment you are actually making.

2. Count informal care explicitly, and state the valuation method. Unpaid family care is often the largest single input in the system, and leaving it at zero makes a societal-perspective analysis meaningless. Decide up front whether you are using the opportunity-cost method (the carer's forgone earnings) or the replacement-cost method (the market price of equivalent paid hours), because they can differ substantially and can flip a result. State the method prominently, test the alternative in sensitivity analysis, and never let unpaid care vanish simply because no invoice was raised.

3. Choose the financing model deliberately, and name what each distributes. Social LTC insurance, means-tested safety nets, and reliance on family are not technical variants of one thing; they place the risk on contributors, on the individual, and on the family respectively. Be explicit about which risk you are socializing and which you are leaving with the household, and about the equity and gender consequences that follow (see Chapter 3.4 — Equity). Borrowing another country's mechanism without its financing base — as when a means-tested system aspires to universal outcomes — produces predictable shortfalls.

4. Do not expect the private LTC insurance market to rescue you. The near-universal thinness of the private long-term care insurance market is a structural result of adverse selection, long-horizon uncertainty, and consumer myopia, not a marketing problem to be solved with a better product. If you rely on it to fill a funding gap, model low take-up and high premiums as the base case. Where the state wants private cover to play a role, design the risk-pooling and the public backstop that make it viable, rather than assuming the market will self-correct.

5. Design the eligibility threshold with its cliff-edge in mind. Whether eligibility turns on a means test, a dependency level measured in ADL limitations, or both, the boundary creates sharp incentives and hard cases just either side of it. A steep asset threshold can penalize saving and produce deep unfairness between two similar people on opposite sides of the line. Consider tapers, caps on lifetime cost, and disregards that soften the cliff, and cost the behavioural responses — asset transfer, gaming of assessments — that any threshold invites.

6. Attack cost-shifting at the boundary, not just within each budget. Because the health–social-care line is administrative, each side can save money by reclassifying a need as the other's, with no benefit to the person and often a worse outcome. Track where cost lands across both budgets together, and treat a "saving" that merely moves cost across the boundary as a failure, not a success. The classic pattern — a delayed hospital discharge caused by no available care package — is a boundary failure that a single-budget view will never diagnose.

7. Invest in supporting carers as economic infrastructure, not as welfare. Family carers are the load-bearing structure of the whole system; when a carer's health breaks or they leave the labour market, the state usually inherits a far larger bill. Respite, carer's allowances, training, and flexible-work support are best appraised as investments that protect the informal-care supply, with the avoided formal-care and health costs on the benefit side. Costing the carer's own forgone earnings and worse health makes the return on this support visible.

8. Prioritize reablement and prevention, but honour the long lag. Home adaptations, falls prevention, rehabilitation, and reablement can delay or reduce dependency and are frequently cost-effective. But the outcome — a care admission that does not happen — arrives years after the spend and in a different budget, so standard short-horizon return-on-investment tools understate the value and the accountability is split. Use long time horizons, model the avoided downstream care, and align the budget that pays with the budget that benefits, or the investment will be perennially crowded out.

9. Value outcomes with measures built for older and dependent people. For much LTC, the goal is not to extend life or cure disease but to sustain dignity, control, and quality of daily living, which the standard quality-adjusted life year captures poorly. Use social-care-specific and capability-based measures — capability and wellbeing instruments for older people such as ICECAP-O are developed in Chapter 3.5 — Capabilities — so that "a good day" and independence count in the evaluation. Measuring only health outcomes will systematically undervalue care that works.

10. Integrate around the person, and pool the budget to match. Fragmented assessment, records, and funding force dependent people and their carers to navigate seams that serve the organizations, not them. Aligning or pooling health and social-care budgets, sharing a single assessment, and organizing services around the individual reduce cost-shifting and duplication — but integration has its own transaction costs and rarely saves money quickly, so promise better coordination and outcomes rather than instant savings. Evaluate it honestly, because the evidence on cash savings from integration is mixed.

11. Plan the paid care workforce as a binding constraint. LTC is labour-intensive, often low-paid, and increasingly short of workers as demand rises and other sectors compete for the same people (see Chapter 3.6 — Health Workforce and Labour Markets). Wages, migration policy, training, and turnover are not side issues; they set the ceiling on how much formal care can actually be delivered at any price. A financing plan that assumes an infinitely elastic supply of care workers will fail in the labour market before it fails in the budget.

12. Anticipate demographic and family-structure change, not just today's numbers. Falling birth rates, later childbearing, higher female labour-force participation, and geographic dispersion all shrink the future supply of family carers even as the number of dependent people rises. Project the dependency ratio and the availability of informal care together, because a plan that quietly assumes daughters will keep providing unpaid care at historical rates is planning for a world that is disappearing.

Questions to discuss with your team

1. Whose responsibility is long-term care — the individual, the family, or the collective — and does our funding model match the answer we would defend? Every LTC system encodes an answer to this, usually without stating it: a means test says the individual pays until impoverished, social insurance says contributors share the risk, and silence says the family absorbs it. Surface your team's real position by testing hard cases — a lifelong renter versus a homeowner, an only child versus a large family, a woman who left work to care versus one who did not — and ask who you are comfortable asking to pay. Distinguish what your budget currently does from what you would defend in public, because they are often not the same. Be explicit about the gender consequences, since "the family" overwhelmingly means women. An honest answer commits to a principle, accepts who it disadvantages, and checks that the money actually flows the way the principle implies.

2. What is the true cost of the care we are planning, once unpaid family care is counted? Most LTC business cases quietly value informal care at zero, which flatters any option that shifts work onto families and penalizes any option that pays for it. Work a concrete comparison: a package that keeps someone at home with heavy family input versus a residential placement, costed once with unpaid care at zero and once with it valued by opportunity cost and by replacement cost. Notice how the ranking can change, and discuss which perspective — health system, wider public sector, or whole society — your decision should take. An honest answer names the perspective, states the valuation method for unpaid care, and does not treat a carer's forgone job and health as free just because no invoice arrives.

3. Where is cost-shifting happening across the health–social-care boundary, and are we rewarding it? When health and social care hold separate budgets, each can improve its own position by pushing cost or blame across the line, and performance systems often reward exactly that. Map a common journey — a frail older person admitted, treated, and stuck awaiting a care package — and follow where cost and delay actually land in each budget. Ask whether any recent "saving" on one side simply reappeared as a larger cost on the other, or as a worse outcome for the person and their carer. Consider whether pooled budgets, joint assessment, or shared accountability would change the incentives, and be honest that integration costs effort and rarely saves money fast. An honest answer measures across both budgets at once and stops calling a transfer a saving.

4. Are we appraising prevention and reablement over a horizon long enough to see the return, and does the budget that pays also benefit? Home adaptations, falls prevention, and reablement can delay or avoid dependency, but the payoff — a care admission that never happens — arrives years later and usually in a different budget from the one that spends. A short-horizon return-on-investment tool will therefore understate the value and hand the accountability to no one, so these programmes are perennially first to be cut when money is tight. Interrogate the time horizon your business cases actually use, and whether the budget bearing the cost is the one that reaps the avoided downstream care. Discuss whether the modelled downstream saving is real and cashable or merely notional, and how you would hold it to account across a decade and across organizational boundaries. An honest answer commits to a long horizon, models the avoided care explicitly, and either aligns the paying and benefiting budgets or names the split as a risk to the investment.

5. Have we planned the paid care workforce as a binding constraint, or are we assuming staff will simply appear? Long-term care is labour-intensive and often low-paid, and every financing plan implicitly assumes a supply of care workers that the labour market may not deliver as demand rises and other sectors compete for the same people (see Chapter 3.6 — Health Workforce and Labour Markets). A scheme can be fully funded on paper and still fail because the hours cannot be staffed at the wage on offer. Examine what your plan assumes about pay, recruitment, turnover, training, and any reliance on migration, and stress-test it against a tight labour market rather than an infinitely elastic one. Ask who bears the cost of a wage rise needed to fill vacancies, and whether the money follows. An honest answer treats the workforce as a hard ceiling on delivery, budgets for the wages and turnover that ceiling implies, and does not quietly assume care will be provided at any price.

6. Are we measuring what matters to dependent people, or only the health outcomes our tools happen to capture? For much long-term care the goal is not to cure or extend life but to sustain dignity, control, and a good daily life, which the standard quality-adjusted life year measures poorly and can miss entirely. If your evaluations count only health gain, you will systematically undervalue care that works and over-reward interventions that happen to move a clinical measure. Discuss which outcomes your commissioning actually rewards, and whether capability and wellbeing instruments for older people — such as ICECAP-O, developed in Chapter 3.5 — Capabilities — and carer outcomes belong in the appraisal. Consider the carer's own wellbeing as an outcome in its own right, not merely an input to be preserved. An honest answer names the outcome measure it uses, admits what that measure cannot see, and adds social-care and capability measures where health outcomes alone would mislead.

In practice: a health economics example

The fictional middle-income country of Veranova is designing its first national long-term care scheme. Its population is ageing faster than its income is rising, urban migration is pulling adult children away from ageing parents in rural provinces, and women — the traditional carers — are entering paid work in growing numbers. Today, almost all dependency is absorbed by families, with a thin private nursing-home sector for the wealthy and no social care financing to speak of. The Ministry of Health and the Ministry of Social Welfare have been asked to recommend a financing model within a fixed medium-term envelope.

The team frames three options. Option A is a means-tested safety net: the state pays for care only for those with almost no assets, everyone else self-funds. It is cheap to the treasury but leaves the broad middle exposed to catastrophic cost, offers no risk pooling, and — because the private long-term care insurance market is virtually non-existent and, on the evidence from other countries, unlikely to emerge at scale — leaves most families no way to insure the risk. Option B is a social LTC insurance scheme on the German and Japanese pattern: a mandatory contribution funds a graded, defined benefit for anyone crossing a dependency threshold, with co-payments and a cash-benefit option so families can be paid for some of the care they already give. Option C is enhanced status quo: modest subsidies and respite for family carers, with no universal entitlement.

The analysts insist on a societal perspective, because the true issue is who bears the cost, not merely the treasury's line. They measure dependency in activities of daily living limitations and model each cohort's trajectory over years, not a single episode. Critically, they value informal care explicitly. Under the opportunity-cost method, rising female labour-force participation means the forgone earnings of carers — and the tax the state loses when a carer leaves work — climb steeply over the projection. This reframes Option A: it looks cheapest to the treasury only because it hides an enormous and growing cost on families and on the wider economy.

Option B costs more public money up front, but the modelling shows it does three things the others do not. It pools a catastrophic, uninsurable risk that the market cannot; it channels a cash benefit to family caregivers, partly formalizing and protecting the informal-care supply the country will increasingly need; and, by funding home care and reablement, it reduces avoidable hospital use — a cross-boundary gain that only the societal view reveals. The team flags the risks honestly: the paid-care workforce barely exists and must be built, wage and training costs will rise, and the contribution base is narrow in an economy with a large informal sector. They recommend a phased social-insurance scheme with an explicit carer cash benefit and a matching workforce plan, and they refuse to book the "saving" from lower family care as free — because, valued properly, it was never free at all.

Four sector lenses

Startup

A digital-health or care start-up in this space typically sells into the seams the system leaves open: carer-support apps, remote monitoring and falls detection, care-coordination platforms, or marketplaces matching home-care workers to families. Its economics turn on who will actually pay — the dependent person, the family, an insurer, or a public commissioner — and that buyer is often not the user. The strongest evidence a start-up can bring is a credible, long-horizon claim about delayed dependency or avoided admissions, valued in the budget that benefits; the hardest problem is that the payer who would save is rarely the payer who buys. Start-ups should design for the reality that much of their value accrues to third parties years later, and build the data to prove it.

Small business

A small established provider — a single care home, a family-run domiciliary care agency, a GP partnership, or a modest equipment supplier — lives on thin margins set largely by the fees a public commissioner is willing to pay. Unlike a start-up chasing growth on investor money, it survives on steady occupancy or hours and on a local reputation that took years to build, so its central risks are a commissioner's fee freeze, a bad inspection, and the loss of a few key care workers it cannot easily replace. Workforce is the daily constraint: recruiting, paying, and keeping staff in a tight local labour market often decides whether the business can accept the next placement at all. It has little power over the financing model or the health–social-care boundary and mostly absorbs their consequences, so its economic skill is careful costing of its own service and honest negotiation of a sustainable rate rather than reshaping the system around it.

Enterprise

A large care provider, hospital group, or insurer manages LTC as a portfolio of long-duration, correlated risks and thin-margin, labour-intensive services. Its central challenges are workforce — recruiting, paying, and retaining care staff in a tight market — and the boundary, since a provider that spans health and social care can either exploit cost-shifting or capture the gains from integrating across it. An insurer that has watched the private long-term care insurance market fail knows that pricing decades of uncertain care risk is close to impossible without a public backstop or a mandate. At enterprise scale, the winning move is usually to organize around the whole trajectory of need rather than the profitable slice, and to treat carer support and workforce stability as core operations, not overheads.

Government

A ministry or national payer owns the financing model, the eligibility threshold, and the boundary itself, and therefore owns the biggest choices in the chapter. It must decide how to distribute an uninsurable risk — through social insurance, a means-tested safety net, or continued reliance on the family — knowing each choice has stark equity and gender consequences (see Chapter 3.4 — Equity). It also controls the levers of integration: pooled budgets, joint assessment, and aligned accountability that can blunt cost-shifting, though never for free. Government carries the fiscal tail of an ageing society and the political heat of any threshold, so its task is to make the distribution of cost explicit and defensible rather than letting it fall by default on whoever is least able to resist — usually unpaid women.

Common failure modes

• Valuing informal care at zero. Leaving unpaid family care out of the account makes options that dump work on families look cheap. Fix: count it explicitly by opportunity-cost or replacement-cost method, and test both.
• Costing dependency as an episode. A single-year snapshot understates a commitment that runs for years. Fix: model the whole trajectory and the long high-cost tail over a long horizon.
• Calling a cost-shift a saving. Moving cost across the health–social-care boundary looks like efficiency on one ledger and is a loss on the other. Fix: measure across both budgets together and reward net gains only.
• Betting on private LTC insurance. Assuming a healthy private market will fill the gap ignores a near-universal market failure. Fix: model low take-up and high premiums as the base case, and design a public role.
• Short-horizon ROI on prevention. Judging reablement and falls prevention on next year's numbers hides value that lands years later in another budget. Fix: use long horizons and align the paying and benefiting budgets.
• Threshold cliff-edges. A hard means or dependency test creates gross unfairness and gaming just either side of the line. Fix: use tapers, caps, and disregards, and cost the behavioural response.
• Assuming an infinite care workforce. Financing plans fail in the labour market when care wages and turnover are treated as afterthoughts. Fix: plan the paid workforce as a binding constraint alongside the money.
• Integration sold as instant savings. Promising quick cash from pooling budgets sets integration up to be judged a failure. Fix: promise better coordination and outcomes, and evaluate honestly.

Maturity model

Checklist

• The costing covers the whole trajectory of dependency, not a single episode, over a long horizon.
• Informal care is counted explicitly, with the valuation method (opportunity-cost or replacement-cost) stated and the alternative tested.
• The perspective (health system, wider public sector, or societal) is fixed and stated before results are read.
• The financing model is chosen deliberately, and what it distributes to individuals, families, and contributors is made explicit.
• Any reliance on private LTC insurance models low take-up and high premiums as the base case.
• Eligibility thresholds are designed with cliff-edges, tapers, caps, and behavioural responses considered.
• Cost is tracked across both the health and social-care budgets, and no cross-boundary transfer is booked as a saving.
• Carer support is appraised as an investment, with avoided formal-care and health costs on the benefit side.
• Prevention and reablement are judged over a long horizon with the benefiting budget aligned to the paying one.
• Outcomes include social-care and capability measures suited to older and dependent people, not health outcomes alone.
• The paid care workforce is planned as a binding constraint on delivery.
• Demographic and family-structure change is projected, not assumed to hold at historical rates.

Key sources

• World Health Organization — reports on ageing, healthy ageing, and long-term care systems.
• Organisation for Economic Co-operation and Development (OECD) — Health at a Glance and dedicated long-term care work (e.g. Help Wanted? Providing and Paying for Long-Term Care).
• Germany's statutory long-term care insurance (Pflegeversicherung) and Japan's Long-Term Care Insurance system, as exemplars of the social-insurance model.
• England's social care means-testing framework and successive independent reviews of care funding, as exemplars of the means-tested model.
• Academic literature on the valuation of informal care (opportunity-cost and replacement-cost methods) and on the failure of private long-term care insurance markets.
• ICECAP-O and adult social care outcome measures, for capability and wellbeing measurement in older people (see Chapter 3.5 — Capabilities).

References

1. Long-term care — Wikipedia — https://en.wikipedia.org/wiki/Long-term_care
2. Long-term care insurance — Wikipedia — https://en.wikipedia.org/wiki/Long-term_care_insurance
3. Social care — Wikipedia — https://en.wikipedia.org/wiki/Social_care
4. Activities of daily living — Wikipedia — https://en.wikipedia.org/wiki/Activities_of_daily_living
5. Dementia — Wikipedia — https://en.wikipedia.org/wiki/Dementia
6. Informal care — Wikipedia — https://en.wikipedia.org/wiki/Informal_care
7. Caregiver — Wikipedia — https://en.wikipedia.org/wiki/Caregiver
8. Care work — Wikipedia — https://en.wikipedia.org/wiki/Care_work
9. Social insurance — Wikipedia — https://en.wikipedia.org/wiki/Social_insurance
10. Means test — Wikipedia — https://en.wikipedia.org/wiki/Means_test
11. Adverse selection — Wikipedia — https://en.wikipedia.org/wiki/Adverse_selection
12. Dependency ratio — Wikipedia — https://en.wikipedia.org/wiki/Dependency_ratio
13. Health at a Glance — OECD — https://www.oecd.org/en/publications/health-at-a-glance_19991312.html
14. Ageing and health — World Health Organization — https://www.who.int/health-topics/ageing