Why Japan’s Care Insurance Is Not the Bankruptcy Story You Read

Four claims about Japanese elder care that circulate in English, checked against the 2025 payment statistics and the 2026 premium schedule

SeaPRwire – Every few months an English-language outlet discovers that Japan’s long-term care insurance is about to collapse. The story usually opens with the cost figure, notes that the population is ageing, quotes a projection for 2040, and concludes that the system is unsustainable. The Japanese-language coverage of the same week is usually about a municipal premium vote, a revision to the fee schedule, or an argument over whether the copay should rise from 10% to 20% for higher earners.

Both descriptions are of the same system. Only one of them is a budget document, and the gap between the two is where four durable misunderstandings live.

The claim defended here is narrow: Japan’s care insurance is not a fund that can go bankrupt, because it is not a fund. It is a pay-as-you-go social insurance with a three-year repricing cycle, and the political fight is not about solvency but about which generation absorbs the next repricing. Once that is clear, the four popular claims below stop being alarming and start being interesting.

Four Claims That Survive Because They Are Half Right

None of these claims is invented. Each has a true core, which is why each travels.

• Care costs are exploding and the system will run out of money.
• Japan puts its elderly in institutions and visits them twice a year.
• Foreign care workers will close the labour gap.
• Japan’s elderly are the richest old people in the world.

Take them one at a time, with the numbers the Japanese government published in 2025 and 2026.

The Fund Is Not Running Out, It Is Repricing

Start with the institution. The Long-Term Care Insurance Act passed in 1997 and the scheme began operating in April 2000, which makes it younger than most of the people it now supports. It is financed half by premiums and half by public funds, with users paying 10% of the cost at the point of service, rising to 20% or 30% for higher incomes. Everyone over 40 pays in; everyone over 65 is entitled after certification, and the 40 to 64 cohort is entitled only for 16 specified age-related conditions.

That structure explains why “running out” is the wrong frame. There is no accumulated pot being drained. Money comes in each year from premiums and taxes and goes out each year as benefits. What can rise is the contribution, and what can fall is the benefit.

Both have been moving. Total care spending reached 12.2208 trillion yen in fiscal 2025, up 282.7 billion yen or 2.4% on the year, and roughly 2.8 times the 4.3783 trillion yen recorded in fiscal 2001. The number of people receiving care services rose 48,400 to 5.7795 million, and preventive services reached 1.3612 million.

On the contribution side, the nationwide weighted average premium for the 65-plus group was set at 6,225 yen a month for the ninth planning period covering fiscal 2024 to 2026. That is 3.5% above the 6,014 yen of the previous period and 2.1 times the 2,911 yen charged when the scheme started in fiscal 2000. The spread across municipalities is wider than the average suggests: Osaka charged 9,249 yen and the village of Ogasawara 3,374 yen in the same period. Of all municipalities, 45.3% raised the premium, 37.2% held it flat, and 17.5% cut it by drawing on reserves.

The 40 to 64 cohort pays differently. Its contribution is bundled into medical insurance at a rate that rose from 1.59% to 1.62% from March 2026 premiums, split between employee and employer. That rate is uniform across prefectures, unlike the medical portion, which is why a payslip is a better guide to the intergenerational bargain than any headline about the 65-plus premium.

And the fee schedule was revised upward in 2026 by 2.03% on a temporary basis, split between 1.95% for staffing from June 2026 and 0.09% for meal cost from August 2026. The staffing component is explicitly a pay rise: care workers get about 10,000 yen a month more, with a further 7,000 yen for providers that meet productivity conditions.

Put those together and the picture is not a collapse. It is a machine that raises its own price every three years, and the price is rising faster than the wages of the people paying it. The 2040 projection of 27.6 trillion yen, equal to 3.5% of GDP, is the reason the machine keeps being adjusted, not evidence that it has failed.

Where 5.4 Million People Receive Care

The second claim is the one with the strongest emotional pull and the weakest empirical base: that Japan institutionalised its elderly.

The December 2025 monthly report from the ministry counts three delivery channels. Home-based services reached 4,476,325 recipients. Community-based services, the small-scale day care and night visits designed to keep people in their own neighbourhoods, reached 943,430. Facility services reached 974,983.

Read those three together. Roughly 5.4 million people received care at home or in a community setting, against 975,000 in a facility. That is about 85% outside institutions. Among facility users, 584,910 were in special nursing homes, 340,721 in rehabilitation facilities for the elderly, and 52,039 in medical care hospitals.

Table 1: Recipients of Japanese long-term care insurance by delivery channel, December 2025 monthly report

Source: Japan Institute of Life Insurance , from MHLW monthly report, December 2025

Two caveats before this becomes an argument. A person can receive more than one type of service, so the total overstates the number of individuals. And “home-based” includes people living in paid residential homes, which are private housing with care attached, not the family apartment. Still, the ratio is decisive: for every person in an institution, roughly five and a half receive care somewhere else.

Why does the myth survive? Because institutional care is visible and expensive and photogenic, while a visiting helper at 8am is not. And because the waitlist for a special nursing home is a real and painful thing that generates local news, which then gets read abroad as the norm.

Eighty Thousand Foreign Workers Against 2.3 Million

The third claim is the most recent and the most optimistic: that Japan will import its way out of the shortage.

The numbers are genuine and growing fast. At the end of July 2026, about 80,000 people were residing in Japan under the Specified Skilled Worker (i) status in the care sector, according to provisional immigration figures reported by The Mainichi. The cap for that status in care was 126,900 through March 2029. In October 2026 the government told an expert panel it would add 20,000 places by shifting them from a separate care quota of 33,800 under the skill-development employment system that begins in April 2027, keeping the overall care cap at 160,700.

So the pipeline is real, and it is being expanded because demand outran the plan. None of that makes it sufficient. Japan has more than two million registered care workers, and the ministry’s ninth plan projected a need for 2.72 million by fiscal 2040, roughly 570,000 more than the workforce in fiscal 2022. Eighty thousand people against a workforce above two million is under 4%.

The structural limits are the ones that do not move with policy. The care field requires a Japanese-language test at roughly A2 level plus a care-specific skills test, which sets a floor on who can enter. The programme is capped by Cabinet decision and has already been closed once in another sector: food service stopped accepting new entrants in April 2026 when it hit its ceiling. And the sector’s own representative body publicly criticised the 20,000-place shift as evidence that the estimate was wrong rather than as a solution.

The honest version of the claim is that foreign workers are closing part of the gap and will close more of it. The false version, which circulates more widely, is that they will close it.

Not Everyone Over 65 Is Comfortable

The fourth claim is the mirror image of the first: that Japan’s elderly are wealthy, so the cost is affordable.

Japan does hold a large share of household financial assets in older hands, and that fact gets repeated as though it described every household. The OECD’s pension review puts Japan’s relative poverty rate for people over 65 at 20% or more, against an OECD average of 14.8%. Relative poverty here means income below half the median household disposable income, after taxes and transfers.

The welfare data points the same way. About 1.64 million households were receiving public assistance as of December 2025, more than half of them elderly households, and more than 90% of those were single-person households. Applications reached 256,438 in 2025, the highest since the current counting method began in 2013.

The insurance statistics add a demographic detail that matters for policy. Of the 7.08 million people certified as needing care or support in fiscal 2023, 4.76 million were women. The first-category insured pool that year was 35.89 million people, and 20.18 million of them were 75 or older. A system whose beneficiaries are disproportionately old women living alone is a system whose costs are driven by household structure, not only by medical need.

What the Certification Numbers Say About Health

There is a counterintuitive fact inside the certification data that neither the bankruptcy story nor the wealth story can accommodate.

In fiscal 2023, 6.95 million people were certified as needing care or support, up 280,000 from three years earlier. The 2021 projection had expected 7.15 million by that year. The system came in 200,000 below its own forecast, and the ministry’s explanation was demographic rather than fiscal: the certification rate among people in their eighties has been falling, which suggests that people are reaching that age in better condition than the previous cohort.

By April 2026 the certified population had reached 7.61 million, and the number actually receiving services was 5.85 million, of whom 1.82 million were men and 4.03 million women. The gap between certified and receiving is normal and useful: certification is a gate, not a subscription, and some people are certified at a low level and use nothing that month.

Cost per user is where the system’s real pressure shows. In the April 2025 review, the average monthly cost for care services was 208,400 yen per user, and for preventive services 28,600 yen. The spread across prefectures ran from 233,600 yen in Tottori to well below that elsewhere, which reflects how much of the cost is driven by local supply rather than by local need. A prefecture with more facility beds spends more per user, because the service exists to be used.

That is the mechanism behind the 12.22 trillion yen figure, and it is a supply story as much as an ageing story. Japan did not build its way into a cost problem by having more old people than expected. It built capacity, priced it, and then paid for it.

Why the Fee Schedule Is the Real Policy Lever

If the fund cannot go bankrupt, what actually changes? The answer is the fee schedule, revised every three years, and the 2026 revision is instructive.

The 2024 revision was 1.59%, of which 0.98% went to staffing. The 2026 revision added 2.03% on a temporary basis outside the regular cycle, split between 1.95% for staffing from June 2026 and 0.09% for meal costs from August 2026. The staffing money is the point: about 10,000 yen a month more for care workers, extended from certified care workers to the whole workforce, plus 7,000 yen for providers that meet productivity conditions, and new top-ups for home visit services that had been excluded.

Read the sequence. In 2024 the government put two thirds of a small increase into wages. In 2026 it went outside the cycle to put almost all of a larger increase into wages. That is not the behaviour of a system trying to expand benefits. It is the behaviour of a system that has concluded its binding constraint is staff, and that paying staff is cheaper than building beds.

For a household, the same lever works in reverse. Every yen added to the fee schedule comes back through the premium, the payroll rate, or the copay. The 2026 revision is therefore a decision to raise the price of care in order to keep the supply of carers, and the bill arrives in 2027 when municipalities set the next premium.

The Three-Year Clock Nobody Outside Japan Watches

The single fact that makes the Japanese debate legible is the calendar. The scheme began in April 2000. Since then every three years has produced a revision: the second period from 2003, the third from 2006, the sixth from 2015, the ninth covering 2024 to 2026.

Each revision sets two things at once, the benefit basket and the premium path, and each is preceded by a municipal projection of how many people will need care in the coming period. In 2026 the ninth period is in its final year and the tenth, covering 2027 to 2029, is being prepared. That is why the temporary 2.03% fee revision in 2026 exists: the government wanted to lift wages before the tenth-period projections were locked.

For anyone outside Japan trying to read the system, the clock is a better guide than any single number. A story published in 2026 about a record premium is describing a decision taken in 2024 for a period ending in 2026. A story about a 2040 projection is describing a modelling exercise that the tenth and eleventh revisions will invalidate. Neither is wrong. Both are being read as though the system were static, when the only constant is that it gets re-priced every thirty-six months.

Why These Four Keep Circulating

There is a common mechanism, and it is not stupidity.

Each claim contains a fact that is true and a conclusion that does not follow. Care spending really is at a record. Institutional care really is scarce and rationed. Foreign care workers really did grow from a few thousand to 80,000 in under a decade. Some elderly households really are rich. A reader who checks the first clause and stops has a defensible reason to believe the second.

The second mechanism is translation. The Japanese term is kaigo hoken, long-term care insurance, and the English word “insurance” invites the reader to imagine a fund with reserves and an actuarial balance sheet. It is closer to a payroll-financed benefit with a statutory review every three years. Once you call it a fund, “running out” becomes a coherent sentence, and a coherent sentence is much easier to publish than a three-year repricing cycle.

The third is that the Japanese debate is about distribution and the English debate is about survival. In Japan, the argument in 2026 is whether the copay should rise, whether the 40 to 64 rate should rise faster than the 65-plus premium, and whether the fee revision should favour staffing over facilities. Those are arguments about who pays, and they are conducted with the assumption that the system continues. Exported without that assumption, they read as symptoms of collapse.

What Is Left When the Claims Fall

Strip the four claims away and a less dramatic picture remains, with one genuinely hard problem inside it.

The system is solvent in the accounting sense, because it cannot be otherwise: benefits are set annually against contributions, and when the two diverge the fee schedule, the premium, or the copay moves. What is not settled is the fairness of the sequence. A person aged 45 in 2026 pays 1.62% of payroll for a benefit they will draw in 2046 under terms nobody can specify today, while a person aged 75 pays a premium that has doubled in real terms since the scheme began and receives care at a 10% copay that may rise.

That is the fight worth watching, and it is not a bankruptcy story. The 2026 temporary revision, at 2.03%, is the clearest signal of where the pressure is going: it put money into wages rather than into new capacity, which means the system has decided that the binding constraint is staff, not beds. If that is right, the next three-year revision will be about how to pay care workers enough to keep them, and the answer will come from the same three places as always, the premium, the payroll rate, and the copay.

Data basis

Data as of October 2026. Spending and recipient counts come from the Ministry of Health, Labour and Welfare’s fiscal 2025 survey of long-term care benefits, which covers claims reviewed between May 2025 and April 2026; the 12.2208 trillion yen total includes both insurance benefits and user copayments, and is therefore larger than the insurance payout alone. Delivery-channel counts come from the December 2025 monthly report as tabulated by the Japan Institute of Life Insurance, and are counts of service recipients rather than of unique individuals, so a person using day care and home help appears twice. Premium figures are the nationwide weighted average of the base amount for first-category insured persons, published by the ministry for the ninth planning period, fiscal 2024 to 2026; municipal figures vary around that average by income band and by local benefit level. The 1.62% second-category rate applies to standard monthly remuneration from March 2026 premiums. Foreign worker figures are provisional immigration statistics as at the end of July 2026, reported by The Mainichi on 20 October 2026. Poverty figures are the OECD’s relative measure at 50% of median equivalised disposable income.

Limitations

Three numbers are weaker than they look. The 2040 cost projection of 27.6 trillion yen assumes a benefit level and a demographic path that will both be revised. The 570,000 worker shortfall is a planning estimate from 2024, and the government has already revised its intake cap once. And the 85% home-care share counts service episodes, not people, so the true share of individuals cared for outside institutions is somewhat lower.

Sources

• MHLW , fiscal 2025 survey of long-term care benefits, 2026
• MHLW , long-term care insurance premium history by planning period, 2026
• Japan Institute of Life Insurance , recipients by service type, December 2025
• Nikkei , care insurance premium hits record 6,225 yen, 2024
• The Mainichi , Japan to shift 20,000 skilled visa slots to care, 20 October 2026
• OECD , old-age income poverty, Pensions at a Glance 2025

This article was prepared by the expert team at Global Infinity Labs, drawing on insights and data provided through its MCP data services.

About Global Infinity Labs

Global Infinity Labs is an AI innovation company specializing in MCP data services.

Email: support@globalinfinitylabs.com