
Who holds the risk? Six ways the state came to carry it, and the tier it left behind
Before any government insured anyone against sickness, old age or want, people insured each other. In Britain in 1910 somewhere between 6.3 and 9.5 million people belonged to a friendly society, and about four working men in five were members at some point in their lives; the Oddfellows alone had over a million. In Australia in 1909 half a million people were lodge members in a population of 4.8 million, which with their dependants was something like half the country, and in the mining towns nearly every man. Germany's guild and factory sickness funds predate Bismarck by a century. The Danish sygekasser, the French sociétés de secours mutuels, the Japanese kō and the Chinese village cooperatives all did the same work: a regular contribution into a pool, a defined benefit when the bad thing happened, and a book that the members, not a treasury, could read.
This essay opens a series on the Five Giants that William Beveridge named in 1942, want, disease, ignorance, squalor and idleness, written from Sydney and read across five desks. Its question is simple and it will recur in every part: when the state took over the job of carrying risk, what did it do with the tier that was already carrying it? The answer differs by country more than any left-right story allows, and the differences explain most of what is happening this autumn in Canberra, Westminster, Paris, Berlin, Washington and Beijing.
Germany built on it
Bismarck's sickness insurance law of 1883 is usually called the first welfare state. It was closer to a franchise. It made membership of a sickness fund compulsory for industrial workers and set minimum benefits, but the funds themselves were the existing local, occupational and company Krankenkassen, run by elected boards of contributors and employers, holding their own money. Accident insurance followed in 1884 and old-age pensions in 1889. The state wrote the rules and the funds held the risk. That architecture has survived two empires, a republic, a dictatorship, a division and a reunification. The number of statutory funds has fallen from thousands to under a hundred through mergers, and a central health fund has pooled contributions since 2009, but a German still chooses a Kasse, pays a contribution that is printed on the payslip, and can read the fund's accounts. When long-term care was finally added in 1995 it was added the same way, as a contributory Pflegeversicherung with a defined benefit, which is why its members know to the euro that their nursing-home co-payment rises by half in January and that the pool runs short in October. France took a similar path in 1945: a state Sécurité sociale for the basic layer, with the mutuelles kept as the complementary tier, so that nineteen in twenty French residents today hold a second, member-owned policy on top of the first. Japan did the same between the Health Insurance Act of 1922 and universal coverage in 1961, and added care insurance in 2000 with contributions from everyone over forty.
The common feature is that the member survived the state. The contribution is visible, the benefit is defined, and the body that holds the money has a name and an annual meeting.
Britain nationalised it
Lloyd George's National Insurance Act of 1911 also used the existing tier, and that is where the British story turns. Sickness benefit and the doctor's bill were administered through "approved societies", which were the friendly societies, the trade unions and, crucially, the big industrial insurers like the Prudential, admitted on equal terms. One contemporary called the Act "the death warrant of the friendly societies", and over thirty years it was. The societies became agents of a state scheme, their surpluses regulated, their doctors' panels fixed, their rivalry with commercial insurers decided by capital. By the time Beveridge wrote his report in 1942, the approved-society system was the thing he most wanted to abolish: it was unequal, because rich societies paid extra benefits and poor ones did not, and it was administratively tangled. The National Health Service Act of 1946 and the National Insurance Act of the same year did abolish it. From 5 July 1948 health was free at the point of use and paid from taxation and a flat national contribution; the societies kept their savings arms and lost their purpose.
What Britain nationalised was a working mutual tier for sickness. What it did not nationalise, because no such tier existed, was care. Care of the old and the disabled at home had been done by families, parishes and the lodge's sick visitor, and the National Assistance Act 1948 simply left it there with a means test attached. Seventy-eight years later the means test is still there, its upper limit frozen at £23,250 since 2010, and the Prime Minister's promise of 27 September to make personal care free from the next parliament is the twenty-third attempt to finish the job. The Scandinavian countries followed the British route a generation later: Denmark's sickness funds, state-subsidised since 1892, were absorbed into tax-funded county health services in 1973, and Sweden's a decade earlier. In the Nordic model the contribution disappeared into general taxation entirely, and the member with it; what remains is a citizen and a municipality.
Australia replaced it twice
Australia's lodges were, per head, among the strongest in the world, and lodge medicine, a doctor on an annual capitation contract to the society, was the ordinary way a working family saw a doctor until the 1920s. The lodges lost that first to the doctors, who fought capitation for thirty years and won fee-for-service from the last national holdout, the Victorian Manchester Unity, in 1921; then to the Commonwealth, which built pharmaceutical benefits in the 1940s and hospital and medical benefits under the National Health Act 1953 through subsidised private funds, several of them the old societies in new form. Medibank in 1975 and Medicare in 1984 replaced that again with a tax-funded, universal scheme carrying a levy that is a tax in all but name. Australia is therefore a Beveridge country for health with a Bismarckian residue, the private health funds, kept alive by rebate and penalty. For care it did something neither model had tried: the NDIS of 2013 wrote an individual entitlement into statute and funded it from general revenue, with no contribution, no fund and no membership, and on 1 October this year it trimmed participation supports again, because an entitlement with no pool behind it is only ever as large as the next budget allows.
The United States never built it
America is the control case. It had friendly societies, lodges and ethnic mutual-aid societies on the same scale as Britain in 1910, and lodge practice was fought by organised medicine with the same result. But the national takeover never came. Social Security in 1935 covered old age and, later, disability; health was left to the market, and the market was shaped by an accident: wartime wage controls in the 1940s let employers compete on benefits instead of pay, and a tax ruling made employer-paid insurance untaxed income. Employment became the membership. Medicare and Medicaid in 1965 covered the old and the poor; the Affordable Care Act of 2010 subsidised the rest. Just over half of Americans still get their cover through a job, and the consequence for care is that Medicaid, a means-tested poverty programme, is the largest payer of long-term care in the country, and when states tighten its work rules, as several are doing this autumn, the fallback is a crowdfunding page. GoFundMe is what a friendly society looks like when nobody owns it.
China built it, lost it, and is rebuilding it by law
The most dramatic version of the cycle is Chinese. The rural cooperative medical schemes of the 1950s and 1960s were village-level mutual funds, financed by the collective and small member contributions, with barefoot doctors on retainer; by 1976 they covered about nine in ten villages. When the communes were dissolved in the early 1980s the funds dissolved with them, and within a decade coverage had fallen to a few per cent of the rural population. The state rebuilt from 2003 with a voluntary, subsidised residents' scheme, merged it in 2016, and on 28 August this year legislated the Medical Security Law, which from 1 January 2027 makes enrolment a legal duty and tax authorities the collectors, while leaving the premium and the benefit to be set by notice. Between the collapse and the law, something else happened: Xiang Hu Bao, an online mutual-aid pool run by Ant Group, enrolled over a hundred million members in three years and was shut down in 2021, because the operator held the ledger and the regulator could not see it. China has now had all three tiers in living memory, mutual, market and state, and has chosen the state, with a duty to pay and a promise to be named later.
The pattern, and the giant it leaves standing
Read side by side, the six systems sort into three answers to the same question. Germany, France and Japan kept the member and put the state behind the fund. Britain, Scandinavia and, for health, Australia abolished the member and put the citizen in front of the ministry. The United States never answered, and China answered three times. The differences matter less for health, where every rich country except America now covers nearly everyone somehow, than for the giant none of them finished: care. Care was never mutualised at scale before the state arrived, so there was nothing to nationalise and nothing to build on. Germany and Japan built a contributory tier for it late; Britain, Australia and America left it to the family and the means test; China is where Britain was in 1946. That is why five of our six desks produced the same story this autumn, a government shifting the cost of care onto the citizen and keeping the decision in the ministry, and why the one that didn't, Germany, produced instead a fund whose members can read the shortfall.
The other four giants went through the same cycle in their own time: income support from the trade-union out-of-work fund to the benefit office, housing from the building society to the housing authority, education from the mechanics' institute to the department. Later essays in this series will take each of them, country by country. The question will be the same each time. Not whether the state should carry the risk, which in a rich country it plainly must for the catastrophic part, but what happened to the tier that let people carry the ordinary part together, and whether anything was gained by removing it.
For communities, now
Nothing in this history is a reason to wait. The lodge of 1909 needed a hall, a secretary and a locked tin; the equivalent today is forty households and a ledger every member can read. The ordinary risks that no national scheme fully covers, the carer's week off, the gap between an assessment and a service, the taxi to the clinic, the bridging month when a plan is cut, are exactly the risks the friendly societies were built for, and they can be pooled on a street this year. The pool is held by no one in particular; each member's balance, contribution and claim sits in a ledger every member reads, end-to-end encrypted in the app, and the record survives the neighbour who started it. The state took over the tier because the tier worked. It can work again, underneath whatever the state decides next.
Sources: Historical Association, "The world in 1913: friendly societies"; Green and Cromwell, Mutual Aid or Welfare State (Australian lodge membership, 1909); History Cooperative, "Getting a grip: the roles of friendly societies in Australia and Britain reappraised"; Bismarck's Health Insurance Act 1883 and Pflegeversicherung 1995 (German federal law gazette); National Insurance Act 1911, National Health Service Act 1946, National Assistance Act 1948 (legislation.gov.uk); Beveridge, Social Insurance and Allied Services (1942), on approved societies; Danish Ministry of Health on the 1973 sygekasse reform; Australian National Health Act 1953, Medibank (1975) and Medicare (1984) records; US Social Security Act 1935, IRS treatment of employer health benefits (1943, 1954), Census Bureau health insurance coverage; World Bank and Lancet histories of China's cooperative medical schemes; Xinhua on the Medical Security Law (28 Aug 2026); YourCare leaders on the NDIS, Germany, France, the United States, China and the United Kingdom (Sep–Oct 2026).
Get Early Access
YourCare is preparing for iOS App Store submission. Join the waitlist.
Join the Waitlist