Ranking three · Question one of two

Who owns the delivery system

Not who pays — who owns the buildings, the beds and the production. Hospitals, hospital beds, long-term care facilities and pharmaceutical manufacturing. The United States ranks 36th of 39.

39 countries 4 measured components + primary care, qualitatively OECD-only toggle in the nav
This page and the financing page answer two different questions

This page: who owns the delivery apparatus. The hospitals, beds, care homes and drug factories — the physical means of providing care.

The public financing page: who pays the bills. The share of health spending that is public money, regardless of who owns the provider.

They are close to independent of each other, which is the single most useful thing on either page. Japan finances 84.8% of its health care publicly and owns 13.4% of the delivery system. The correlation between the two across 39 countries is only r = +0.25.

A

How public delivery ownership is measured here

Four components, averaged · and an honest account of the coverage

There is no published "share of the health delivery system that is government-owned" statistic. It had to be assembled. OECD Health Statistics reports ownership sector — publicly owned, private not-for-profit, private for-profit — for four parts of the delivery apparatus, and the index on this page is the mean of whichever are available for each country.

ComponentCountriesWhat it counts
Hospital beds33Share of all hospital beds located in publicly owned hospitals. The best single measure — beds are the unit of acute capacity.
Hospitals33Share of hospital establishments that are publicly owned. Counts a rural clinic-hospital the same as a teaching hospital, so it reads lower than beds where public hospitals are large.
Long-term care beds24Share of residential long-term care beds in publicly owned facilities — nursing and elderly care.
Long-term care facilities31Share of long-term care establishments that are publicly owned.
19 countries have all four components; 8 have three; 13 have two; one has a single component. The component count is shown against every country in the ranking below, because an index built from one number is not comparable to one built from four.
What is missing, and it matters

Primary care and clinics are not in this numeric index, and that is a real gap rather than an oversight. OECD publishes no ownership sector for ambulatory-care providers, and the WHO Global Health Observatory has no facility-ownership indicator at all, for any country. Nobody counts clinic buildings.

Rather than leave primary care out entirely, section C handles it qualitatively — classifying 33 countries by whether GPs are salaried employees or self-employed contractors, which is the ownership question that can actually be answered. The numeric index here is acute and residential capacity; read it together with section C, not on its own.

Two countries are also missing from the OECD ownership tables entirely despite well-known answers. The United Kingdom is the important one: NHS trusts own essentially all acute capacity, so its absence would badly distort the picture. It appears in the hospital-beds series below as a clearly marked documented estimate. Sweden reports only its long-term care figure, which is why its index rests on a single component.

B

The ranking

Public share of the measured delivery apparatus · all countries with data
Public ownership of the delivery apparatus
Mean of available components · hover for the component breakdown
Lithuania, Denmark, Latvia, Norway and Slovenia own most of what delivers their care. The Netherlands is the extreme case in the other direction: its hospitals and care homes are entirely private non-profit, giving it a delivery index of zero despite 68% public financing. The United States sits at 13.9%, ahead only of Japan, South Korea and the Netherlands.
#CountryIndexComponents Hospital bedsHospitalsLTC bedsLTC facilities
Blank cells are components the country does not report. Countries with fewer components have less reliable indices — Sweden's 80.6 rests on a single long-term-care figure.

America owns 21% of its hospital beds, 23% of its hospitals, and about 6% of its long-term care capacity. It is one of the most privately delivered health systems in the developed world.

Every component, every country that reports it

The index above averages the four components. Below is each one on its own, showing every country with data rather than a top-ten — the four series have different country coverage, and a country can sit high on one and low on another.

Publicly owned hospital beds
The United Kingdom appears here as a documented estimate, marked "(est.)" and shown with a dashed outline elsewhere on this page — it does not report to the OECD ownership table, but NHS trusts own essentially all acute capacity. Independent-sector hospitals held roughly 2% of acute beds and 6% of all hospital beds in England as of 2018.
Publicly owned hospital establishments
Counting buildings rather than beds. This reads lower than the beds measure wherever public hospitals are the large ones — a country's public teaching hospitals hold far more beds than the private clinics that outnumber them.
Publicly owned long-term care beds
Publicly owned long-term care facilities
Long-term care is the most privatised part of the delivery system almost everywhere, including in countries with near-total public ownership of hospitals. Canada owns 99.4% of its hospital beds and 24.6% of its care homes; Norway is the rare country that keeps both public.
C

Primary care: who employs the doctor

A qualitative measure where no percentage exists · 33 countries

The hospital numbers say nothing about primary care, and there is no published "share of clinics that are publicly owned" anywhere. But there is a usable signal. OECD reports GP remuneration split by worker status — salaried employees and self-employed — and which series a country reports tells you its predominant model. A country that reports only self-employed GP earnings has a primary-care sector of independent contractors who own their own practices, however the money reaches them.

Denmark is the instructive case, and it runs against intuition

Danish GPs are not state employees working from state-owned premises. They are private practitioners who own their practices, paid by a mix of fee-for-service and capitation out of public funds. The arrangement dates to nineteenth-century craft-guild and citizens' sickness funds; when the counties took over health insurance in 1973, "the general structure and payment systems were maintained. The GPs were still private practitioners." Copenhagen was the historical exception, where GPs were salaried.

Finland is the true contrast: the 1972 Primary Health Care Act obliged municipalities to provide general health care, and primary-care staff are "employed directly by the municipalities on a fixed salary basis." Public financing does not imply public employment, and neither implies public premises.

Two dimensions of delivery ownership
Hospital bed ownership against the primary-care employment model · 33 countries with both measures
United States * documented estimate, not OECD-reported
The top-left cell — public hospitals and salaried primary care — is the fullest, and is mostly Nordic, Iberian and Central European. The top-right cell is the one that breaks the usual story: Denmark, Norway and Ireland own their hospitals but buy primary care from independent contractors. The United States sits in the bottom row with private hospitals and private practice, alongside Japan and the Netherlands.
#CountryPrimary-care modelSalaried GP paySelf-employed GP payPublic hospital beds
Pay in USD PPP, latest available year. A dash means OECD publishes no series for that worker status in that country — which is itself the signal being used. This is an indicator of the predominant model, not a headcount of doctors.
What this qualitative layer adds, and what it still misses

It resolves the biggest blind spot in the hospital-only index — but it is about employment, not premises. Whether a salaried Finnish GP works from a municipally owned health centre (usually yes) or a Danish contractor rents private consulting rooms (usually yes) is documented in the country literature rather than in any dataset. Nobody publishes the share of clinic buildings in public hands, and this site does not claim to.

C

Production and the supply chain

State-owned pharmaceutical manufacturing · assembled from primary sources

Manufacturing is the other half of delivery ownership, and no international body collects it. This is a documented list rather than a ranking, and it is the least systematic material on the site.

CountryState pharmaceutical capacity
CubaBioCubaFarma — state conglomerate covering drugs, vaccines, diagnostics and medical equipment; domestic supply plus export. The most complete public pharmaceutical system in the world.
ChinaSinopharm (~120,000 staff) and China Resources Pharma (~72,700) — central state-owned enterprises spanning manufacture, R&D, wholesale distribution and retail pharmacy. Public ownership along the entire chain.
IndonesiaBio Farma holding company (with Kimia Farma and Indofarma) — 3.1 billion dose annual capacity, among the world's seven largest vaccine makers, WHO licences for 20 vaccines.
BrazilFiocruz/Farmanguinhos and Butantan — federal and state public laboratories; Farmanguinhos supplies about 40% of the Health Ministry's drug purchases from official laboratories.
ThailandGovernment Pharmaceutical Organization — state enterprise under the Ministry of Public Health; 200+ products including antiretrovirals, vaccines and test kits, plus biologics via Siam Bioscience.
BangladeshEssential Drugs Company Ltd — 100% state-owned, three plants, under the Ministry of Health and Family Welfare.
India / KeralaCentral public-sector drug units are largely defunct, but Kerala State Drugs & Pharmaceuticals is fully state-owned and supplies government hospitals.
United StatesNone. No state-owned pharmaceutical manufacturer, no public generic producer, no public vaccine plant. Federal involvement is confined to funding research (NIH) and purchasing — the NIH funds the science, private firms own the resulting patents.
D

Does owning the delivery system change anything?

Tested against health outcomes, spending per person, and efficiency

This is the question worth asking, and the answer is more interesting than either side of the argument would like.

Public delivery ownership against health outcomes
39 countries · delivery index against composite health score
A weak negative slope — but it is composition, not causation. The high-ownership end is crowded with post-socialist countries (Latvia, Romania, Bulgaria, Poland, Croatia) that inherited state hospital systems and have lower outcomes for reasons that have nothing to do with who holds the deeds. Drop those eleven countries and the correlation goes to r = −0.01 — exactly zero.
The sensitivity test that decides this question
RelationshipAll 39Excluding post-socialistRichest only

Any finding that survives only in the full sample and vanishes when eleven post-socialist countries are removed is a finding about post-socialism, not about ownership. On outcomes, nothing survives.

Public delivery ownership against spending per person
39 countries · delivery index against health spending per person, PPP $
The direction is consistently negative in every specification — countries that own their hospitals spend less — and for hospital beds alone the relationship strengthens to r = −0.43 once income is controlled for. But it too weakens sharply when post-socialist countries are excluded, and it never reaches statistical significance at these sample sizes. Suggestive, not established.
Countries grouped into thirds by public delivery ownership
26 countries with three or more measured components · group averages
The most publicly delivered third spends 28% less per person than the least ($5,026 against $6,969) and scores modestly lower on health (81.7 against 86.2), which nets out to better measured efficiency. But the low-ownership group is Germany, Japan, the Netherlands, Australia, Israel, Korea and the United States, and the high-ownership group is largely Central and Eastern Europe. The grouping is as much a map of geography and history as of ownership.
What the delivery data actually supports
  • Health outcomes: no relationship. r = −0.15 across all 39 countries, and r = −0.01 once post-socialist composition is removed. Who owns the hospitals does not predict how long people live.
  • Spending: a consistent negative lean, never significant. Every specification points the same way — more public ownership, less spending per head — and the beds-only partial correlation of −0.43 is the strongest signal on this page. With 24–33 countries it cannot be called established.
  • Efficiency: weakly positive, not significant. r = +0.12, rising to +0.21 among well-measured countries.
  • Public delivery is not public financing. The two correlate at only r = +0.25 — see the comparison of the two axes.

The defensible reading: public ownership of delivery is close to outcome-neutral and possibly cost-reducing. It is a claim about cost and control, not about longevity. Anyone arguing that public hospitals produce healthier populations is not supported by this data — and neither is anyone arguing they produce sicker ones.

Limits worth stating plainly

39 countries is a small sample, all of them OECD or OECD-adjacent, so none of this speaks to low-income health systems. The index measures acute and residential capacity only — primary care, clinics and the workforce are missing. And these are cross-sectional associations: a country's ownership mix is an outcome of its history, not a lever pulled at random, so nothing here estimates what would happen if a given country changed its ownership model.

E

References

Every source used on this page · switch style, then copy individually or all at once
Previous← The most efficient systems NextWho pays: public financing →