Health Systems

Why this matters in health economics

Every country, whatever its politics, must answer the same three questions: where does the money come from, how is it pooled, and how does it reach the people who deliver care. The answers are not technical footnotes. They determine who is covered and who is left exposed, whether a serious illness bankrupts a family or merely worries them, and whether a hospital's incentive is to do more, do less, or do the right thing. For a director, the financing model and the payment mechanism are the two levers with the widest reach in the whole of health economics — wider than any drug, pathway, or protocol — because they set the incentives every other decision runs inside.

The stakes are public money at enormous scale and the trust that goes with it. Health spending is one of the largest lines in any national budget, and much of it is other people's money — taxpayers', contributors', premium-payers' — spent by professionals on strangers. How that money is raised decides how progressive the burden is; how it is pooled decides who is protected against the lottery of illness; how providers are paid decides whether the system rewards volume, restraint, or outcomes. Get the design wrong and you can spend more each year and buy less health, while the poorest pay the most at the worst moment.

This chapter is comparative by construction. No country's arrangement is the default, and the instructive contrasts run across income levels: a tax-funded national service, a social-insurance system of competing sickness funds, a private-insurance market, and a low-income setting where families still pay cash at the clinic door. The point is not to crown a winner but to make the trade-offs legible, so you can read your own system's incentives and reform them deliberately. The policy process and regulation that enact these designs belong to Chapter 3.2 — Health Policy; the disciplines of saying no within any budget belong to Chapter 3.3 — Rationing.

Core concepts

A health system is the whole set of organizations, people, and resources whose primary purpose is to promote, restore, or maintain health. The World Health Organization frames its functions as financing, generating resources (workforce, facilities, knowledge), delivering services, and stewardship — and describes people as playing four roles within it: contributors who fund it, citizens to whom it is accountable, providers who staff it, and consumers who use it. Most system analysis reduces to two design choices — how money is raised and pooled, and how providers are paid — layered on top of the goal of universal health care: coverage for everyone who needs care, with financial protection so that using it does not cause hardship.

Financing models are conventionally sorted into a small typology, usually named after their originators. The Beveridge model — after the 1942 Beveridge Report that shaped the United Kingdom's National Health Service (NHS) — funds care from general taxation and typically delivers it through publicly owned providers, free at the point of use; it is the archetype of a single-payer system. The Bismarck model — after Otto von Bismarck's 1880s German reforms — funds care through compulsory social health insurance, usually payroll contributions paid into non-profit sickness funds that pay a mix of public and private providers. Germany's system remains the exemplar. The private insurance model funds care through premiums paid to competing insurers, the dominant mode for the working-age population in the United States. And the out-of-pocket model is not a model so much as the absence of one: where pooling is thin, people pay cash at the point of care, which describes a large share of spending in many low-income countries.

These are ideal types, and every real system is a blend. What matters economically is the degree of prepayment and pooling. A pooled system — whether the pool is filled by taxes or contributions — spreads the cost of illness across the healthy and the sick and across a person's own lifetime, which is precisely the protection an out-of-pocket system cannot offer. Pooling is also where the failures of Chapter 1.3 — Market Failure bite: a voluntary pool unravels through adverse selection, which is why systems that rely on competing insurers use risk equalization (risk adjustment) to transfer funds towards insurers who enrol sicker members, so that competition is about efficiency rather than risk-selection.

The second design choice is provider payment, and here the mechanism is the message. Fee-for-service pays for each item of activity and so rewards volume — it invites the supplier-induced demand of Chapter 1.3 — Market Failure. Capitation pays a fixed sum per enrolled person per period regardless of activity, which rewards prevention and restraint but invites under-provision and risk-selection. Diagnosis-related groups (DRGs) pay a fixed price per case according to the patient's diagnosis and complexity — a form of prospective, activity-based payment that rewards efficiency per admission but can invite "up-coding" and more admissions. A global budget gives a provider a fixed total for a period, which controls aggregate cost tightly but blunts any incentive to treat more patients or shorten queues. No payment method is neutral; each pushes behaviour somewhere, and mature systems blend them deliberately to counterbalance the distortions.

Layered over financing and payment is the public–private mix: who owns the providers, who bears the financial risk, and whether private cover sits on top of, or instead of, public entitlement. Where private insurance buys faster or better access to services the public system also provides, a two-tier system emerges, with consequences for equity that any honest comparison must state.

A pressure sits beneath every one of these designs: health spending tends to grow faster than the economy that funds it. Ageing populations need more and longer care; new technology expands what medicine can do while rarely doing it more cheaply; public expectations ratchet upward as living standards rise; and health, being labour-intensive and hard to automate, suffers the Baumol cost effect — wages must track the wider economy even where measured productivity cannot, so the relative price of care drifts up over time (see Chapter 1.3 — Market Failure). This is a structural fiscal-sustainability challenge that faces tax-funded, social-insurance, and private systems alike, not a temporary overspend that a single round of cuts will settle.

Economists frame the room a government has to meet that pressure as fiscal space for health — the capacity to raise health spending without endangering the sustainability of the public finances. The World Health Organization identifies a handful of sources, each with a ceiling. Economic growth lifts revenue almost automatically, but it is neither guaranteed nor evenly shared. Re-prioritizing the budget towards health competes with education, defence, and debt service, and cannot be repeated indefinitely. Health-specific earmarked revenues — payroll contributions, or taxes on tobacco, alcohol, and sugary drinks — can raise money and, in the case of health taxes, improve health at the same time, but the yields are modest against a whole health budget. External aid is vital in the poorest settings yet volatile, and it shrinks as countries grow richer. Efficiency gains — buying the same health for less — are real but slow and politically contested. No single source is sufficient; durable financing assembles several.

How that space is filled shapes sustainability and equity together, which is why revenue-raising and pooling are not merely accounting choices. A base of broad general taxation tends to be the most progressive and the most stable across a person's working life and retirement; social-insurance contributions levied on wages pool well but narrow as populations age and formal employment shrinks, leaving fewer contributors for each beneficiary; and out-of-pocket payment, the residual in every system, is both the least sustainable and the most regressive, because it charges the sick most at the moment they can least afford it. Shifting the mix towards prepaid, pooled revenue is the through-line of the universal-health-coverage agenda (see Chapter 3.4 — Equity and Chapter 4.2 — Global Health and Trade), and it is as much a strategy for sustainability as for fairness.

Facing this pressure, systems reach for cost containment — but a blunt cap can be a false economy. Freezing tariffs, holding budgets flat, or delaying adoption controls this year's outlay while quietly shifting cost onto households through longer waits and informal charges, storing up sicker and more expensive patients for later, or blocking a technology that would have paid for itself. The disciplined alternative is to contain cost by removing low-value care rather than by squeezing everything equally — the difference between rationing on value and rationing by attrition (see Chapter 3.3 — Rationing and Chapter 2.5 — Budget Impact and Affordability). Sustainable financing is ultimately about buying more health per unit spent, not simply spending less.

Best practices

1. Separate the two decisions — how you raise money and how you pay providers — and design each on its own merits. These are independent levers that are often confused. A tax-funded system can pay by DRGs; a social-insurance system can use capitation; a private insurer can run a global budget with a provider group. Naming which lever you are pulling stops you importing another country's whole model when you only needed one of its parts.

2. Judge a financing model by pooling and prepayment, not by its label. The economically decisive feature is how much of total spending is prepaid and pooled rather than paid out of pocket at the point of need, because that is what delivers financial protection. A "public" system with high user charges and a "private" system with broad mandatory pooling can protect people similarly; the label matters less than the share of care that is prepaid and the breadth of the pool.

3. Choose the payment mechanism to fit the behaviour you want, then counterbalance its distortion. Fee-for-service where you want more of an under-supplied service; capitation where you want continuity and prevention; DRGs where you want hospital efficiency per case; global budgets where you must cap aggregate cost. Because each mechanism has a characteristic failure, pair it with the monitoring that catches that failure — activity audits against fee-for-service, under-referral checks against capitation, coding audits against DRGs, waiting-time and access measures against global budgets.

4. Blend payment methods rather than searching for a perfect one. The strongest systems layer a stable base payment with a marginal activity signal and a quality component — for example capitation for a registered population, a fee for specific priority activities, and a modest pay-for-performance element. The blend exists precisely because no single mechanism gets the incentive right; tune the proportions to the distortion you most fear.

5. Build risk equalization into any system of competing insurers or funds. Where more than one payer enrols members, the profit-maximizing move is to attract the healthy and shed the sick. Risk equalization — transferring funds towards payers with sicker enrolees, adjusted for age, sex, and diagnosis — is what redirects competition from risk-selection towards efficiency and service. Germany's and the Netherlands' systems depend on it; a multi-payer design without it will select rather than serve.

6. Protect against catastrophic and impoverishing spending explicitly. The test of a financing system is not average spend but what happens to a family hit by serious illness. Cap out-of-pocket costs, exempt the poor and the chronically ill from charges, and watch the incidence of catastrophic health expenditure as a headline metric. In systems that lean on out-of-pocket payment, extending even a thin layer of pooled prepayment for the poorest often buys more protection per unit than any change at the top.

7. Design cost-sharing to deter low-value care without deterring the poor and the sick. User charges dampen demand, but the RAND Health Insurance Experiment showed they dampen effective and ineffective care alike, hitting the poorest hardest (see Chapter 1.3 — Market Failure). If you use cost-sharing, exempt low-income and chronic-disease groups, aim charges at genuinely discretionary use, and keep them well below the level at which people skip necessary care.

8. Make the public–private boundary a deliberate choice, and state its equity effect. Decide whether private cover is substitutive (replacing public entitlement), complementary (covering charges), or supplementary (buying speed or comfort), because each has a different effect on the public system and on fairness. A supplementary tier that lets those who can pay jump the queue relieves pressure but entrenches two-tier access; name the trade-off rather than letting it emerge by default.

9. Compare systems on outcomes, access, and financial protection together — never on cost alone. A system that spends less may be rationing harder, shifting cost to families, or under-treating; a system that spends more may be buying speed and choice. Use several dimensions at once — population health outcomes, waiting times and unmet need, out-of-pocket burden, and administrative overhead — and read them as a set. International datasets such as the OECD's Health at a Glance exist to support exactly this kind of multi-dimensional comparison.

10. Adjust every cross-country comparison before you draw a lesson from it. Headline spending and outcome figures conflate demography, disease burden, price levels, and how each country counts. Standardize for age and need, convert on purchasing-power terms and state the year, and check that "hospital admission" or "nurse" means the same thing in both systems before concluding one is more efficient. An unadjusted league table is a source of confident errors.

11. Treat workforce as the binding constraint, not a variable you can flex quickly. Clinicians take a decade to train and are internationally mobile, so a financing or payment reform that assumes staff will simply appear or redeploy will stall. Model the workforce implications of any system change explicitly, and remember that payment design shapes where staff choose to work — fee-for-service can pull them towards lucrative procedures and away from under-served areas and specialties.

12. Plan for sustainability against structural cost growth, not just this year's deficit. Ageing populations, new technology, and the Baumol cost effect (see Chapter 1.3 — Market Failure) push health spending up structurally in every system. A durable design needs a financing base that grows with the economy and a payment system that rewards value rather than volume; reforms that only cut this year's prices without changing the incentive will see cost re-emerge.

Questions to discuss with your team

1. What behaviour is our current provider-payment method actually rewarding, and is it the behaviour we want? Every payment mechanism pushes providers somewhere, and the push is often invisible until you name it. Trace a typical patient through your system and ask what each provider earns more by doing — more procedures, more registered patients, more admissions coded at higher complexity, or simply staying within a fixed envelope. The honest answer usually reveals a mismatch: you may say you want prevention and continuity while paying in a way that rewards activity, or say you want access while paying a global budget that rewards keeping the queue long. A good discussion identifies the one or two places where the incentive most clearly points the wrong way, and what a blended payment would have to reward to correct it, rather than pretending the current method is neutral.

2. When a family in our system faces a serious, expensive illness, what protects them financially — and where is that protection thinnest? This question moves the conversation from average cost to the distribution of risk, which is where financing models succeed or fail. Identify who in your covered population is most exposed to catastrophic or impoverishing spending: the uninsured, the under-insured, those facing charges the exemptions miss, or those whose condition falls outside the benefit package. The tension is that broadening protection costs money and someone must fund the pool, so every extension of coverage is also a decision about who contributes. An honest answer names the specific groups whose protection is thinnest today and the mechanism — a cap, an exemption, a broader pool — that would close the gap, and is candid that closing it has a price and a payer.

3. If we borrowed a design idea from another country's system, what would we have to adjust for it to work here? Policy tourism is seductive and dangerous: a mechanism that works in one context depends on institutions, prices, and a workforce that may not travel. Pick a real example your team admires — German risk-equalized sickness funds, a DRG-based hospital tariff, primary-care capitation — and work through what it presupposes: the data to risk-adjust, the coding capacity to run DRGs honestly, the primary-care base to capitate. The tension is between the appeal of a proven model and the cost of the scaffolding it needs, which is often invisible from the outside. An honest answer distinguishes the transferable principle from the local scaffolding, and states plainly what your system lacks and would have to build before the borrowed idea could function.

4. As our costs grow faster than the economy that funds them, where will the money come from — and which sources have we not honestly tested? This question forces the conversation off this year's balance and onto the structural pressure that faces every system: ageing, new technology, rising expectations, and the Baumol cost effect all push spending up in real terms. Walk through the sources of fiscal space in turn — growth, re-prioritizing the budget, earmarked or health-specific taxes, external aid where relevant, and efficiency gains — and be candid about the ceiling on each. The tension is that the politically easy sources are finite: you cannot re-prioritize towards health indefinitely, growth is neither guaranteed nor evenly shared, and efficiency savings are slow and contested. A durable answer assembles several sources rather than betting on one, and names which of them your team has quietly assumed will always be there. It should also say what happens to protection and equity if the assumed source fails, because the residual in every system is out-of-pocket payment, the least sustainable and most regressive of all.

5. Where is the boundary between public entitlement and private cover drawn in our system today, and what is it doing to equity? Most systems let a private tier grow at the edges without ever deciding deliberately what role it should play, so the first task is simply to describe the boundary as it actually is. Establish whether private cover in your setting is substitutive (replacing public entitlement), complementary (covering charges the public system leaves), or supplementary (buying speed, choice, or comfort), because each has a different effect on the public pool and on fairness. The tension is real and rarely stated: a supplementary tier that lets those who can pay jump the queue relieves pressure on the public system while entrenching two-tier access, so it helps and harms at once. An honest answer names which type dominates in your system, who gains and who is left behind, and whether the boundary is a deliberate policy or an accident that has simply accumulated.

6. When we are told to contain cost, do we remove low-value care or squeeze everything equally — and how would we know the difference? Cost containment is unavoidable, but the method matters more than the target, and a blunt cap is often a false economy that merely moves the cost out of sight. Trace what your last round of savings actually did: whether it stripped out genuinely low-value activity, or froze tariffs and held budgets flat in a way that lengthened waits, raised informal charges, and stored up sicker patients for later. The tension is that squeezing everything equally is administratively easy and looks decisive, while removing low-value care specifically is slow, contested, and requires evidence about what is low-value in the first place. An honest answer distinguishes rationing on value from rationing by attrition, watches the out-of-pocket burden as the cost you may simply be shifting onto households, and admits where the system is currently taking the easy route (see Chapter 3.3 — Rationing and Chapter 2.5 — Budget Impact and Affordability).

In practice: a health economics example

The fictional high-income Republic of Nordvik runs a tax-funded national health service on the Beveridge model: care is free at the point of use, hospitals are publicly owned, and for two decades each hospital has been paid by a global budget — a fixed annual sum negotiated with the health ministry. Aggregate spending has been well controlled, which the finance ministry values. But waiting times for planned surgery have grown steadily, and a public inquiry has found that hospitals, having no marginal reward for treating one more patient, quietly manage demand by letting queues lengthen. The health minister asks whether Nordvik should move to activity-based funding using diagnosis-related groups, as several neighbouring systems have.

The health-economics team frames the choice as a payment-mechanism reform, holding the financing model constant — Nordvik stays tax-funded and single-payer; only how money reaches hospitals changes. Under a DRG tariff, each hospital would be paid a fixed national price per case, adjusted for diagnosis and complexity. The team is clear about what this would fix and what it would break. The global budget's core failure is that it rewards inactivity at the margin, so a DRG tariff should shorten queues by making each additional patient bring income. But DRGs carry their own distortions: hospitals gain by admitting more cases, by coding them at higher complexity than the clinical facts warrant ("up-coding"), and by favouring profitable case-mixes over unprofitable ones. Aggregate cost, so well controlled under global budgets, could rise sharply.

The team models a hybrid rather than a wholesale switch. A large share of each hospital's income remains a fixed base payment for readiness and unavoidable fixed costs; a marginal DRG payment rewards additional activity in the specialties where queues are longest; and an aggregate ceiling caps total activity-based payment so the national budget stays bounded. To defend against up-coding, they commission independent coding audits and a clinical-coding standard before the tariff goes live, not after. To protect access for complex, unprofitable patients, they set higher tariffs for high-complexity groups and monitor whether any hospital is quietly avoiding them.

The team is candid with the minister about the residual tensions. Activity-based payment will lengthen no queues but may pull staff and theatre time towards the well-tariffed elective procedures and away from unglamorous chronic care that the tariff prices poorly — a workforce and case-mix effect to watch. The reform does nothing for the underlying driver of demand, so without attention to prevention and primary care the tariff simply pays to process more illness. And the equity question is explicit: faster elective care is welcome, but if a private tier grows alongside the reformed public system, Nordvik risks drifting towards two-tier access. The minister accepts the hybrid, the audit scaffolding, and a standing review of case-mix and waiting times by income group — an acknowledgement that changing how providers are paid changes behaviour in ways that must be governed, not just launched. The legislative machinery to enact the tariff belongs to Chapter 3.2 — Health Policy.

Four sector lenses

Startup

A digital health start-up rarely sets financing or payment policy, but it lives or dies by them: its entire commercial model is a bet on how it will be paid. A remote-monitoring venture that assumes fee-for-service reimbursement will find no billing code in a capitated or global-budget system, while the same product may thrive where a capitated provider is hungry for anything that reduces admissions it already funds. Founders should map, before building, which payer bears the cost their product reduces and which payer captures the saving — the two are often different actors, and a saving that accrues to someone who is not the buyer is a saving no one will pay for. The credible pitch names the payment mechanism it plugs into and the incentive it aligns with.

Small business

A small but established provider — a GP partnership, a single private clinic, a community pharmacy, or a small diagnostics supplier — takes the system's financing and payment rules as fixed facts it must run a steady business inside, not a bet it is placing. Its margin lives in the details of how it is paid: which items the fee schedule covers, how a capitation formula treats the age and morbidity of its registered list, whether a DRG tariff prices its actual case-mix fairly, and how promptly the payer settles. Unlike a start-up, it is not chasing a new market but protecting a durable one, so its risks are a tariff revision that quietly underpays its work, an exemption rule that leaves it carrying unfunded demand, or a shift in the public–private boundary that moves patients in or out of its reach. Its influence on system design is slight individually, which is why such providers act through professional associations and contract negotiations rather than alone. The practical discipline is to understand its own payment terms precisely and to keep its case-mix and costs legible, so it can show when a rule underfunds the care it is expected to deliver.

Enterprise

A large hospital group or insurer experiences the system's design as its operating reality and its strategic battleground. A provider negotiates the blend of base, activity, and quality payment that determines its income and its risk; an insurer manages the pool, the risk-equalization transfers, and the cost-sharing rules that shape both selection and use. At enterprise scale the organization is large enough to shift the system's behaviour — a dominant hospital can respond to a DRG tariff by reshaping its case-mix, an insurer can compete on efficiency or quietly on risk-selection — so regulators watch it closely. The strategic question is whether the enterprise is optimizing within the incentives honestly or gaming them, because payers and the public increasingly tell the difference.

Government

Government designs, funds, and stewards the whole machine, and owns the choices no single participant can make: the financing model, the breadth of the pool, the payment architecture, and the public–private boundary. Its comparative advantage is reach — it can mandate universal coverage, equalize risk across insurers, set a national tariff, and cap out-of-pocket exposure — and its discipline must match that reach, because every lever has a characteristic failure that only stewardship catches. It also carries the sustainability problem alone: structural cost growth, an ageing population, and a slow-to-train workforce are its to plan for across decades, not budget cycles. The design and enactment of these choices is developed in Chapter 3.2 — Health Policy; the point here is that the system's shape is, ultimately, a governmental decision in every country on earth.

Common failure modes

• Confusing the financing model with the payment mechanism. Assuming that going "public" or "private" settles how providers are paid, when the two are independent. Fix: design financing and payment separately, and name which lever a reform actually pulls.

• Importing a whole foreign model to solve one problem. Copying another country's system wholesale because one feature appeals. Fix: extract the transferable principle and build the local scaffolding it needs, rather than transplanting the scaffolding blind.

• Treating a payment method as neutral. Believing any single mechanism can reward the right behaviour without distortion. Fix: expect a characteristic failure from every method, blend deliberately, and pair each with the monitoring that catches its failure.

• Multi-payer competition without risk equalization. Letting competing insurers or funds profit by selecting the healthy. Fix: build risk adjustment before, not after, opening the system to competition.

• Judging systems on cost alone. Declaring a cheaper system more efficient while ignoring rationing, unmet need, and cost shifted to families. Fix: compare outcomes, access, and financial protection together, and adjust for need and price before drawing a lesson.

• Blunt cost-sharing. Charges that deter the poor and the chronically ill as much as they deter low-value care. Fix: exempt low-income and chronic-disease groups and aim charges at discretionary use.

• Ignoring the workforce. Reforming payment or financing while assuming staff will appear or redeploy on demand. Fix: model workforce effects explicitly and remember that payment shapes where clinicians choose to work.

• Chasing cost containment while ignoring value. Capping budgets or freezing prices across the board to hit a number, which shifts cost onto households and stores up sicker patients rather than saving money. Fix: contain cost by removing low-value care, not by squeezing everything equally, and watch the out-of-pocket burden as the cost you may simply be moving off the books.

Maturity model

Checklist

• State your financing model and your provider-payment mechanism separately, and confirm each is chosen on its own merits.
• For each major provider-payment method in use, name the behaviour it rewards and the monitoring that catches its characteristic failure.
• Confirm payment is blended where a single method's distortion would otherwise dominate.
• Measure the share of care that is prepaid and pooled, and track catastrophic and impoverishing health spending.
• Confirm any competing-payer arrangement has risk equalization on age, sex, and diagnosis.
• Check that cost-sharing exempts the poor and the chronically ill and targets discretionary use.
• State whether private cover is substitutive, complementary, or supplementary, and name its equity effect.
• Compare your system with others on outcomes, access, and financial protection together — adjusted for need and price, not cost alone.
• Model the workforce and sustainability implications of any financing or payment reform before enacting it.
• Distinguish the transferable principle from the local scaffolding before borrowing any foreign design.

Key sources

• World Health Organization — The World Health Report 2000: Health Systems — Improving Performance, and WHO health financing publications (pooling, prepayment, strategic purchasing, universal health coverage).
• World Health Organization — The World Health Report 2010: Health Systems Financing — The Path to Universal Coverage, and WHO guidance on fiscal space, revenue raising, and pooling for health.
• OECD — Health at a Glance — comparative indicators on spending, outcomes, access, and financial protection across countries.
• Justin Keen, Donald Light & Nicholas Mays, Public–Private Relations in Health Care.
• Alan Maynard (ed.), The Public–Private Mix for Health.
• Barbara McPake & Charles Normand, Health Economics: An International Perspective — system typologies and provider payment.
• Economics Network, Health Economics for Teachers — health systems module — https://economicsnetwork.ac.uk/health/teachers

References

1. Health system — Wikipedia — https://en.wikipedia.org/wiki/Health_system
2. Universal health care — Wikipedia — https://en.wikipedia.org/wiki/Universal_health_care
3. Beveridge Report — Wikipedia — https://en.wikipedia.org/wiki/Beveridge_Report
4. Single-payer healthcare — Wikipedia — https://en.wikipedia.org/wiki/Single-payer_healthcare
5. National health insurance — Wikipedia — https://en.wikipedia.org/wiki/National_health_insurance
6. Out-of-pocket expense — Wikipedia — https://en.wikipedia.org/wiki/Out-of-pocket_expense
7. Risk equalization — Wikipedia — https://en.wikipedia.org/wiki/Risk_equalization
8. Fee-for-service — Wikipedia — https://en.wikipedia.org/wiki/Fee-for-service
9. Capitation (healthcare) — Wikipedia — https://en.wikipedia.org/wiki/Capitation_(healthcare)
10. Diagnosis-related group — Wikipedia — https://en.wikipedia.org/wiki/Diagnosis-related_group
11. Two-tier healthcare — Wikipedia — https://en.wikipedia.org/wiki/Two-tier_healthcare
12. Fiscal space — Wikipedia — https://en.wikipedia.org/wiki/Fiscal_space
13. World Health Organization — The World Health Report 2000: Health Systems — Improving Performance — https://www.who.int/publications/i/item/924156198X
14. World Health Organization — The World Health Report 2010: Health Systems Financing — The Path to Universal Coverage — https://www.who.int/publications/i/item/9789241564021
15. OECD — Health at a Glance — https://www.oecd.org/health/health-at-a-glance/
16. Economics Network — Health Economics for Teachers — https://economicsnetwork.ac.uk/health/teachers