Supply of Healthcare

Why this matters in health economics

Every service you commission or run is the output of a production process: staff, buildings, equipment, consumables, and information combined to produce diagnoses, procedures, and care. If you do not understand that process — how output responds to inputs, how cost behaves as volume rises, where scale helps and where it stops helping — you cannot judge whether a proposed hospital is the right size, whether a merger will save money, or whether a rural service is unavoidably expensive or merely badly run. Supply is the other half of the picture whose first half is Chapter 1.2 — Demand for Health and Healthcare; a plan that models one without the other is half a plan.

The stakes are concrete and public. Building a hospital that is too large strands capital in empty wards; building one too small forces patients to travel and denies the volume that makes some specialties safe. Setting a tariff below the true marginal cost of an extra case pushes providers to avoid the patients who need care most. Misreading a provider's objectives — assuming a for-profit chain and a public teaching hospital will respond to the same incentive in the same way — produces policy that backfires. These are supply-side judgements, and getting them wrong wastes money that could have bought health elsewhere.

Supply also carries distinctive risks that a director must see coming. Much of the cost of a hospital is fixed and sunk, so capacity decisions are hard to reverse and easy to get wrong for a generation. Some services are natural local monopolies, so competition cannot be relied on to discipline them. And the organizations that supply care pursue a tangle of objectives — mission, prestige, surplus, teaching, research — that rarely reduce to simple profit. Reading those objectives correctly is the difference between policy that works and policy that merely looks rational on paper.

Core concepts

The health-service production function. At its simplest, supply rests on a production function: the relationship between the quantity of factors of production a provider uses — labour, capital, consumables, technology — and the quantity of health services it produces. It is an engineering-and-organization fact, not a financial one: it tells you the most output obtainable from a given bundle of inputs. Two hospitals with identical inputs can sit on different points relative to their production function, and the gap between what a provider does produce and what it could produce from the same inputs is technical inefficiency — a recurring target for improvement (see Chapter 2.1 — Economic Evaluation for how efficiency is valued).

Inputs and the special place of labour. The inputs to care are capital (buildings, scanners, theatres), consumables (drugs, devices, prostheses), information systems, and — dominant in most health services — people. Labour typically accounts for the majority of a health provider's costs, but its economics as a distinct market with its own wage-setting, licensure, skill mix, and shortages belongs to Chapter 3.6 — Health Workforce and Labour Markets; here it is treated as one input among several in the production process. What matters for supply is the mix: the same operation can be produced with more senior-clinician time and less technology, or the reverse, and the least-cost mix depends on relative input prices.

Diminishing returns and returns to scale. Two ideas govern how output responds to inputs. Diminishing returns describes what happens when you add more of one input while holding the others fixed — a fifth surgeon in a two-theatre unit adds less than the second did. Returns to scale describes what happens when you scale all inputs together: output may rise more than proportionately (increasing returns), in step (constant), or less than proportionately (decreasing). The two are constantly confused, and the confusion produces bad capacity decisions.

Economies of scale and the U-shaped cost curve. Economies of scale exist where the average cost of a unit of care falls as the volume produced rises — fixed costs spread over more cases, bulk purchasing, specialization of staff. But scale economies are exhausted at some point, and beyond it diseconomies set in as coordination, bureaucracy, and complexity grow, giving the long-run average cost curve its characteristic U-shape and defining a minimum efficient scale. The empirical lesson from hospital studies worldwide is sobering: scale economies are real but modest and are exhausted at a moderate size, so the very large hospital is rarely the cheapest per case.

Economies of scope. Distinct from scale, economies of scope exist where producing several services together costs less than producing each separately — shared theatres, laboratories, imaging, and records across specialties. Scope economies are the economic case for the general hospital and for integrated care; they also have limits, and a provider that adds unrelated services can find scope diseconomies of distraction and complexity.

Cost structure: fixed, variable, average, marginal. A provider's costs split into fixed costs, which do not vary with the number of patients in the short run (the building, the core establishment, the scanner), and variable costs, which rise with activity (consumables, some staffing, drugs). From these come the two numbers that drive most supply decisions: average cost, total cost divided by output; and marginal cost, the cost of producing one more case. Because so much of a hospital's cost is fixed, its marginal cost is often well below its average cost — a fact that quietly shapes every pricing and volume decision (see the cost curve family for how these relate).

Capacity, occupancy, and the cost of standing ready. Health services must hold capacity to meet peaks and emergencies, so capacity utilization — how fully the fixed plant is used — is central to their cost per case. Very high occupancy looks efficient but leaves no slack for surges and harms safety; very low occupancy is safe but expensive. The optimum is not full; part of the fixed cost of a health system is the deliberate, valuable cost of standing ready.

Provider objectives: public, non-profit, for-profit. Supply behaviour depends on what the provider is trying to maximize. A for-profit hospital is modelled as maximizing profit, but most of the world's care is produced by public providers and by non-profit organizations whose objectives are mixed — output, quality, prestige, teaching, research, staff welfare, mission — and who reinvest any surplus rather than distributing it. A public hospital answers to a budget and a set of duties rather than a share price. These objective functions predict different responses to the same payment or reform, and assuming universal profit-maximization is a common and costly analytical error.

Provider market structure. How providers compete depends on market structure. Health-provider markets are rarely perfectly competitive; more often they resemble monopolistic competition (many differentiated providers) in dense urban areas, oligopoly among a few large systems, or — commonly — a natural monopoly where a single hospital serves a whole area because scale economies and low local demand leave room for only one. Providers also integrate vertically, joining primary, community, and hospital care under one roof. Monopoly as a market failure to be regulated, and supplier-induced demand, belong to Chapter 1.3 — Market Failure; here the concern is the structure itself and what it means for how supply is organized.

Best practices

1. Model the production function before you model the money. Start from how the service is actually produced — what inputs, in what mix, yield what output — before layering on costs and tariffs. A financial model that does not rest on a credible account of the physical production process will mislead you, because it cannot tell technical inefficiency (producing too little from your inputs) from genuinely high cost. Ask whether a provider is off its production frontier before you conclude its costs are unavoidable.

2. Separate fixed from variable cost, and know your marginal cost. Almost every supply decision — whether to take on extra cases, close a ward, set a tariff, or run a service overnight — turns on which costs actually change. In a hospital, marginal cost is usually far below average cost, so an extra case funded at average-cost tariff looks profitable when it may barely cover its true marginal cost, or vice versa. Insist that business cases state fixed and variable cost explicitly and identify the relevant marginal cost for the decision at hand.

3. Size services to the minimum efficient scale, not the maximum imaginable. Because average-cost curves are U-shaped, there is a volume band where cost per case is lowest; below it you forgo scale economies, above it you invite diseconomies. Use the evidence that hospital scale economies are modest and exhausted at moderate size to resist the reflex that bigger is always cheaper. Where volume is genuinely too low for safety or efficiency, consolidation may help — but justify it with the actual cost curve, not an assertion.

4. Distinguish scale from scope when you plan consolidation. Merging two hospitals to gain scale (more of the same activity) is a different bet from co-locating services to gain scope (shared theatres, imaging, and records across specialties). Many consolidations promise scale savings that the evidence says are small, while the real gains — if any — come from scope and from concentrating specialist volume where outcomes improve with practice. Name which economy you are pursuing, and test whether the structure you propose actually delivers it.

5. Weigh the volume–outcome relationship alongside cost and access. For some complex procedures, providers and clinicians who do more of them achieve better outcomes, which argues for concentrating that work. But concentration raises travel and can create local monopoly, so it trades access and competition against quality. Make the trade-off explicit and service-specific: centralize where the outcome gain is real and large, keep local where it is not (see Chapter 3.4 — Equity for the access dimension).

6. Read the provider's objective function correctly before you design an incentive. A payment reform that assumes profit-maximization will misfire on a public hospital pursuing throughput and teaching, or a non-profit protecting its mission and its staff. Ask what each provider actually maximizes — surplus, output, quality, prestige, research, staff welfare — and predict the response to your lever from that, not from a textbook firm. The same tariff can drive a for-profit chain to expand a profitable service and a mission-driven charity to cross-subsidize an unprofitable one.

7. Design for the right level of capacity, including deliberate slack. Treat spare capacity as a purchased good, not pure waste: a health system that runs at maximum occupancy has no resilience for surges and pays in safety and delay. Decide the occupancy target explicitly, size fixed capacity for the peak you must meet, and cost the reserve honestly rather than pretending full utilization is the goal. The pandemic-era lesson is that surge capacity has real option value.

8. Match input mix to relative input prices and to skill. The least-cost way to produce a given service depends on what inputs cost locally and on what each input can safely do. Where clinician time is scarce and expensive, shifting appropriate tasks to other staff or to technology can lower cost without lowering quality — but the boundaries of safe substitution and the labour-market consequences belong to Chapter 3.6 — Health Workforce and Labour Markets, which you should engage before redesigning the mix.

9. Recognize when supply is a natural monopoly and regulate accordingly. In many areas, scale economies and limited demand mean one hospital is the efficient number, so competition cannot discipline price or quality and something else must. Do not force artificial competition where the market cannot sustain it; instead use regulation, transparent tariffs, quality reporting, and accountable governance. The failure aspects of monopoly power are owned by Chapter 1.3 — Market Failure; the structural fact that some services are single-supplier by nature is a supply reality to plan around.

10. Separate the tariff from the true cost of production. How providers are paid — fee-for-service, capitation, diagnosis-related-group tariffs — is owned by Chapter 3.1 — Health Systems, but you cannot set or respond to a tariff sensibly without knowing the underlying production cost. A tariff above marginal cost invites providers to expand supply; one below it invites them to shed or avoid the activity. Cost the production process independently so you can see whether a price is driving supply in the direction you intend.

11. Benchmark cost per case against comparable producers, adjusting for case mix. Unit-cost comparisons expose technical inefficiency, but only if they compare like with like: a tertiary centre treating complex cases will and should cost more per case than a district hospital. Adjust for case mix, teaching, and research before concluding a provider is expensive, and use the comparison to ask why, not merely to name a number. Wide unexplained variation in cost for similar output is the signal worth chasing.

12. Treat capital decisions as near-irreversible and sunk. A hospital building lasts decades and cannot be resized cheaply, so a mistake in scale or location persists far longer than an operating error. Give major capital cases the scrutiny their irreversibility warrants — stress-test the demand assumptions from Chapter 1.2 — Demand for Health and Healthcare, model a range of futures, and prefer flexible, expandable designs where uncertainty is high. Sunk cost, once spent, should not distort the operating decisions that follow.

Questions to discuss with your team

1. For our biggest service, do we actually know its fixed cost, its variable cost, and the marginal cost of one more case — and are we making decisions as if we do? Most organizations manage to an average cost or a tariff and rarely decompose it, yet almost every marginal decision — take extra referrals, open a Saturday list, close a bay — depends on which costs truly change. The honest discussion starts by trying to state, for one real service, what proportion of its cost is fixed and what an additional case genuinely adds. Expect the exercise to be uncomfortable: cost accounting in health is hard, and the answer is often "we don't know within a wide band." A good outcome is not a false-precision number but a shared understanding of where the fixed–variable line falls and how far marginal cost sits below average cost, so the next volume or closure decision is made with eyes open rather than on tariff arithmetic alone.

2. Where we are proposing to make a service bigger or to merge providers, exactly which economy are we banking on — and does the evidence support it? Consolidation is perennially justified by "economies of scale", but the hospital-cost evidence says those are modest and soon exhausted, while the real gains, where they exist, tend to come from scope and from concentrating specialist volume. The team should force the distinction: are we chasing lower cost from doing more of the same, lower cost from sharing infrastructure across services, better outcomes from higher volume, or simply fewer management teams? Each has different evidence behind it and different risks — merged organizations can grow diseconomies of coordination and reduce local competition. An honest answer names the specific mechanism, quantifies it against comparable evidence rather than assertion, and is candid about what is being traded away, especially patient access and any remaining competitive discipline.

3. What does each of our major providers actually maximize, and does our incentive design assume something different? A payment or performance lever built for a profit-maximizer will land differently on a public hospital, a religious charity, a research-intensive academic centre, and a private chain, because their objective functions differ. The discussion should try to characterize, provider by provider, what really drives behaviour — surplus, throughput, prestige, teaching, research, staff protection, mission — and then ask whether the incentives currently in place pull in that direction or against it. The tell-tale signs of a mismatch are levers that "should" work but do not, or that produce gaming and unintended supply shifts. An honest answer accepts that objectives are mixed and partly unobservable, and designs incentives that are robust to that uncertainty rather than betting on a single tidy model of the firm.

4. What occupancy are we actually running at, is it deliberate, and have we costed the slack we hold — or the slack we have quietly eliminated? Capacity is where supply economics meets safety and resilience, and it is routinely mismanaged from both ends: some services run so hot that any surge tips them into delay and harm, while others carry unexamined empty capacity that no one has justified. The team should establish, for the services that matter, what the occupancy target actually is and why — not the number that emerged by accident, but the level chosen against the peak the service must meet and the surge risk it must absorb. The uncomfortable part is admitting that the efficient occupancy is not full, that deliberate spare capacity is a purchased good with real option value, and that the cost of that reserve should appear on the books rather than being treated as waste to be squeezed out. An honest answer names an explicit target, states what peak and what surge it is sized for, and accepts that a system stripped of all slack has bought a lower running cost by selling its resilience.

5. Where a service is effectively a single supplier for the area it serves, are we relying on a competition that cannot exist — and if so, what is disciplining it instead? In many places the efficient number of hospitals is one, because scale economies and limited local demand leave no room for a second, and yet governance and policy often behave as though patient choice or provider rivalry will keep quality and cost honest. The discussion should identify, service by service, which of our providers are genuine natural monopolies for their catchment, and then ask the harder question: if competition cannot discipline them, what does — transparent tariffs, published quality data, accountable governance, external inspection, or nothing at all? The tension is that forcing artificial competition on a market that cannot sustain it wastes money on duplication and churn, while leaving a monopoly unregulated invites drift in cost and quality. An honest answer distinguishes the services where competition is a real and useful discipline from those where it is a comforting fiction, and puts a deliberate regulatory and governance substitute in place wherever it is the latter.

6. Which of our capital decisions are effectively irreversible, and are we scrutinizing them — and protecting our future flexibility — accordingly? A hospital building lasts decades and cannot be cheaply resized or relocated, so a mistake in the scale or the site of fixed capital persists far longer than any operating error, binding successors who had no say in it. The team should identify which decisions on the table are genuinely near-irreversible and sunk, and ask whether those are getting scrutiny proportionate to their permanence — stress-testing the demand assumptions from Chapter 1.2 — Demand for Health and Healthcare across a range of plausible futures rather than a single central forecast. The real tension is between committing now to capture scale or scope and preserving the option to adapt as demography, technology, and clinical models shift under a building designed for today. An honest answer treats major capital as a bet made under deep uncertainty, prefers flexible and expandable designs where the future is genuinely unclear, and is candid that sunk cost, once spent, must not be allowed to distort the operating decisions that follow.

In practice: a health economics example

Scenario: a fictional national health ministry in a middle-income country in the Andean region, here called the Republic of Andara, deciding how to reconfigure hospital care across a mountainous province.

The Ministry of Health of Andara must decide the future of hospital care in Cordillera Province, where three ageing district hospitals — in the provincial capital and two smaller towns three and five hours away by mountain road — all run below capacity and all need major capital investment. A consultancy has proposed closing the two smaller hospitals and building one large modern hospital in the capital, promising "economies of scale" and a lower cost per case. The Minister asks the ministry's health economist to test the case before any capital is committed.

She begins with the production and cost structure. The three hospitals together run at low occupancy, so on paper a single larger hospital spreading its fixed costs over more activity looks cheaper. But she interrogates the scale claim against the evidence that hospital scale economies are modest and exhausted at moderate size: the proposed single hospital would sit well beyond the volume band where average cost is lowest, risking diseconomies of coordination rather than savings. She also separates the economies at stake — most of the genuine gain would come from scope (shared imaging, laboratories, and specialist theatres) and from concentrating the small number of complex procedures where volume improves outcomes, not from raw scale across routine work that the district hospitals already do adequately.

Then she prices the access cost the consultancy ignored. Closing the two outlying hospitals would push emergency and maternity patients onto three-to-five-hour mountain journeys — care that is price- and distance-inelastic and genuinely needed (echoing the derived-demand logic of Chapter 1.2 — Demand for Health and Healthcare). For time-critical conditions, the travel would cost lives and would fall hardest on the poorest and most remote communities, an equity cost owned by Chapter 3.4 — Equity that any honest appraisal must carry. She also notes that a single provincial hospital would be a natural monopoly with no competitive discipline, so quality would depend entirely on governance and regulation rather than choice.

She models three options: the consultancy's single large build; refurbishing all three at their current scope; and a networked design — one strengthened hub in the capital for complex, high-volume specialist work, with the two district hospitals retained and modernized for emergency, maternity, and routine care, linked by shared diagnostics, common records, and telemedicine. She is explicit about the trade-offs. The single build minimizes headline cost per routine case but strands the province's most vulnerable patients and bets on scale economies the evidence does not support. The three-way refurbishment protects access but forgoes the real outcome gains from concentrating complex work. The networked option captures scope economies and the volume–outcome benefit for complex care while preserving local access — at a higher capital cost and greater coordination demand. She recommends the networked design, stress-tests its demand assumptions across a range of futures, prefers expandable construction given the irreversibility of the capital, and treats the retained local capacity not as waste but as purchased resilience for a province where distance itself is a clinical risk.

Four sector lenses

Startup

A young digital or specialist provider — a diagnostics start-up, a single-specialty day-surgery clinic, a telehealth service — often has an unusually clean cost structure: low fixed costs, high variable share, and the freedom to choose its input mix from scratch. Its supply advantage is precisely that it can pick a narrow, high-volume segment and ride the volume–outcome and scale curves within it, undercutting general providers on that slice. The risks are that it lacks the scope economies of a general hospital, that it may cream-skim the profitable, low-complexity cases and leave the hard ones to others, and that its small scale leaves it exposed if demand or reimbursement shifts. A start-up should be honest about which economy it is genuinely capturing and which costs it is externalizing to the wider system.

Small business

An established small provider — a GP partnership, an independent day-surgery unit, a single-site community hospital, a family-run care home, a local diagnostics supplier — knows its cost curve from years of running it, which is precisely what a start-up does not. Its supply position is defined by a fixed plant and a settled input mix that are hard to flex quickly: it sits at whatever scale it has, usually below the point where the sharpest scale economies would bite, and its unit cost is exposed when volume dips or when a tariff is set off a larger provider's average cost. Unlike the start-up, it is not chasing a growth curve but defending a viable steady state, so its real decisions are about occupancy, skill mix, and whether to share infrastructure — joint diagnostics, back-office, or referral networks — to capture scope economies it cannot generate alone. It competes on local access, continuity, and responsiveness rather than scale, and its chief vulnerability is that a single natural-monopoly rival or an unfavourable tariff can render its modest volume uneconomic overnight.

Enterprise

A large hospital group, integrated delivery network, or insurer-provider organization manages supply as a portfolio across sites, and lives or dies by its cost curves and capacity management. It can pursue scope economies through shared infrastructure and vertical integration, concentrate specialist volume across its network, and negotiate input prices at scale — but it also risks the diseconomies of coordination, bureaucracy, and lost local responsiveness that come with size. Its market power raises regulatory and competition concerns (the failure side belongs to Chapter 1.3 — Market Failure), so it must show that consolidation delivers real production efficiencies rather than merely pricing power. Accountability to boards, regulators, and bond markets means its supply decisions must be costed, benchmarked, and defensible.

Government

A ministry or national payer shapes the supply of care for a whole population and carries duties no firm does: universal access, geographic equity, and stewardship of irreversible public capital. It decides how many hospitals of what size sit where, sets the payment rules that drive provider behaviour (see Chapter 3.1 — Health Systems), and must regulate the natural monopolies that competition cannot discipline. Its hardest supply trade-off is between the cost efficiency of concentration and the access and equity value of local provision, and it must make that trade-off transparently and defend it publicly. Its horizon is long, its capital decisions bind successors for decades, and its accountability is democratic — so it must resist both the false economy of over-concentration and the waste of propping up genuinely sub-scale, unsafe services.

Common failure modes

• Assuming bigger is cheaper. Treating scale economies as unlimited and justifying mega-hospitals on cost grounds the evidence does not support. Fix: use the actual U-shaped cost curve and the modest, quickly-exhausted scale economies hospital studies show; size to minimum efficient scale.

• Confusing scale with scope. Selling a merger as scale savings when any real gain would come from shared infrastructure or concentrated specialist volume. Fix: name the specific economy pursued and test whether the proposed structure delivers that one.

• Pricing off average cost when marginal cost is what matters. Making take-it-or-leave-it volume decisions on average-cost tariffs while true marginal cost sits far below. Fix: decompose fixed and variable cost and identify the marginal cost relevant to each decision.

• One-size-fits-all incentive design. Applying a profit-maximizer's incentive to public and non-profit providers and being surprised when it misfires or is gamed. Fix: characterize each provider's actual objective function and design levers robust to mixed, partly-unobservable motives.

• Worshipping high occupancy. Driving utilization to the maximum and calling spare capacity waste, then having no resilience for surges. Fix: set an explicit occupancy target, cost deliberate slack as purchased resilience, and size fixed capacity for the peak you must meet.

• Ignoring the access cost of concentration. Centralizing services on cost or volume grounds while treating patient travel as free. Fix: cost access and equity impacts explicitly, and centralize only where the outcome gain clearly outweighs them.

• Forcing competition where the market is a natural monopoly. Mandating rivalry in areas that can sustain only one provider, then puzzling at the churn and cost. Fix: recognize natural monopoly and regulate for price, quality, and accountability instead.

Maturity model

Checklist

• For each major service, decompose cost into fixed and variable, and identify the marginal cost relevant to current decisions.
• Before any consolidation, state which economy (scale, scope, or volume–outcome) you are pursuing and test it against comparable evidence.
• Size proposed capacity to the minimum efficient scale band, not to the largest option available.
• Cost the access and equity impact of any centralization explicitly, including patient travel and its incidence.
• Characterize each major provider's actual objective function before designing or applying an incentive.
• Set an explicit occupancy target and cost deliberate spare capacity as purchased resilience, not waste.
• Cost the production process independently of the tariff, so you can see which direction the price drives supply.
• Benchmark cost per case against comparable providers, adjusted for case mix, teaching, and research.
• Identify natural-monopoly services and ensure they are regulated for price, quality, and accountability rather than left to competition.
• Treat major capital decisions as near-irreversible: stress-test demand, model a range of futures, and prefer flexible designs.
• Engage workforce planning (Chapter 3.6) before changing input mix or skill substitution.

Key sources

• Alan Williams' health economics "plumbing diagram" — locating supply of healthcare within the wider map of the discipline (see Chapter 1.1 — Introduction to Health Economics).
• Folland, Goodman & Stano, The Economics of Health and Health Care — standard textbook treatment of production, cost, and provider behaviour.
• Morris, Devlin & Parkin, Economic Analysis in Health Care — production functions, costs, and the theory of the provider.
• Economics Network, Health Economics for Teachers — teaching materials on the structure of the healthcare industry, supply, and cost.
• OECD, Health at a Glance, and WHO health-systems publications — cross-country evidence on hospital capacity, occupancy, and provider mix.

References

1. Production function — Wikipedia — https://en.wikipedia.org/wiki/Production_function
2. Diminishing returns — Wikipedia — https://en.wikipedia.org/wiki/Diminishing_returns
3. Returns to scale — Wikipedia — https://en.wikipedia.org/wiki/Returns_to_scale
4. Economies of scale — Wikipedia — https://en.wikipedia.org/wiki/Economies_of_scale
5. Economies of scope — Wikipedia — https://en.wikipedia.org/wiki/Economies_of_scope
6. Average cost — Wikipedia — https://en.wikipedia.org/wiki/Average_cost
7. Fixed cost — Wikipedia — https://en.wikipedia.org/wiki/Fixed_cost
8. Variable cost — Wikipedia — https://en.wikipedia.org/wiki/Variable_cost
9. Marginal cost — Wikipedia — https://en.wikipedia.org/wiki/Marginal_cost
10. Capacity utilization — Wikipedia — https://en.wikipedia.org/wiki/Capacity_utilization
11. Market structure — Wikipedia — https://en.wikipedia.org/wiki/Market_structure
12. Natural monopoly — Wikipedia — https://en.wikipedia.org/wiki/Natural_monopoly
13. Economics Network — Health Economics for Teachers — https://economicsnetwork.ac.uk/health/teachers
14. Health at a Glance — OECD — https://www.oecd.org/health/health-at-a-glance/