Health Workforce and Labour Markets

Why this matters in health economics

You can build a hospital in three years, buy a scanner in three months, and license a drug in a single decision — but a specialist doctor takes a decade or more to train, and a health system that is short of people cannot spend its way out of the shortage quickly. Staff typically account for the majority of a health system's running costs, and the workforce is the input that every other input depends on: a ward without nurses is an empty building, and a diagnostic pathway without radiologists is a queue. For a director, workforce is where budgets, safety, and access meet.

The stakes are unusually sharp because the health labour market is not a textbook competitive market. Entry is deliberately restricted by law through licensure; a single national employer often dominates hiring; training is long, expensive, and frequently subsidized by the public purse; and the people involved carry duties of care that make ordinary market signals — walk out for higher pay elsewhere — ethically and practically loaded. Each of these features bends wages, employment, and effort away from the competitive benchmark, and each opens a policy lever that a leader can pull or misuse.

Get the workforce economics wrong and the costs are concrete and severe: rota gaps filled by expensive agency staff, safety incidents when skill mix is stretched too thin, burnout that drives experienced clinicians out early, and training pipelines that deliver the wrong specialists a decade too late. Get it right and you unlock the single largest lever on both cost and quality. This chapter treats the workforce as a market to be understood and shaped, not merely a headcount to be filled. General provision and production of services belong to Chapter 1.4 — Supply of Healthcare; the international movement of clinicians belongs to Chapter 4.2 — Global Health and Trade; this chapter owns the labour market itself.

Core concepts

The health labour market and derived demand. Labour economics studies how wages and employment are set by the interaction of workers supplying effort and employers demanding it. Demand for health workers is a derived demand — hospitals want nurses not for their own sake but for the care they produce (the demand side is developed in Chapter 1.2 — Demand for Health and Healthcare). Supply depends on wages, working conditions, training capacity, and the alternative uses of a clinician's skills. The health human resources field studies the planning and stock of this workforce.

Monopsony and wage-setting. A monopsony is a market with a single dominant buyer — here, a single dominant employer of labour. When one organization, such as a national health service, employs most of a country's nurses, it faces the whole upward-sloping labour supply curve rather than a competitive going wage, and can set pay below the level a competitive market would produce. The signature of monopsony is a persistent, unfilled vacancy that the employer does not clear by raising wages, because raising the wage of the marginal recruit means raising it for everyone. Monopsony as a market failure is catalogued in Chapter 1.3 — Market Failure; here it is a practical wage-setting reality.

Human capital and the economics of training. A clinician embodies human capital — a stock of skill built by costly, time-consuming investment (education, training, experience) that yields a return over a career. Because medical education is long and partly publicly funded, training decisions are investment decisions with long lags: the number of places opened today sets the specialist supply a decade hence, and errors are slow to correct. Physician supply and nursing shortage dynamics both turn on this pipeline lag.

Licensure and entry restrictions. Occupational licensing makes it illegal to practise a profession without a state-recognized credential. It protects patients from unqualified practitioners — a genuine response to the information asymmetry of Chapter 1.3 — Market Failure — but it also restricts entry, and any barrier to entry tends to raise the price and incomes of insiders. Scope of practice rules define what each licensed profession is legally permitted to do; where they are drawn decides who may safely and lawfully perform a task.

Skill mix, the division of labour, and task-shifting. The division of labour — splitting a job into tasks assigned to differently-skilled workers — is as central to health production as to any other. Skill mix is the balance of staff types on a team; task-shifting moves a task from a higher-cost, more-restricted professional to a lower-cost, appropriately-trained one (for example, a nurse practitioner or pharmacist taking on work once reserved to doctors). Done well it raises productivity and access; done badly it stretches safety or simply relabels a shortage.

Compensating differentials. A compensating differential is the extra pay a worker requires to accept worse conditions — night shifts, dangerous exposure, an unpopular rural posting. Where a system cannot or will not pay the differential, it gets the predictable shortage: empty rural posts, unfilled unsocial-hours rotas, and hard-to-staff specialties.

Retention, burnout, and the effort margin. Occupational burnout — exhaustion and disengagement from chronic workplace stress — is an economic variable, not just a wellbeing one. It raises turnover, sickness absence, and error rates, and it lowers the effective labour supply of the people who remain. Retention is usually cheaper than recruitment: replacing an experienced clinician costs the sunk training, the recruitment spend, and the lost productivity of a vacancy and a novice.

Best practices

1. Model the pipeline in decades, not budget years. Training lags mean today's intake decisions set supply ten to fifteen years out, so plan the workforce on the timescale of the asset you are building. Build a stock-and-flow model — entrants, attrition at each stage, retirements, migration — and stress-test it against demographic demand, rather than reacting to this year's vacancy list. The most expensive workforce errors are pipeline errors made a decade earlier and impossible to reverse quickly.

2. Diagnose whether a shortage is real or a pricing signal. A persistent vacancy at a fixed wage is often not an absolute scarcity of people but a wage or conditions set below the market-clearing level — the fingerprint of monopsony wage-setting. Before commissioning more training places (a ten-year fix), ask whether the post would fill at a higher wage, a better rota, or improved conditions (a faster one). Distinguishing a supply shortage from a price problem changes which lever you pull and how quickly it works.

3. Use the whole compensation package, not just headline pay, to fill hard posts. Rural, unsocial-hours, and high-stress roles need a compensating differential; if you will not pay it in money, you will pay it in vacancies. Pensions, housing, training access, protected time, and predictable rotas are all currency, and non-pay levers can be more cost-effective per post filled than across-the-board rises. Target the differential at the specific post that is hard to fill rather than inflating the whole pay bill.

4. Set skill mix by matching task to the lowest safe cost. Apply the division of labour: identify tasks a more-restricted professional performs that a trained, less-costly one could safely do, and shift them — with the training, supervision, and legal cover to make it safe. This raises productivity and access, but only counts as a saving if the freed-up senior time is redeployed to higher-value work, not lost. Task-shifting that simply dilutes a team to plug a gap trades a staffing problem for a safety one.

5. Treat licensure as a safety instrument with a price, and keep scope of practice evidence-based. Occupational licensing protects patients but restricts entry and raises insider incomes; the honest question is whether a given restriction's safety benefit justifies its cost in access and price. Review scope of practice boundaries against evidence, not professional territory, and be alert to rules defended as safety that function as protection of income. Where evidence shows a task is safe in other hands, an outdated scope rule is a rationing device in disguise.

6. Make retention a costed strategy, because it is cheaper than recruitment. Replacing an experienced clinician means paying again for recruitment, induction, and the lost productivity of a vacancy and a learning curve, so a pound spent keeping a good clinician usually beats a pound spent replacing one. Cost your turnover explicitly, identify where in a career people leave, and target retention spend — flexible working, career progression, workload relief — at those points. A retention rate is a financial metric as much as a morale one.

7. Treat burnout as a driver of effective labour supply, not a soft issue. Occupational burnout reduces the hours, productivity, and safety of the staff you already have, and it accelerates the exits you can least afford. The economics favour prevention: workload design, adequate staffing ratios, administrative-burden reduction, and control over rotas are interventions with a measurable return in retained productive capacity. A workforce run permanently hot looks efficient on a spreadsheet and is quietly decapitalizing.

8. Design training bonds and subsidies to align private and public returns — carefully. When the state pays to train a clinician, it holds an asset that can walk away; return-of-service ("bonded") agreements, tuition subsidies tied to public or rural service, and loan-forgiveness schemes try to recoup that human capital investment. They can work, but heavy-handed bonding breeds resentment, evasion, and departure the moment the bond expires. Set the terms as a fair exchange of subsidy for service, and pair them with conditions people will actually stay for.

9. Expand training capacity where the bottleneck actually is. More medical-school places do nothing if the true constraint is clinical placement capacity, supervising senior staff, or specialty training posts downstream. Find the binding constraint in the pipeline — often the number of training posts, not applicants — and invest there; opening the front door while the middle of the pipe is blocked produces frustrated trainees, not more consultants. Physician supply responds to the narrowest point, not the widest.

10. Plan the whole labour market, not one profession at a time. Doctors, nurses, pharmacists, therapists, and support workers are substitutes and complements; a decision to expand one changes the demand for the others. Plan the health human resources system as an interacting whole, so that a new-role expansion (say, physician associates) is matched by the supervision capacity and scope rules it needs, and does not simply create a bottleneck elsewhere. Single-profession planning is how systems end up with too many of one skill and too few of another.

11. Contain the agency and locum premium by fixing its cause. Heavy reliance on temporary staff at premium rates is usually a symptom — of unfilled substantive posts, poor rotas, or pay compression — not a cause, and capping agency rates without fixing the substantive offer just moves the shortage. Measure the true cost of temporary staffing, including its effect on continuity and team morale, and redirect the premium toward the substantive conditions that would make the permanent post attractive. A large agency bill is a message about your permanent offer.

12. Fold automation into workforce plans as augmentation, and count the transition costs. Technologies that automate documentation, triage, or image reading change the marginal product of clinical labour and can relieve the tasks that drive burnout (the evaluation of these tools belongs to Chapter 5.3 — AI Health Economics). Treat them as ways to release scarce clinical time toward higher-value work rather than as headcount substitution, and budget for the training and workflow redesign that realizing any saving requires. An efficiency that is never redeployed or that adds hidden oversight work is not a saving.

Questions to discuss with your team

1. Is our persistent vacancy a shortage of people, or a price we are unwilling to pay? This is the first fork in every workforce decision, and getting it wrong is expensive. If the post would fill at a higher wage, better rota, or improved conditions, then it is a pricing problem with a fast fix; if no realistic package would fill it, it is a genuine supply shortage needing a ten-year training response. The tension is that as a dominant employer you have both the power to hold wages down and the duty to keep posts safely filled, and the cheap short-run choice (leave the wage low, backfill with agency) can be the expensive long-run one. Concrete angles: what would it actually take to fill this specific post, and have we ever tested it? What are we already paying in agency premiums and safety risk to keep the substantive wage low? An honest answer names which kind of problem you have, with evidence, rather than defaulting to "we need more training places" because that defers the cost to a successor.

2. Where could we safely shift tasks to a different skill mix, and what would have to be true for that to be a real saving rather than a diluted rota? Task-shifting is one of the few levers that can improve access and cost at once, but it is also the easiest to fake — relabelling a gap as a "new model of care" while quietly reducing safe staffing. The genuine version requires training, supervision, legal scope cover, and the redeployment of freed senior time to higher-value work; the counterfeit version just spreads fewer skilled hands over the same patients. The honest debate weighs a real productivity gain against a real safety and professional-relations risk, and confronts scope-of-practice rules that may be defended as safety but function as territory. Concrete angles: which specific tasks, done by whom now, could whom do instead, and what is the evidence on safety? Where would the freed time go, and how would we know the saving was real? A good answer is specific about tasks and safeguards, not a slogan about "working at the top of licence".

3. What is our turnover actually costing us, and are we investing in retention as seriously as in recruitment? Systems reliably over-invest in filling posts and under-invest in keeping the people already in them, partly because recruitment is a visible line item and turnover is a diffuse one. The tension is that retention spend — flexible working, workload relief, career progression, burnout prevention — competes with front-line service budgets and shows its return slowly and indirectly. Yet replacing an experienced clinician means paying again for training you had already banked, plus recruitment and the productivity cost of a vacancy and a novice. Concrete angles: do we know our turnover rate by staff group and where in careers people leave? Have we costed a single experienced departure end-to-end? What would we spend to prevent it, and is that less than replacement? An honest answer treats the retention rate as a financial metric with a business case, not a wellbeing aspiration bolted on at year end.

4. Are we planning the workforce on the timescale of the asset we are building, or reacting to this year's vacancy list? A specialist takes a decade or more to train, so today's intake decisions set supply ten to fifteen years out, yet workforce planning is chronically hostage to the annual budget cycle and the current rota gap. The tension is that the horizon on which the workforce must be planned is far longer than the horizon on which anyone is held accountable, so the slow, structural fix is repeatedly deferred in favour of the visible short-run patch. It is also easy to open the wrong end of the pipe — expanding applicant intake when the true bottleneck is clinical placements, supervisors, or specialty training posts downstream. Concrete angles: do we have a stock-and-flow model — entrants, attrition at each stage, retirements, migration — stress-tested against demographic demand, or just a headcount? Where is the binding constraint in our pipeline, and are we investing there or at the widest point? Are we planning professions as an interacting whole, so that expanding one is matched by the supervision and scope it needs? An honest answer commits to the ten-year view even though the payoff lands after the current leadership has moved on.

5. How are we filling the posts nobody wants — rural, unsocial-hours, high-stress — and are we paying the differential honestly or in vacancies? Some posts are intrinsically harder to fill, and economics is blunt about it: a worse job needs a compensating differential, and a system that will not pay it in money pays it in empty rosters, agency premiums, and safety risk instead. The tension is that a dominant employer with a compressed, nationally-set pay structure often cannot or will not target extra reward at the specific hard post, so the shortage concentrates exactly where care is already thinnest. Non-pay currency — housing, protected training time, predictable rotas, career access, pensions — can be more cost-effective per post filled than a headline rise, but it takes deliberate design. Concrete angles: which specific posts are chronically hard to fill, and what would it actually take to fill each one? Are we targeting the differential at those posts, or inflating the whole pay bill and still not filling them? What are we currently paying, in agency and unsafe cover, to avoid paying the differential openly? An honest answer names the true price of the unpopular post rather than pretending a uniform wage will staff a non-uniform job.

6. Is our agency and locum bill a problem to cap, or a message about our permanent offer? Heavy reliance on temporary staff at premium rates feels like the disease, so the instinct is to cap the rate — but the premium is usually a symptom of unfilled substantive posts, poor rotas, or pay compression, and capping the rate without fixing the offer just relocates the shortage. The tension is that the agency line is politically visible and infuriating, which makes attacking it directly attractive even when the money would do more redirected into the substantive conditions that would make the permanent post attractive. Automation of documentation, triage, or image reading sits alongside this as a way to relieve the tasks that drive people to leave, but only if the freed time is genuinely redeployed rather than quietly reabsorbed. Concrete angles: what is our temporary staffing truly costing, including its effect on continuity and team morale? What would the substantive offer have to be for these posts to fill, and is that cheaper than the premium we now pay? Where could technology release scarce clinical time, and have we budgeted the training and workflow redesign to realize it? An honest answer treats a large agency bill as feedback on the permanent offer, not as an expense to be legislated away.

In practice: a health economics example

The Republic of Maridia is an upper-middle-income country whose care is delivered mainly through a tax-funded National Health Service that employs roughly four in five of the country's nurses — a textbook near-monopsony in the nursing labour market. The Ministry of Health faces a chronic problem: some 12% of hospital nursing posts sit vacant, wards routinely run below safe staffing, and the gaps are plugged by private-agency nurses at roughly double the substantive hourly cost. A newly appointed director of workforce is asked to fix the vacancies without breaching the pay envelope set by the finance ministry.

The instinctive proposals are to open two new nursing schools and to cap agency rates. The director insists on diagnosing the problem first. The pattern — many applicants to nursing courses, high early-career attrition, and vacancies that persist at the current wage — points less to an absolute shortage of people than to a wage and conditions set below the market-clearing level, exactly what monopsony predicts. Building schools would add supply in three to four years to a market that is haemorrhaging trained nurses in their first five years of practice. Capping agency rates without changing the substantive offer would simply relocate the shortage.

The director commissions a costing from the standpoint agreed with the finance ministry — a health-system perspective (the choice of perspective is developed in Chapter 2.1 — Economic Evaluation). The full cost of the status quo is made visible: the agency premium, the recruitment and induction cost of replacing each departing nurse, the productivity loss of vacancies and novices, and the safety incidents associated with stretched skill mix. Against that, three interventions are modelled. First, a targeted retention package — a modest unsocial-hours differential, guaranteed rota predictability, and funded post-registration career progression — aimed at the first-five-years attrition point. Second, a task-shifting programme that trains senior nurses as nurse practitioners and moves defined tasks from junior doctors to them, with supervision and revised scope of practice rules, freeing nursing time on the wards. Third, a smaller expansion of clinical placement capacity, since the pipeline analysis shows placements, not applicants, are the binding constraint.

The numbers favour retention first. Because the agency premium and replacement costs are so large, the modelling shows the retention package roughly paying for itself within two to three years by shrinking turnover and agency reliance — even though it raises the substantive pay bill, which is precisely why a pure pay-envelope view had blocked it. The task-shifting programme improves access and relieves junior-doctor workload but only registers as a saving where the freed senior-nurse time is redeployed to direct care rather than absorbed. The placement expansion is retained as the slow, structural fix, sized to the true bottleneck rather than to headline school numbers. The uncomfortable part of the analysis is explicit: the retention differential, once given, is hard to withdraw and will ripple across the pay structure, and the finance ministry must accept a higher substantive wage bill now to remove a larger agency-and-turnover bill later. The director's recommendation names that trade-off rather than hiding it — the monopsony wage was cheap on the pay line and expensive everywhere else.

Four sector lenses

Startup

A digital-health start-up offering, say, a nurse-led remote monitoring service lives or dies by a handful of scarce clinicians and cannot out-bid large employers on base pay. Its levers are the non-pay differential: flexibility, autonomy, purpose, equity, and freedom from the administrative burden that drives occupational burnout in large institutions. It can design roles around new scope of practice and task-shifting from the start, unencumbered by legacy demarcations, and its workforce model is often the product itself. The risk is thinness — a single clinician's departure can halt the service — so it must plan retention and cross-cover with a seriousness disproportionate to its size.

Small business

An established general-practice partnership, single clinic, or care home employs a small, stable team and competes for the same scarce clinicians as far larger neighbours, but without their pay scales, internal labour market, or training pipeline. Its edge is the opposite of the start-up's novelty: a known workplace with continuity, local roots, and a manageable rota, where a good employer reputation does much of the recruiting. Skill mix is its sharpest lever — a well-judged blend of practitioners, associates, and support staff can hold viability that an all-doctor model could not — but with thin margins it feels every locum shift and every departure acutely. Its retention task is personal and undramatic: predictable hours, a fair share of unsocial work, and a career path within a small structure, since it cannot absorb the turnover a large employer treats as background noise.

Enterprise

A large hospital group or insurer is usually the local monopsony buyer of labour and holds real wage-setting power, along with the responsibility that comes with it. Its scale lets it run internal labour markets, structured career ladders, and its own training pipelines, and lets it invest in retention and skill-mix redesign at a scale that pays back. The same scale makes it slow: pay structures ripple across thousands of staff, professional relations are institutionalized, and a large agency bill can become a permanent feature rather than a signal. Its central task is to manage the whole interacting health human resources system rather than fight vacancies profession by profession.

Government

A ministry or national payer sets the rules of the labour market itself: it funds and caps training places, defines licensure and scope of practice, and as the dominant employer in tax-funded systems sets the wage that shapes the whole market. Its decisions carry the longest lags and the widest equity consequences — where it opens training places and sets rural differentials determines who gets care a decade out (equity is developed in Chapter 3.4 — Equity). It must balance the fiscal attraction of monopsony wage restraint against the safety and access costs it creates, and it owns the interface with international recruitment and its ethics (Chapter 4.2 — Global Health and Trade). Its hardest discipline is resisting the false economy of a cheap wage line that reappears, larger, as agency spend, turnover, and unsafe care.

Common failure modes

• Treating a pricing problem as a supply problem. Commissioning a decade of new training to fix a vacancy that would clear with a better wage or rota. Fix: diagnose whether the post would fill at a higher price before building the pipeline.
• Cheap-wage, expensive-everything-else. Holding the substantive wage down and paying the difference in agency premiums, turnover, and safety risk. Fix: cost the status quo in full, not just the pay line.
• Task-shifting as disguised dilution. Relabelling a staffing gap as a "new model of care" while cutting safe staffing. Fix: require training, supervision, scope cover, and evidence of redeployed senior time before counting a saving.
• Scope-of-practice rules as protected territory. Defending demarcations as safety when the evidence shows a task is safe in other hands. Fix: review scope against evidence, and name protection of income where it is operating.
• Opening the wrong end of the pipe. Expanding applicant intake when placements, supervisors, or training posts are the real constraint. Fix: find and invest in the binding bottleneck.
• Running the workforce permanently hot. Booking apparent efficiency by understaffing, then losing it to burnout, absence, and exits. Fix: treat safe staffing and workload as capital preservation, not slack.
• Bonding without belonging. Relying on punitive return-of-service contracts while ignoring the conditions people would actually stay for. Fix: pair fair bonds with a workplace worth staying in.

Maturity model

Checklist

• Build a stock-and-flow workforce model on a 10-to-15-year horizon, not a budget-year one.
• For each persistent vacancy, test whether it is a shortage or a pricing problem before committing to training.
• Cost the full status quo — agency premium, turnover, vacancy productivity loss, safety risk — not just the pay line.
• Target compensating differentials at the specific hard-to-fill posts, using pay and non-pay levers.
• Review scope-of-practice boundaries against safety evidence, and identify any that protect income.
• Require training, supervision, legal cover, and measured redeployment before booking a task-shifting saving.
• Measure turnover by staff group and career stage, and cost at least one end-to-end departure.
• Fund a retention business case aimed at the point in careers where people actually leave.
• Identify the binding constraint in the training pipeline and invest there, not at the widest point.
• Design any training bond as a fair exchange of subsidy for service, paired with conditions people will stay for.
• Treat automation as time released toward higher-value work, and budget its transition costs.

Key sources

• World Health Organization — Global Strategy on Human Resources for Health and the National Health Workforce Accounts, the standing international framework for workforce planning and data.
• OECD — Health at a Glance, for comparative data on health-workforce numbers, remuneration, and skill mix across countries.
• World Health Organization — Global Code of Practice on the International Recruitment of Health Personnel, the reference for ethical recruitment (applied in Chapter 4.2 — Global Health and Trade).
• Economics Network — Health Economics for Teachers, health-policy and health-systems modules covering workforce and labour-market topics.
• Academic literature on monopsony in health labour markets and on nurse and physician labour supply, in journals such as the Journal of Health Economics and Health Economics.

References

1. Labour economics — Wikipedia — https://en.wikipedia.org/wiki/Labour_economics
2. Monopsony — Wikipedia — https://en.wikipedia.org/wiki/Monopsony
3. Occupational licensing — Wikipedia — https://en.wikipedia.org/wiki/Occupational_licensing
4. Human capital — Wikipedia — https://en.wikipedia.org/wiki/Human_capital
5. Health human resources — Wikipedia — https://en.wikipedia.org/wiki/Health_human_resources
6. Division of labour — Wikipedia — https://en.wikipedia.org/wiki/Division_of_labour
7. Scope of practice — Wikipedia — https://en.wikipedia.org/wiki/Scope_of_practice
8. Occupational burnout — Wikipedia — https://en.wikipedia.org/wiki/Occupational_burnout
9. Physician supply — Wikipedia — https://en.wikipedia.org/wiki/Physician_supply
10. Medical education — Wikipedia — https://en.wikipedia.org/wiki/Medical_education
11. Nurse practitioner — Wikipedia — https://en.wikipedia.org/wiki/Nurse_practitioner
12. Compensating differential — Wikipedia — https://en.wikipedia.org/wiki/Compensating_differential
13. Global Strategy on Human Resources for Health: Workforce 2030 — World Health Organization — https://www.who.int/publications/i/item/9789241511131
14. WHO Global Code of Practice on the International Recruitment of Health Personnel — World Health Organization — https://www.who.int/publications/i/item/wha68.32
15. Health at a Glance — OECD — https://www.oecd.org/en/publications/serials/health-at-a-glance_g1gh2ea9.html
16. Health Economics for Teachers — Economics Network — https://economicsnetwork.ac.uk/health/teachers