Kenneth Arrow in the Waiting Room

Audio article

Uses the speech voice supplied by your browser or device.

The plastic chair in a clinic waiting room does not invite economic theory.

It has a thin metal frame, a tired backrest, and a surface polished by other people’s waiting. Somewhere nearby a printer coughs. A file is opened. A receptionist says a name with the practiced neutrality of someone who has seen too many anxious faces in one morning. Outside, Calcutta keeps doing what Calcutta does: traffic, dust, small bargains, a man arguing near a pharmacy counter, someone buying strips of tablets after doing private arithmetic in the head.

This is not the usual entrance to Kenneth Arrow.

Arrow normally arrives in a cleaner room. He arrives through welfare economics, general equilibrium, the impossibility theorem, and the elegant cruelty of mathematical proof. He was already famous before he turned to medical care. At thirty-two, he had produced the result now called Arrow’s Impossibility Theorem, a blow to any simple dream that collective choices can always be made fairly, consistently, and rationally. He had won the John Bates Clark Medal. In 1972 he would become the youngest person to receive the Nobel Prize in Economic Sciences.

But in 1963, in the American Economic Review, Arrow published a paper with a dry title and a very live nerve: “Uncertainty and the Welfare Economics of Medical Care.” It ran for thirty-three dense pages. It did not behave like a speech. It did not offer a slogan. It did not say, in the pleasing manner of public argument, that one side was stupid and the other side was kind.

It did something more useful.

It explained why medical care is not a normal market.

That sounds simple until you sit with it. Not imperfect in the ordinary way. Not merely expensive. Not merely badly run. Not merely captured by interests, though capture is real. Not merely confused by policy, though policy often adds its own fog. Arrow’s point was colder and more durable: the problems of medical care are built into the thing itself.

The patient is not a shopper. The doctor is not a shopkeeper. The insurer is not a harmless cashier. The hospital is not a hotel with needles. The drug company is not selling a normal consumer product. The regulator is not an optional decoration added after the market has done its work.

Medical care begins in vulnerability. That is where the textbook starts to tear.

The Patient Is Not Shopping

The standard economic story likes a clear buyer and a clear seller. The buyer has preferences, information, time, and some freedom to walk away. The seller has a product. Price does its little dance between them. Competition disciplines arrogance. Choice rewards quality. Everyone may not be happy, but the machine has a certain dignity.

Now place that story inside a clinic.

The patient does not arrive with sovereign consumer confidence. The patient arrives unsure, uncomfortable, hurried, and dependent on knowledge that belongs to someone else. Pain narrows the mind. Fear narrows it further. Even without panic, the patient often has no clean way to know what is wrong, which test matters, whether a procedure is necessary, or whether the advice being given is evidence, habit, caution, incentive, or plain uncertainty in a white coat.

The information gap is not like buying a used car. A careful car buyer can learn enough to ask better questions. A patient cannot become an oncologist during a short consultation. A patient cannot learn surgical technique while lying on a trolley. A patient cannot read a pathology report into practical wisdom by staring at it harder.

Arrow understood that this creates delegation. The patient must hand over judgment. Not because the patient is foolish. Not because the patient is lazy. Because the knowledge required to decide is specialized, costly to acquire, and often unavailable at the moment when the decision must be made.

That is why trust enters the room.

Trust is not an extra courtesy in medical care. It is part of the operating system. You can pay for a consultation, a test, a bed, or a procedure. You cannot buy the inner assurance that the person advising you is acting for your good when you cannot verify the advice. That assurance has to be produced by ethics, training, reputation, licensing, professional culture, institutional rules, and repeated human conduct.

The market can price a service. It cannot manufacture trust on demand.

The Doctor Has Too Many Roles

The doctor in Arrow’s account is not a villain. That matters. A poor argument needs a villain. A serious argument needs a structure.

The doctor is at once expert, agent, and seller. Expert, because the doctor knows more than the patient. Agent, because the doctor is expected to act in the patient’s interest. Seller, because medical work is paid for, organized, billed, reimbursed, and often rewarded according to volume, procedure, status, or institutional position.

These roles do not line up neatly.

If a doctor recommends surgery, the patient usually cannot know whether the recommendation reflects the best available evidence, professional caution, local custom, defensive practice, financial incentive, or the simple fact that surgeons see the world through surgical possibilities. Many doctors act honorably. That does not remove the structural tension. A system cannot be designed on the assumption that character alone will hold up every beam.

This is why medicine developed ethical codes that go beyond ordinary commerce. The Hippocratic tradition, restrictions on advertising, rules against fee-splitting, licensing, peer review, professional associations, and standards of conduct are not quaint ornaments from a more solemn age. They are attempts to solve a market problem that price alone cannot solve.

When the buyer cannot judge quality, the seller’s conscience becomes economically relevant.

That sentence sounds strange only because economics often pretends conscience is outside the model. In medical care it is inside the machine.

Insurance Helps And Distorts

Insurance is the sensible human answer to financial catastrophe. Pay a predictable amount now so an unpredictable disaster does not destroy you later. It works cleanly enough for some risks: fire, car damage, life insurance. The event happens or it does not. The loss can usually be observed.

Medical care is messier.

Need is not always binary. It is continuous, uncertain, and contestable. One more test may be useful or wasteful. One more scan may reveal something important or create a new maze. One expensive drug may be a breakthrough, while another may offer only a small benefit at a very high price. The insurer cannot perfectly observe need. The patient does not face the full price. The doctor may not know the full price either. The bill that finally appears may look like an encrypted insult.

Arrow emphasized two classic insurance problems.

The first is moral hazard. Once insured, people may use more care than they would if they paid the full price at the counter. This does not require dishonesty. It follows from the reduced marginal cost. If an additional test costs the patient little at the point of use, the patient and doctor have less reason to economize, even when the system as a whole pays heavily.

The second is adverse selection. People who expect to need care are more eager to buy insurance than people who expect not to need it. If participation is voluntary, the insured pool becomes costlier, premiums rise, healthier people leave, and the market can unravel. This is why many developed countries either require broad participation or fund care through taxation.

Insurance also breaks the price signal. In an ordinary market, the price a buyer sees is connected to the cost of the thing. In medical care, the patient may pay a small copay for something billed at thousands. The doctor may bill one number, the insurer may allow another, and the hospital may record a third. No single actor faces the whole cost in a clean way.

Insurance is therefore both necessary and distorting. Without it, many people could not bear the cost of serious care. With it, the ordinary discipline of price becomes muffled.

This is not a small administrative inconvenience. It is one reason medical care keeps refusing to become a tidy market.

The Other Actors In The Room

Arrow’s drama contains more than patient, doctor, and insurer.

The hospital matters. It is capital, labor, technology, emergency readiness, teaching, research, billing, regulation, community expectation, and institutional survival packed into one building. Calling it a firm is not wrong, but it is thin. A hospital must hold expensive equipment, specialized staff, and standby capacity for events that may or may not happen at any convenient hour.

The pharmaceutical company matters. Drug development is expensive, uncertain, and regulated. Many candidates fail. The few that succeed must carry the cost of the failures. Companies argue that high prices fund research and development. Critics reply that public funding often supports foundational science, that marketing can absorb enormous money, and that patents create monopoly power over products some patients may need urgently.

Both sides contain truth. That is what makes the issue hard rather than merely noisy.

The regulator matters. In the United States, agencies such as the FDA and CMS, along with state medical boards and licensing bodies, try to protect patients, ensure quality, and control costs. They work with incomplete information, limited resources, political pressure, and the risk that regulated industries may learn to bend regulation toward themselves.

The government matters everywhere. It funds research. It subsidizes medical education. It regulates insurance. It provides care directly in systems such as the VA and the NHS, finances care through programs such as Medicare, sets prices in many countries, and uses taxation to redistribute the burden of illness.

Government is not an alien visitor interrupting a clean private market. In Arrow’s framework, government involvement arises because the private market is already unstable, incomplete, and morally exposed.

What Is Being Bought?

A haircut is a service. Legal advice is a service. A meal is a service mixed with a good. Medical care is also a service, but the resemblance misleads.

When you buy medical care, you often buy uncertainty arranged in a professional form. The outcome is not guaranteed. The quality may be hard to observe. The need is often involuntary. The timing may be urgent. The price may not correspond to value in any intuitive sense. And the effects may spill beyond the patient.

A surgery may help, fail, or harm. A diagnosis may be correct, incomplete, or overtaken by new information. A treatment that looks sensible for a population may not work for a particular body. A cheap vaccine can protect many people beyond the person who receives it. A careless use of antibiotics can help resistant bacteria become everyone’s problem. A public health surveillance system may benefit people who never know it exists.

These spillovers are externalities. That word sounds like something dropped from a bureaucrat’s shelf, but it means a simple thing: some costs and benefits fall on people outside the transaction. Markets handle private wants more easily than shared consequences. Medical care is full of shared consequences.

Arrow’s point was not that each feature is unique. Other markets have uncertainty. Other markets have expert sellers. Other markets have externalities. The difference is the concentration. Medical care combines urgency, fear, expert dependence, hidden quality, insurance distortion, social spillover, and moral seriousness in one place.

That combination is the architecture.

A Short History Of The Problem

The problem did not begin in 1963.

For most of human history, care was mixed with family duty, charity, religion, craft, guesswork, and social authority. The Hippocratic tradition emerged in ancient Greece around the fifth century BCE and gave the physician an ethical identity, but much actual care remained untested, local, and uneven. The patient could not evaluate the healer. Often the healer could not evaluate the treatment either.

The nineteenth century changed the power of medicine. Germ theory, developed by Pasteur, Koch, and others, made infection intelligible. Anesthesia in the 1840s made surgery bearable. Lister’s antisepsis in the 1860s made surgery safer. Medical schools expanded. Professional associations formed. Licensing laws appeared, partly to protect the public from quackery and partly to protect the profession from competition.

The information gap did not disappear. It became more legitimate. The trained physician really did know more. That made trust more necessary, not less.

Health insurance developed later. In the United States, Baylor University Hospital in Dallas offered a prepaid hospital plan to teachers in 1929. Blue Cross and Blue Shield plans expanded in the 1930s. During World War II, wage controls encouraged employers to offer health insurance as a fringe benefit. In 1954, the tax treatment of employer-sponsored insurance helped cement the arrangement.

By the time Arrow wrote in 1963, the United States had a patchwork of private insurance, public provision for some groups, and many people left outside reliable coverage. Medicare and Medicaid would arrive in 1965. Elsewhere, countries had already chosen different paths: the UK’s National Health Service began in 1948; Germany’s social insurance model had older roots and kept expanding; France and Canada developed mixed and single-payer arrangements of their own.

Arrow’s paper did not immediately remake policy. It was too analytical for that. But it helped create the intellectual terrain of health economics. Later work examined moral hazard, adverse selection, supplier-induced demand, cream-skimming, risk selection, price regulation, payment incentives, and the strange fact that more market language does not automatically produce a better medical system.

The Affordable Care Act of 2010, with its individual mandate, exchanges, and subsidies, can be read as one American attempt to address adverse selection while keeping a largely private insurance structure. Debates over Medicare for All, public options, and single-payer systems are debates over how radical the response to Arrow’s diagnosis should be.

Different Countries, Same Underlying Problem

Arrow wrote from the American context, but the problem travels.

The United States spends more on healthcare per person than any other country, roughly twice the average of other wealthy nations, while performing worse on many broad measures such as life expectancy, infant mortality, chronic disease management, and uneven access. Tens of millions have been uninsured or underinsured. The system relies heavily on private insurance, which brings administrative costs, billing complexity, marketing, underwriting, claims disputes, and barriers created by deductibles, copays, and coinsurance.

Administrative costs in the United States have often been estimated around 8 percent of total healthcare spending, compared with roughly 1 to 3 percent in countries with single-payer or more integrated systems. Fee-for-service payment rewards volume. Pharmaceutical prices are often far higher than in peer countries. The defense is innovation. The objection is monopoly power, public subsidy of basic science, and pricing that can become detached from practical cost.

The United Kingdom chose another bargain. The National Health Service, created in 1948, funds care through general taxation and provides much of it free at the point of use. It reduces the fear of financial ruin from ordinary access to care. It also faces waiting lists, rationing, budget politics, and the familiar problem that demand at a zero point-of-use price can outrun supply.

Germany uses social insurance funded largely through payroll contributions, with sickness funds and risk pooling. France combines public insurance with private supplements. Canada uses single-payer financing for medically necessary hospital and physician services while leaving other areas more mixed.

None of these systems is pure. None is painless. But they are all institutional answers to the same Arrow problem: medical care cannot be left to ordinary consumer choice and price competition without producing serious failures.

In lower-income countries, the problems become harsher. There may be fewer trained physicians, weaker regulation, more out-of-pocket payment, less insurance, fragile public provision, counterfeit drugs, and limited administrative capacity. The World Health Organization, the World Bank, and many NGOs try to help with finance, technical assistance, and public health programs, but imported models often break against local constraints.

Arrow’s framework still helps. It does not supply a ready-made answer. It explains why the answer is never simple.

The Machinery Of Failure

The first mechanism is information asymmetry.

One party knows more than another. In healthcare the gap is extreme. The doctor knows more about diagnosis and treatment. The patient knows more about symptoms, tolerance, preference, and ordinary life. The insurer knows more about coverage rules, pricing, and reimbursement. The pharmaceutical company knows more about its product and trial data. No one has the whole map.

The asymmetry is also expensive to reduce. Clinical trials cost money and time. Peer review is imperfect. Quality measurement can be gamed. A patient portal may release more data, but data is not interpretation. A reputation may reflect bedside manner, marketing, institutional prestige, or simple luck.

The second mechanism is uncertainty.

Arrow distinguished risk, where probabilities are known, from uncertainty, where they are not. Medical care often lives closer to uncertainty. There is diagnostic uncertainty: what is the problem? There is therapeutic uncertainty: will the treatment work? There is prognostic uncertainty: what happens later? There is financial uncertainty: what will the bill become once the system is done translating suffering into codes?

Insurance is designed for risk. Medical care often supplies uncertainty wearing the clothes of risk.

The third mechanism is trust.

Trust is an economic institution, not merely a pleasant emotion. In ordinary commerce, contracts, warranties, and courts can repair some failures. In medical care, contracts are incomplete. You cannot contract for a cure. You cannot sue away every bad outcome, because bad outcomes can occur even when care is careful. Law can punish gross negligence. It cannot supervise every clinical judgment.

So trust fills the gap. The patient trusts the doctor. The doctor trusts the patient to report honestly. Society trusts the profession to discipline itself. The profession trusts institutions to support good practice. When commercialization, conflicts of interest, scandal, or political theater erodes trust, the economic system is damaged too.

The fourth mechanism is externality.

Vaccination protects others. Antibiotic misuse can endanger others. Sanitation, disease surveillance, and health education are public goods. A public good is something from which people benefit even if they do not individually pay for it, which means ordinary markets tend to underprovide it.

Public health has always known this. Economics sometimes has to relearn it in more expensive language.

Technology Changes The Surface, Not The Structure

Arrow wrote before much of modern medicine’s present machinery existed.

Diagnosis now uses CT scans, MRIs, PET scans, biomarker tests, genetic testing, and AI systems that can detect patterns in medical images. These tools can improve accuracy and standardize some decisions. They can also produce false positives, raise costs, widen inequality between rich and poor systems, and create new expert gaps. A genetic test may say someone has a 40 percent lifetime risk of breast cancer. It cannot say exactly what will happen, when, or what choice will feel bearable.

Drug development has moved from serendipity and trial-and-error toward molecular biology, computational chemistry, targeted pathways, cell cultures, animal models, and human trials that may involve tens of thousands of participants and cost billions. The results include astonishing successes: hepatitis C cures, HIV transformed into a manageable chronic condition, and cancer treatments that can extend life. The same system also produces marginal benefits at astronomical prices, heavy marketing, and patent-protected monopoly power.

There are alternative models: more public funding, prizes for useful discoveries, compulsory licensing for essential medicines. Each has trade-offs. Arrow does not tell us which is best. He tells us why the drug market is not ordinary. The patient usually does not choose the product. The doctor prescribes it. The insurer often pays. The patent holder may control supply. The need may be urgent.

Information technology has its own promise. Electronic health records can reduce errors, improve coordination, and provide data for research and quality improvement. Telemedicine can help rural and underserved populations reach clinicians. Wearables can monitor signals in real time.

They also create burdens. EHRs can be cumbersome and can pull attention away from the person in the room. Telemedicine cannot replace every physical examination. Wearables can be inaccurate or needlessly alarming. Data can leak. Insurers may find new ways to classify risk. WebMD and a wrist device do not make a patient equal to a physician with years of training.

Information helps. It does not abolish expertise.

Payment reform tells the same story. Fee-for-service remains powerful in the United States and rewards more visits, tests, and procedures. Value-based care tries to pay for quality and outcomes instead of volume. Accountable care organizations, bundled payments, pay-for-performance, and capitation all try to bring incentives closer to patient welfare.

Each model solves one problem and creates another. Quality is hard to measure. Metrics can be gamed. Providers may avoid very sick patients. Capitation can encourage underprovision. Fee-for-service can encourage overuse. The perfect payment system remains unavailable, which is another way of saying Arrow was onto something.

What Arrow Did Not Say

It is easy to misuse Arrow.

He did not prove that markets have no role in healthcare. Competition among hospitals may improve some forms of quality. Consumer choice among insurance plans may matter. Price transparency may help for services that are shoppable, non-urgent, and understandable.

But those are partial tools. They do not turn the whole sector into a grocery aisle.

Arrow also did not write a campaign pamphlet for socialized medicine. The paper was analytical, not a manifesto. It identified structural problems and institutional responses: insurance, professional ethics, regulation, public provision, and non-market norms. Arrow himself was a social democrat, but the paper can illuminate the British NHS, the German social insurance system, the Canadian single-payer model, and a reformed American mixed system without becoming identical to any of them.

The paper has limits. It focused heavily on acute medical care. Preventive care, chronic disease management, long-term care, rehabilitation, and community health have different rhythms. Some decisions allow more time. Some relationships reduce information gaps over repeated visits. Some systems depend more on household labor and local infrastructure than on hospitals and insurers.

The paper also assumes a developed-country setting. In places with weak institutions, thin insurance markets, underfunded public systems, and uneven professional capacity, the same structural problems appear in rougher form. The policy answer has to fit the institutions available, not the institutions imagined.

The unresolved question remains: how much market, how much government, how much professional self-rule, how much public financing, how much patient choice?

Some argue for a largely public system. Some argue for regulated markets with universal coverage. Some argue for health savings accounts, catastrophic insurance, and price transparency. Cross-country evidence suggests that systems with more public involvement often achieve better outcomes at lower cost, but the comparison is tangled with culture, population health, state capacity, and history.

There may be no one correct mix for all countries.

That is not an excuse for confusion. It is an argument for humility.

The Calcutta Version Of The Argument

In Calcutta, theory has to pass through ordinary inconvenience before it becomes real.

You can explain information asymmetry in a seminar. Or you can watch a person at a pharmacy counter nod at instructions they only half understand, because the line behind them is growing and the language of medicine has become a small locked room. You can explain moral hazard with diagrams. Or you can notice how no one knows the true price of anything until the bill arrives from somewhere else. You can explain trust as an economic institution. Or you can see how quickly a household becomes quiet when a doctor’s expression changes.

Arrow’s paper matters because it does not flatter our slogans.

It does not say the market is evil. It says the market has conditions under which it works, and medical care violates many of them.

It does not say government is magic. It says government enters because the private structure cannot bear the full load.

It does not say doctors are saints. It says professional ethics exist because the patient cannot evaluate the seller.

It does not say technology will save us. It says information is not the same as judgment.

It does not say trust is soft. It says trust is infrastructure.

This is why the paper still feels alive more than six decades later. We keep wanting healthcare to become simpler than it is. The right wants shopping. The left wants guarantee. The technologist wants data. The insurer wants risk classification. The hospital wants solvency. The patient wants relief, clarity, and not to be ruined.

All of these wants collide in one fluorescent room.

Healthcare policy is hard because every solution injures some value. Free care at the point of use creates rationing pressures. Cost-sharing discourages unnecessary use but also necessary use. Private insurance can expand options but creates administrative waste and selection games. Public systems can control costs but risk queues, underfunding, and political neglect. Patents reward invention but create monopoly prices. Price controls protect buyers but may reduce incentives if done crudely. Transparency helps only where the buyer can understand, compare, wait, and choose.

The patient in the chair is not an abstraction. Neither is the taxpayer. Neither is the doctor at the end of a long shift. Neither is the researcher whose failed compound disappears after years of work. Neither is the regulator trying to guard the public with imperfect tools.

Arrow’s gift was to make the difficulty visible without pretending to dissolve it.

Where It Leaves Us

Kenneth Arrow died in 2017 at ninety-five. The paper remains because the waiting room remains.

We still argue about medical care as if one clean principle could settle it. Markets. Rights. Choice. Efficiency. Equity. Innovation. Discipline. Compassion. Each word contains something necessary. Each becomes foolish when it tries to rule alone.

The lesson is not that medical care should have no market. The lesson is that the market must be handled as one instrument among others, and a dangerous one when used without regard for vulnerability, uncertainty, and trust.

The patient is not a consumer in the ordinary sense. The doctor is not a salesperson in the ordinary sense. The insurer is not a neutral pipe through which money flows. The hospital is not a normal firm. The drug patent is not a normal monopoly. The public health system is not charity. It is shared protection against risks that cross household walls.

Arrow did not give us a cure for healthcare policy. He gave us a diagnosis of its permanent discomfort.

That may be the most useful thing an economist can sometimes do: not solve the human problem, but stop us from lying about its shape.

The waiting room chair remains hard. The printer coughs again. Someone stands when their name is called. The door opens for a moment, then closes, and the rest of us return to waiting with more dependence than the market knows how to price.

---

P.S. References

Arrow, K. J. (1963). Uncertainty and the Welfare Economics of Medical Care. American Economic Review.