The Starfish and the Mussel Bed

In the summer of 1963 a young zoologist named Robert Paine began visiting a stretch of rocky shore at Mukkaw Bay, on Makah land at the northwest tip of Washington State, with a crowbar. At low tide he pried every ochre starfish he could find off a strip of rock about eight meters long and threw each one as far out to sea as his arm allowed. He came back and did it again, month after month, for years. A matching strip nearby he left alone.

The rock he left alone kept what it had: barnacles, limpets, chitons, whelks, anemones, a sponge, several kinds of algae, and mussels, some fifteen species sharing the surface. On the cleared strip the mussels spread. Starfish eat mussels, and with nothing eating them the mussels did what mussels do, which is to grow shoulder to shoulder until there is no bare rock left. Within a few years the cleared strip held about half as many species as before, and it was heading toward one: a solid black bed of Mytilus californianus from the top of the strip to the bottom.

Paine published the result in 1966 and later gave the starfish its name, a keystone species. What he had shown was stranger than it sounds. He had made the shore safer, removed its chief predator, and the reward for that safety went to a single occupant.

This is the first time this collection of essays has taken up health care, and it starts on a tide flat on purpose. The argument now running through American politics, between the single public system Bernie Sanders has proposed for a decade and a market in medicine that almost nobody has seen working, is usually held as an argument about compassion and cost. Underneath it is an older question about monopoly: where it comes from, and what kind of surroundings it needs.

One winner in a quiet jar

Thirty years before Paine, a Moscow biologist in his early twenties named Georgii Gause had run the same experiment in glass. He put two species of Paramecium in a tube of fluid, fed them a measured ration of bacteria each day, held the temperature fixed, and counted. Each species grown alone did fine. Grown together, one always crowded the other out, and within a few weeks the tube held a single species. Gause tried it many ways. As long as the two were living on the same food under the same unchanging conditions, there was one survivor.

Garrett Hardin gave the result its modern name in 1960, the competitive exclusion principle: complete competitors cannot coexist. In a steady environment, competition is a process with an end, and the end is one.

The next year G. Evelyn Hutchinson, at Yale, asked the obvious question from the other side. A lake in summer holds dozens of species of plankton, all floating in the same water and living on the same few nutrients and the same light. By Gause’s rule there should be one. Hutchinson called it the paradox of the plankton, and his answer was that the lake is never the quiet tube. The season turns, the wind mixes the water, a cold week favors one alga and a warm week another, and the conditions change before any winner has finished winning. Joseph Connell made a general case of it in 1978, looking at rain forests and coral reefs: the richest communities are found where disturbance is frequent enough to keep interrupting the contest, and where disturbance stops, variety drains away.

A market is not a jar of paramecia, and firms can do what a protozoan can’t, which is change their own nature in a season. Still, the pattern carries. Competition under fixed conditions rewards whoever fits those conditions best, and each year the conditions hold, that fit compounds: costs fall with scale, relationships deepen, rivals with a different idea find no opening that suits them. What keeps a market plural is much the same as what keeps a lake plural. New technology arrives, tastes shift, a supplier fails, a cheaper method turns up from abroad, and the leader’s advantages turn out to have been advantages only under last decade’s weather. Joseph Schumpeter’s phrase for this was creative destruction, and it is the economy’s starfish.

The rare natural monopoly

Monopoly, Natural and Otherwise went looking for monopolies that stand without help and found few. A natural monopoly, in the economist’s strict sense, is a business where one provider can serve everyone more cheaply than two, and nearly every real example turned out to be a pipe, a wire or a rail laid along a single strip of land. Most of the others, looked at closely, had a public hand holding a door shut: a charter, a franchise, a dock lease, an exemption passed as a rider.

Medicine has very little of the natural kind. A county of eight thousand people may support one hospital and no more. A region needs one trauma center with a helicopter, not three. That is about where the list ends. A knee replacement, a cataract operation, the management of diabetes, a course of physical therapy, a pregnancy, a blood panel, a prescription: none of these gets cheaper because one organization does all of them for a whole state. Clinics the size of a dentist’s office do most of this work well.

So if American medicine has ended up concentrated, and it has, natural monopoly isn’t much of an explanation. The quiet jar is a better one.

The stabilizers

There is a view of government, held in different forms by both American parties, that its job is to steady things. Markets swing, the weather turns, firms fail, and people get hurt in the swings, so the state stands between citizens and the fluctuation: it smooths prices, guarantees supply, props up what is failing, and fixes standards so that nothing surprising happens. The impulse is decent. The Forest Service acted on it in 1935 when it resolved to put out every fire by ten o’clock the next morning, and this collection has described elsewhere what eighty years of steadied forest looks like when it finally burns.

Health care in the United States has been steadied, one measure at a time, for most of a century, and nearly every measure had a good reason behind it.

In 1910 Abraham Flexner, a schoolteacher hired by the Carnegie Foundation, published a survey of all 155 medical schools in the United States and Canada. Many were as bad as he said: diploma mills with no laboratory and no patients. The report, taken up by the American Medical Association and the state licensing boards, set a single standard on the model of Johns Hopkins, and schools that couldn’t meet it closed. By the mid-1930s there were 66. Five of the seven medical schools that trained Black physicians were gone, along with most of those that took women, most of the homeopathic and eclectic schools, and the night schools a working man could afford. The profession that came out the other side was better trained and a great deal smaller, and its numbers have been set by committee ever since. Since 1997 Congress has capped the residency positions Medicare pays for, and with them the number of new doctors.

In 1942 Congress passed the Stabilization Act, which froze wages for the length of the war. Employers who couldn’t raise pay competed for workers with benefits instead, the War Labor Board ruled that health insurance didn’t count as wages, and the tax authorities agreed that premiums an employer paid weren’t the worker’s income. The arrangement was written into the tax code in 1954 and has never left. An American’s insurance has been chosen by an employer ever since, bought with untaxed money, and lost with the job. The patient stopped being the buyer.

In 1964 New York began requiring anyone who wanted to build a hospital or add beds to prove to a state board that the community needed them. Congress pushed the idea on every state in 1974, on the theory that an empty bed is waste and planners could prevent it. The federal mandate was repealed in 1986, but about thirty-five states still have a certificate-of-need law in some form, and in most of them the existing hospitals may appear at the hearing to argue that a competitor isn’t needed. In 2010 the Affordable Care Act added a rule barring new physician-owned hospitals from Medicare, which in practice bars them.

Add the rest. Prices for most of what a doctor does descend from a schedule Medicare sets with the advice of a committee of the American Medical Association, written in billing codes the Association owns. Each new rule about records, reporting and coverage is a cost that falls the same on a two-doctor practice and a two-hundred-hospital system, and only one of them has a compliance department. A big hospital in trouble is too important to its city to be allowed to close.

Every one of these holds something still. Together they have built a tube with a fixed ration and a fixed temperature, and the result is what Gause would have predicted. By the measure antitrust lawyers use, the hospital markets of most American metropolitan areas are now highly concentrated. Independent practices have been bought up for two decades, and one insurance company, through its subsidiary, employs or contracts with roughly one American physician in ten. Americans often hear this described as what a free market in medicine produces. It is nearer to Paine’s cleared strip: a shore protected from disturbance, covered edge to edge by a few large tenants.

The finished monopoly

Seen this way, Medicare for All is less the opposite of the present system than its completion. The bill Senator Sanders has introduced in each Congress since 2017 would enroll every resident in one federal plan, pay every doctor and hospital from it, and forbid private insurance that duplicates what it covers. It takes the direction of the last eighty years and follows it to the end of the road.

The case for doing so deserves its full weight. One plan means one risk pool with everyone in it, so nobody is priced out for being sick. One payer means no billing departments fighting claims departments, and its advocates point to Medicare’s overhead of a few percent against well over ten for private insurers. One buyer can name its price to drug makers and hospital chains, which no fragmented buyer can. And there is the plain comparison: the United States spends about seventeen percent of its income on health care, the other rich countries ten to twelve, and they cover everyone and live longer. Against a private oligopoly that the state built anyway, a public monopoly at least answers to voters.

That last argument is the strongest, and it is strong only against the cleared strip. A single system is a single environment. It has one budget, set each year by a legislature with other things on its mind, and when the budget runs short there are no prices to rise, so the queue lengthens instead. Canada’s Supreme Court, striking down Quebec’s ban on private insurance in 2005, put it in one sentence: access to a waiting list is not access to health care. The Fraser Institute’s survey of Canadian physicians put the median wait from a family doctor’s referral to treatment at thirty weeks in 2024, and the waiting list of England’s National Health Service has stood above seven million since 2022.

A single system also has one standard of treatment, decided by one body. A patient whose case fits the standard is served. A patient whose case doesn’t, or who would trade a risk differently than the committee did, has no second door. And one buyer naming its price is a monopoly on the buying side, with the usual effect: fewer people find it worth their while to sell.

It’s worth saying that “every other rich country” doesn’t describe one thing. Britain’s system is socialized in the full sense, with state-owned hospitals and salaried staff. Canada’s provinces are single payers buying from private doctors. Germany, the Netherlands and Switzerland cover everyone through competing insurers. Germany’s sickness funds, about ninety-five of them, descend from the workers’ mutual funds Bismarck’s law of 1883 was built on. The comparison shows that universal coverage is possible at lower cost. It doesn’t show that a monopoly is what achieves it.

Is health care special?

The serious argument that medicine can’t be left to markets was made by Kenneth Arrow in 1963. Illness is unpredictable, so the need can’t be budgeted like groceries. The patient knows far less than the doctor and can’t judge the product even after buying it. Recovery can’t be returned for a refund, and a mistake may be final. A customer in an ambulance is in no position to compare prices. For Arrow, the profession’s licenses, ethics codes and nonprofit hospitals were society’s substitutes for a market that couldn’t form.

Each point is true of some part of medicine, and the parts differ. Emergency care is as Arrow describes it, and by most estimates it is only a few cents of the health care dollar. Catastrophic illness is unpredictable, which makes it a matter for insurance, a thing markets have supplied for centuries. The larger share of what medicine does is neither sudden nor mysterious: chronic conditions managed over years, planned surgery, childbirth, tests, drugs taken daily. People choose among these the way they choose a lawyer or a mechanic for a problem they don’t fully understand, by reputation, by second opinion, by a trusted adviser, by a seal on the wall.

Importance doesn’t settle it either. Food matters more than medicine and more urgently, and nobody proposes one national grocer. What sets medicine apart is the gap in knowledge, and here the argument for a single generic standard runs into the history of single standards.

In the autumn of 1793 yellow fever came to Philadelphia, and Benjamin Rush, signer of the Declaration and the most eminent physician in the country, treated it with mercury purges and bleeding, sometimes most of a patient’s blood over a few days. Six years later George Washington, with a throat infection, was bled of something like eighty ounces in half a day by three regular physicians following the best practice of their time. This was the standard of care. During London’s cholera epidemic of 1854, the Homoeopathic Hospital lost about one patient in six and the Middlesex Hospital about half. Homeopathy’s remedies were water. Water was an improvement on bleeding a patient already dying of fluid loss.

A standard is a monoculture of belief. When it’s right, everyone gets the benefit. When it’s wrong, everyone gets the error at once and there is nobody outside it to show the difference. The frontal lobotomy won a Nobel Prize in 1949. For decades ulcers were a disease of stress and acid, until Barry Marshall, unable to persuade his colleagues that a bacterium caused them, drank a culture of it in 1984. In 2001 American hospitals were given a uniform standard requiring that pain be assessed and treated in every patient, and prescriptions of opioids rose in every state together. The regulator’s record has its triumphs too. Frances Kelsey, a reviewer at the Food and Drug Administration, kept thalidomide off the American market in 1960 on her own stubbornness. Both kinds of story are stories about one gate, and how much rides on who is standing at it.

Quacks and specialists, side by side

The alternative has also been tried, and it isn’t pretty to look at.

In the 1830s and 1840s, state after state repealed its medical licensing laws, and by mid-century almost none remained. Samuel Thomson, a New Hampshire farmer, sold the right to practice his patented system of lobelia and steam baths for twenty dollars a family. Homeopaths, hydropaths, eclectics and bonesetters hung their signs beside the regulars. Later came the patent medicines: Mrs. Winslow’s Soothing Syrup, which quieted teething babies with morphine, and Radithor, radium dissolved in water, which killed the Pittsburgh industrialist Eben Byers in 1932 after he had drunk some fourteen hundred bottles.

Two things are worth noticing in that record. The irregulars flourished because regular medicine was bleeding and poisoning its patients, and a good part of what they sold was a refusal of that. And the remedy that worked against the worst of the patent medicines was a modest one. The Pure Food and Drug Act of 1906 didn’t decide what people could take. It required the label to say what was in the bottle, and sales of the morphine syrups fell once mothers could read the word.

For a present-day picture, India is the nearest thing to medicine without a steadying hand. Much of the care in its villages comes from practitioners with no qualification at all, and the harm is real. In the same country, in 1976, a retired eye surgeon named Govindappa Venkataswamy opened an eleven-bed clinic in Madurai and set out to do cataract surgery the way a kitchen turns out meals. Aravind Eye Care now performs several hundred thousand operations a year with outcomes that match Western hospitals, treats around half its patients free or below cost out of what the paying half brings in, and began making its own lenses in 1992 when imported ones cost too much, at a small fraction of the import price. In Bangalore, Devi Shetty’s hospitals do open-heart surgery for a couple of thousand dollars. Neither came out of a plan. They came out of a shore where anything could attach, and plenty that shouldn’t have did.

The United States has pockets of the same thing. The Surgery Center of Oklahoma, owned by its doctors, has posted a single all-in price for each operation on its website since 2009 and takes no insurance. A few thousand direct primary care practices sell unlimited visits for a monthly fee. A century ago the fraternal lodges did something similar, hiring a doctor for a dollar or two a member each year, until the medical societies expelled the physicians who took such contracts.

A free market in medicine looks like this: the brilliant and the worthless next door to each other, the highly specialized and personal treatment available because nobody had to approve it first, and the patient carrying more of the work of telling them apart. Accreditation and Liberty described how that work gets shared without a single gate: fraud is a crime, labels say what is in the bottle, and seals compete, so that a board certification, a hospital’s reputation or a lodge’s choice of doctor means something because it could be lost to a better one.

Discipline

A market like that doesn’t appear when government steps back and looks away. It takes a discipline that runs against the stabilizing instinct at nearly every point, and the discipline is mostly a matter of declining to do things. It means leaving the need for a new clinic to be judged by whoever is willing to build it. It means a failing hospital is sold or closes, a hard thing in a town with one. It means the tax code stops tying insurance to a job, a nurse practitioner or a surgeon trained in Madurai may practice what she can show she knows how to do, and a price is posted before the service is given. Each of these lets some instability back onto the rock.

That is a frame shift, and health care may be where it is hardest. In the frame most people now hold, fluctuation is the harm and a government earns its keep by absorbing it. In the other, fluctuation is what keeps any occupant from taking the whole surface, and the harm to guard against is the stillness. Interfaces argued that a transition has to be made through the layer where two systems touch. Here that layer already exists in the cash-price surgery centers and the monthly-fee practices, which are small, legal, and growing where the rules allow.

One more distinction keeps the choice from being crueler than it has to be. Who pays and who provides are separate questions, and the monopoly argument concerns only the second. Thomas Paine, in Agrarian Justice, proposed paying every person a sum from the ground rent of the country and left what they did with it to them. A society can put money in the hands of the poor and the sick, by a dividend or an account or a lodge’s common fund, and still leave the shore open. A voucher does it with a narrower purpose: public money that can be spent only on care, at whichever clinic or insurer its holder picks. Food stamps work this way, and they are how the United States feeds its poor without a national grocer. The premium subsidies that Swiss and Dutch households of small means receive are vouchers by another name, good at any of the competing insurers. What a single payer joins together, the purse and the provider, can be had apart.

Paine the zoologist kept his strip at Mukkaw Bay clear for several years and then stopped. The starfish came back. They ate the small mussels at the edges, and the bed stayed. A California mussel left alone for a few years grows thicker and longer than an ochre star can open, and the animals on the cleared strip had passed that size. Ecologists returning in later decades found the black band still there, on rock with starfish all around it.

---

Sources

• Robert T. Paine, “Food Web Complexity and Species Diversity,” American Naturalist 100 (1966); “A Note on Trophic Complexity and Community Stability,” American Naturalist 103 (1969), which introduces “keystone species.”
• Robert T. Paine et al., “Perturbation and Recovery Patterns of Starfish-Dominated Intertidal Assemblages in Chile, New Zealand, and Washington State,” American Naturalist 125 (1985), on the persistence of the Mukkaw Bay mussel bed.
• G. F. Gause, The Struggle for Existence (1934).
• Garrett Hardin, “The Competitive Exclusion Principle,” Science 131 (1960).
• G. Evelyn Hutchinson, “The Paradox of the Plankton,” American Naturalist 95 (1961).
• Joseph H. Connell, “Diversity in Tropical Rain Forests and Coral Reefs,” Science 199 (1978).
• Joseph Schumpeter, Capitalism, Socialism and Democracy (1942), ch. 7.
• Abraham Flexner, Medical Education in the United States and Canada, Carnegie Foundation Bulletin No. 4 (1910).
• Paul Starr, The Social Transformation of American Medicine (1982), on licensing repeal, the irregular sects, Flexner, and the origins of employer insurance.
• Stabilization Act of 1942, Public Law 77-729; Internal Revenue Code of 1954, § 106.
• National Health Planning and Resources Development Act of 1974, Public Law 93-641; Patient Protection and Affordable Care Act (2010), § 6001; Balanced Budget Act of 1997, on residency caps.
• Medicare for All Act, introduced by Sen. Bernie Sanders in 2017 (S. 1804) and in later Congresses.
• Kenneth J. Arrow, “Uncertainty and the Welfare Economics of Medical Care,” American Economic Review 53 (1963).
• Chaoulli v. Quebec (Attorney General), [2005] 1 S.C.R. 791.
• Fraser Institute, Waiting Your Turn: Wait Times for Health Care in Canada (2024).
• David T. Beito, From Mutual Aid to the Welfare State: Fraternal Societies and Social Services, 1890–1967 (2000), on lodge practice.
• Pure Food and Drug Act of 1906; Kefauver–Harris Amendment of 1962.
• V. Kasturi Rangan, “The Aravind Eye Hospital, Madurai, India: In Service for Sight,” Harvard Business School case (1993).
• Thomas Paine, Agrarian Justice (1797).

---

A note on how this piece was written: the subject, the argument from ecology to monopoly, and the questions it asks about health care are mine. Much of the research, the examples, and the sentences were drafted by an AI model working from that direction, and then edited by hand. I’d rather say that plainly than have a reader guess at it.