1
Why Markets Cannot Work in Healthcare
Professor Arrowâs Classic
Awarded the Nobel Prize in Economics in 1972 at the age of 51, the youngest winner ever, Kenneth Arrow was an almost legendary figure in economics. A socialist in his youth, Arrow remained a social democrat throughout his life, but some of his most important work has often been misconstrued by conservatives as defending unregulated markets. Nowhere has this curious misconstruction been so powerful, and so pernicious, as in relation to Arrowâs 1963 article, âUncertainty and the Welfare Economics of Medical Care,â which has since been cited nearly 10,000 times.1
Arrowâs article was meant to bring the analysis of information to the study of imperfect competition. Instead, it spawned a new field of health economics, where economists often cite Arrow before proceeding to discuss healthcare as if it is a commodity like cappuccinos or soft drinks. Few remember, if they ever understood, Arrowâs point: medical care cannot be treated like other commodities because of fundamental information asymmetries. Alas, once they leave an authorâs pen, ideas take on a life of their own. If economists have used Arrowâs words to make arguments contrary to his intent, all we can do is to try to recapture his meaning.
Arrow anticipated the failure of the market turn in healthcare. In well-functioning markets, competition among multiple suppliers drives costs and prices down to the cost of production because consumers seek out the highest quality at the lowest prices. This process does not work in healthcare due to information asymmetries that mean âconsumersâ (a.k.a. the sick and disabled) must rely on the judgment of others about what they need to buy, leading inevitably to concentrations of market power.2
The Problem of Healthcare Market Power
Healthcare will often be provided by institutions with market power. Providers operate in large facilities, like hospitals. Hospitals, specialist practices with dedicated machinery, pharmaceutical companies with large research facilities: all operate at a large scale that limits the number of competitors, giving each the power to influence prices.3
Even more important than scale, however, is the role of information. Few consumers can evaluate providers or treatment plans; even trained professionals do little better than amateurs in choosing providers.4 Every doctor is different, patients with identical diagnoses are different, each doctor-patient relationship is different. Feeling unwell, patients go to a doctor, not knowing what problems they may face, or even if they have a problem. They enter physician offices largely unaware of the proper treatment they require. Instead of comparing essentially equivalent bars of soap, where buyers can evaluate price and quality, healthcare âconsumersâ compare providers of unknown and unknowable quality.
Facing such uncertainty, people rely on quality signals, brand names, reputation, the experience of others, advertising, even prices. Buying services that literally can be the difference between life and death, higher prices can increase demand because they are signals of quality. Signals steer patients to a small subset of providers who thus have real market power because people believe they provide better care and save lives. Massachusetts General Hospital, New York Presbyterian, Johns Hopkins, Yale New Haven, the Mayo Clinic, the Cleveland Clinic: these institutions do not compete on a level playing field with other hospitals and providers, and they advertise to further enhance their brands.5 Their reputation gives them the leverage to raise prices.6 The importance of reputation encourages providers to merge with marquee providersâhospitals with a reputation for extraordinary quality.
The Problem of Health Insurance
Arrow also highlighted the way uncertainty complicates the problem of health insurance. We do not know what healthcare we will need or when, or how much it will cost. Most of us, most of the time, spend little on healthcare; half of Americans spent less than $1000 in 2018. Or we may hit the jackpot: 1% spent nearly $200,000 in that same year. The only things that we know with certainty are that we will need healthcare, that it will be very expensive, and that having access to care may make the difference between our living and dying.
It is generally true that people prefer the security of guaranteed small losses to the risk of very large ones. Since classical times, insurance has developed to provide that security by spreading the risk of very bad outcomes among larger groups. Facing the risk of fire, people buy homeowners insurance, paying a certain amount every year to receive money for a new home if needed. Facing mortality risk, people buy life insurance, accepting small losses in premiums for compensation in case of death. Facing the uncertainty of ill health and subsequent healthcare expense, people buy health insurance.
Arrow shows how asymmetric uncertainties create special problems for health insurance. Where no one knows the risks facing individuals, insurers can set premiums based on population averages, confident that the law of large numbers will generally protect them from extreme risk. But where subscribers have privileged information, insurers fear that anyone looking to buy insurance anticipates that they will need coverage. They fear âadverse selection,â where people buy insurance because they know that they are more likely to need coverage than the population average, and so will cost more to cover than the insurance company expects. At the extreme is âmoral hazard,â where people buy insurance precisely because they intend to use the policy.7
Moral hazard and adverse selection lead to destructive insurance company practices, including the creation of an enormous wasteful bureaucracy dedicated to screening subscribers. Companies invest in bureaucrats, underwriters, claims adjusters and investigators, all to police subscribers and to screen for moral hazard and adverse selection. To discourage moral hazard and limit their risk from adverse selection, they limit the value of insurance with provisions for cost-sharing, deductibles, and co-pays, consciously reducing the social welfare gains from insurance. Insurers have also learned that they can profit by screening for certain subscribers. In a process called âcherry picking,â they advertise for those unlikely to use insurance; by âlemon dropping,â they discourage those more likely to need insurance. They have created giant research operations and marketing departments to develop better screening, including multiple plans with alternative benefit structures to steer potential subscribers towards plans more profitable to the insurer, at the expense of providing adequate coverage to the sick and needy.
Social Insurance and Health Insurance
While all insurance suffers from problems of uncertainty, adverse selection, and moral hazard, these problems are much greater for health insurance.8 Life, homeowner or automobile insurers worry that people will destroy their own property, or even arrange their own deaths, to collect insurance. But these insurers usually find enough protection in a relatively small staff of insurance inspectors and, of course, the work of the police. In other cases, however, the risks of adverse selection and moral hazard are so severe that private insurance cannot function and insurance can only be provided by agencies with the legal authority to require universal coverage and to engage in more intrusive inspection to prevent fraud. For unemployment or income insurance, the risks of fraud are so great that private companies do not offer policies.9 Public bodies offer coverage in these cases, through social insurance programs like Unemployment Insurance, because they can mandate universal coverage, eliminating the danger of adverse selection, and have the state authority to monitor behavior to reduce the danger of moral hazard.
As a society, we want people to receive quality healthcare because it raises productivity when people are healthy and because, as a community, we want people to live long, healthy lives. Most countries, then, treat health insurance like income insurance, with public provision to capture the benefits of insurance while avoiding adverse selection with universal provision and moral hazard through state supervision.
The case for social health insurance goes further, beyond efficiency to questions of equity, humanity, and citizenship. Valuing all as equal citizens, we do not want healthcare to be reserved only for those with the most money. Or, even worse, to trust healthcare decisions to private companies whose profits depend on their ability to deny access to healthcare. But this is exactly the situation with private health insurance. Insurers pay close attention to their âmedical loss ratioâ (MLR)âthe share of insurance revenue paid out in health benefitsâbecause a low ratio means that more is available for profits. Every dollar paid out in benefits is a dollar lost from profits.
For managers of insurance companies, a high MLR means they are suffering from adverse selection and moral hazard, but a low MLR shows that they are doing their job of maximizing profits. I...