Medical Market Failures

Academic economists often publish analyses of the consequences of possible policies; in fact, I’d say it is the bread and butter of what economists do.   Such analyses usually rate policies using economist’s standard evaluation criteria: economic efficiency.  When a policy of more government intervention is ranked higher that a policy of less government intervention, that is usually because of an identified “market failure”, i.e., a reason why a low regulation “market” situation would not achieve high economic efficiency.

Some economists also act as pundits, arguing for or against current policy proposals to wider audiences.   Such arguments often draw on a wide variety of kinds of reasoning.  But since these people are economists, and not just pundits, I want to hold them to a higher standard: they should offer market failure arguments for their conclusions, and clearly distinguish these arguments from other reasons.  If a big part of the reason people listen to econo-pundits is that they are economists, having been certified by the economics community as having economic expertise, it would be nice to see signs that they continue to apply their economics expertise to their punditry topics, rather than just using their pulpit to preach policies they like for other reasons.

Econo-pundits have largely met this standard on global warming and carbon regulation; the potential market failure is so obvious: local actions by carbon consumers and producers should ignore global climate effects.  They also meet this standard at least somewhat on the financial crisis, discussing external benefits and costs of spending and regulation.  On medical reform, however, econo-pundits seem to have completely dropped the ball.

All I’ve found is a few posts quoting and responding to Krugman citing Arrow as showing:

The standard competitive market model just doesn’t work for health care: adverse selection and moral hazard are so central to the enterprise that nobody, nobody expects free-market principles to be enough.

But of course to meet the standard I suggest, it is far from sufficient to rattle off a market-failure buzz-phrase or two; one should offer arguments connecting the specific policies one favors or opposes to the market failures one sees as relevant. And so as not to constrain only pro-intervention econopundits, we should expect anti-intervention econopundits to at least say which plausible market failures come closest to justifying intervention.

Let me now try to meet my own standard, by outlining the main possible medical market failures, and the sorts of policies that such failures might justify.

  1. Insurance moral hazard – Insurance is less attractive because it makes people take less care.  Governments can’t help with this, though via limits on what insurance prices can condition on, they can make it worse.
  2. Insurance adverse selection – If those who privately know their risks are lower buy less insurance, too little insurance gets bought.  In this case forcing everyone to buy insurance some can improve welfare.  However, it turns out we don’t see this problem when insurers can price based on everything they know; usually low risk folks buy more insurance.  We do see adverse selection, however, when insurers hands are tied, such as by law or in optional group insurance.
  3. General altruism – If we just generally cared about the welfare of others, such as the poor, we might give them cash; spending the cash to give them medicine instead would be worse.
  4. Health altruism – If we cared more about others’ health than their happiness, we might tax things that hurt health, like smoking, and subsidize things that help health, like exercise, clean air, etc.  We’d have to do a lot of this before we even considered subsidizing medicine, since the correlation between health and medicine is very low.
  5. Emergency room altruism – If we just can’t help but provide basic medical services to folks in crisis who show up empty handed to an emergency room, we might reasonably require folks to buy catastrophic insurance to cover such situations.  Since it would cover far less, this insurance would be far cheaper than most medical insurance today.
  6. Medicine altruism – If we just like knowing that other people can get the medicine they want, regardless of whether it makes them happy or healthy, we might subsidize such medicine for them.
  7. Under-appreciation of  medicine bias – If in fact most people are irrationally biased to underestimate the value of medicine, it might make sense to subsidize medicine.  Of course in this scenario voters would be unlikely to endorse such a subsidy.  And if anything the bias seems to go the other way, which would more justify taxing medicine.
  8. Leaky info – If private producers under-provide info on the quality of medical procedures, professionals, or providers, because they have insufficient property rights to the info they create, governments might subsidize such info products.  This includes research.
  9. Large scale economics – if the efficient scale of some medical production is very large, we might consider allowing very large organizations but regulating their price.
  10. Contracting and agency costs – organizations give their agents, e.g., employees and contractors, rules and incentives to ensure they will behave as intended.  Contracts between customers and producers must try to consider a wide variety of possible circumstances.   These contracts, rules, and incentives require monitoring and enforcement costs.  Unless government organizations have incentives or contracts available to them unavailable to the private sector, they should have no advantages on these issues.
  11. I’ll add more items here as needed.

The item here that I take most seriously is leaky info, especially on new procedures, including drugs.  Something like medicine altruism is also a big factor, though I see it as signaling loyalty to locals, and so hurting foreigners.

GD Star Rating
Tagged as: ,
Trackback URL:
  • Carl Shulman


    I suspect that your spellchecker systematically misleads you: the word ‘analyzes’ is solely a verb, and not also the plural form of ‘analysis.’

  • JRW

    Could you expand on number 2, specifically why tying the hands of insurers leads to more adverse selection?

    • JGWeissman

      If an insurance provider is unable to differentiate between high risk and low risk customers, they will end up charging the low risk customers the same price as the high risk customer, which is more than they would charge if they could take the low risk into account. So the low risk customers will be less likely to buy insurance because it costs more.

    • Jess Riedel

      JGWeissman has it right, but here’s another way to think about it: adverse selection happens when the customer knows more about their health risks than the insurance company. If the insurance company knows and can use all the information in your medical record, it is unlikely that you know much more about your own health risks than the insurance company does. if, however, the insurance company cannot discriminate based on factors that obviously increase risk (e.g. having diabetes) then consumers will be significantly more informed than the insurance company effectively is.

  • Dagon

    economist’s standard evaluation criteria: economic efficiency
    Many economists and most laypeople do not believe this is the only or even primary criterion on which to prefer one policy to another. It could be argued that many uses of the term “market failure” is just a restatement of “market outcome that some people don’t like”.

    as pundits … they should offer market failure arguments
    Why? I’d agree that as economists, this is their strength and they should focus on it. As pundits, though, they are taking on a different role, for a different audience, and with different motivations.

    Do you hold other groups to such limitations in cross-profession endeavors? Are accountants not allowed to have opinions about mandatory vaccination? Are truck drivers not able to make arguments about air travel?

    As consumers of punditry, we need to recognize whether professional certification adds or does not add weight to an argument. That’s our problem, not theirs.

    policies they like for other reasons
    Pointing out flaws (or strengths) in these other reasons in specific is likely more effective than claiming that all non-efficiency arguments coming from an economist are invalid.

    • Jess Riedel

      I think Robin is saying that if your high-profile pundit status is predicated on your knowledge of economics, you should restrict your expert opinion to arguments which are based only on economics. (Or, at the least, you should label when you are using a contentious assumption about which you have no special expertise.)

      In other words, Paul Krugman should not use his aura of specialists to put forward a claim which depends crucially on his personal ethics. If he wishes to do so, he should prominently label that his expert opinion is dependent on the acceptance of a claim in an area where he is merely a layman.

      • Dagon

        if your high-profile pundit status is predicated on your knowledge of economics

        This seems an unlikely predication. There may be some facade of causality between economist status and pundit status, but it seems a pretty weak correlation to me.

      • Jess Riedel

        Dragon: Just because the best economists aren’t spending their time as economic pundits doesn’t mean that economic pundits don’t derive their status from their economic credentials.

  • Jonathan Graehl

    Isn’t that obvious? Suppose a new, incurable, chronic disease X infects half the population (the rest are immune). Suppose a law is enacted requiring that insurers must sell insurance at the same price to the X-infected and X-immune. So, the price for the immune is necessarily increased. As an X-immune, you’re more likely to avoid this comprehensive coverage policy in favor of a combination of self-insurance, and catastrophic coverage (e.g. an annual deductible such that 99% of the time, X-sufferers’ treatment falls under it).

  • It really isn’t market failure when there is no market. Markets require a level of knowledge, information, and transparency that simply does not exist when even prices are hidden.

    • Matt C

      Yup. Try to find out how much a medical procedure will cost you in advance of getting it.

      I did this recently. I called several hospitals and clinics to try to find out the cost of an endoscopy. I never got what I considered a straight answer. I got passed around from person to person and had to leave messages to get called back. They wanted me to tell them exactly which medical codes were going to be applied for the endoscopy. I also had to give them my insurance information for them to be able to tell me anything.

      In the end I got guesses ranging from $600 to $4000 and decided not to bother with the endoscopy.

      At least my dentist can still give a straight answer when I want to know how much something’s going to cost me. For now anwyay.

  • Timothy

    Some version of 7: Under-appreciation of medicine bias. Centre-left voter values medicine and has purchased medical coverage however thinks that others, primarily poor people, would under-value medicine.

    10. positive externalities of health: if I am sick I somehow impose costs on others – eg. financial costs on my employer, health and inconvenience costs on my friends and family – and I underappreciate these costs, causing me to underconsume medical insurance.

    • Your employer, friends, and family can easily influence your behavior; no need for wider government intervention to deal with those interactions.

  • Alan

    Re:5 – Emergency room altruism

    You say – ….. Since it [catastrophic insurance] would cover far less, this insurance would be far cheaper than most medical insurance today.

    But isn’t part of the current high cost exactly due to this problem e.g. ER-based last-ditch care is the most expensive provision method. Wouldn’t therefore the loss of preventative coverage result in very high premiums for such ER care, since it is likely to be more expensive than the ideal regular monitoring and ongoing care?

  • I think your promotion of merging health and happiness metrics is a wise one. I research the economics of happiness and these repeatedly prove to be highly correlated variables. I’m showing a bit of my results here at the following site:


  • 11. Insurance is a credence good – consumers have no way to figure out quality of insurance they’re getting until they got seriously sick and it’s too late to change it anyway. Markets for credence goods are almost always dysfunctional.

    12. Conflict of interest by insurance companies – related to 11, because they get profit by denying coverage on technicalities, and consumers have little ways of figuring out which insurer is safer.

    • Lots of goods are credence goods, with little government intervention.

      • That doesn’t mean the markets are less dysfunctional, just that fewer people get upset about the dysfunction.

        In fact, the example list of credence goods on Wikipedia are all things I hear complained about atypically often.

      • Chris

        I am guessing that the other goods don’t leave you dead or otherwise unable to leverage your experience of a bad product on the other side of the incident.

      • Soulless, credence goods are simply not market failures. It might be bad if it rains on your picnic, but that doesn’t make the rain a market failure. You might prefer a good not be a credence good, but that doesn’t make it a market failure either.

      • Chris

        That is to say, it might be bad if it kills me, but that doesn’t make my death a market failure.

      • Robin, I made no claim about that. Your logic was faulty. Noting the existence of regulation-free credence goods is not in any way a rebuttal to Tomasz’s point 11. That is, given “X is Y, Y has problem Z, ergo it should have intervention”, saying “A, B, and C are also Y, and don’t have intervention” is a non-sequitur.

        However, one of the distinctive features of a credence good is extreme information asymmetry, which is a pretty standard imperfection that reduces market efficiency. Whether or not there’s any useful way to correct it, that does actually seem like a pretty clear-cut case of a minor market failure, unless you’re using some definition of “market failure” that doesn’t boil down to “suboptimal economic efficiency”.

        Your rain example is very witty, but irrelevant.

  • Jeff

    Following up on Tomasz’s #12, I would say 12b: Health insurance is too technical a good to be comparable by most consumers. This would argue for greater regulation of the presentation of benefits.

    13. Cultural issues causing significant irrationality regarding insurance and happiness. Especially demonization of the uninsured in the media. This leads to several negative externalities: significant stress and unhappiness resulting from being denied coverage for, for an example that just happened to me, being too tall; fear of changing jobs, and fear of starting a business because of loss of health insurance; lying to health insurance companies to get coverage, leading to rescission.

    People are strongly attached to their own health as a component of their identity, and highly loss averse. Their behavior regarding insurance isn’t likely to follow the predictions of Homo Economicus for other types of insurance.

  • Pingback: Medical Market Failures | health()

  • Mike

    My motivations for supporting public health care include (3). I do not propose to simply give people money, because I suspect they do not have the discipline to reserve that money for health insurance, but would instead spend it on other needs and wants. I realize this is very patronizing, but the reality is some of us are smarter and better at long-term planning than others, and I suspect it’s largely the latter who lack health insurance.

    My other motivations don’t seem to fit into Robin’s categorization. For one, I want myself and others to be able to purchase insurance regardless of prior conditions or employment status. I am under the impression that the present system makes this difficult.

    I also question how well the market works. It seems to me, a large private insurance company does not have strong enough incentive to cover my medical costs when its obligations are ambiguous. Failing to cover leaves an unsatisfied customer, but I have little recourse, since my insurance is tied to my employment, and my employer offers few plans. I could voice my disappointment, but I worry that employers do not care strongly about customer satisfaction, because many employees do not distinguish job offers based on available insurance plans. Public insurance does not have strong incentive to make me happy, but it has little incentive to deny me as well.

    And I question the efficiency of private insurance. Compared to public insurance, private insurance has large administrative costs. Much of this, I guess, goes toward fighting claims. This may lead to greater “efficiency” within the company, if it cuts expenses more than the cost of administration, but to the extent that it denies medicine that people then pay out of pocket, it is not clear that efficiency is added to the entire system.

    • Mike

      I see Timothy finds my first comment falls under some modification of (7) more closely than (3). My other comments seem in part to fall under Tomasz’s (11) and (12).

    • If small firms have larger admin costs, that is a scale economy, which I listed. Are you sure public plans have little incentive to deny you? Any insurance with limitations will have to have to spend monitoring and enforcement costs to implement those limitations. This is just as true of government provided insurance as of private.

      • Mike

        Both seem to have a complicated set of incentives, but at the core the corporate mission is not expressly to serve the public good, while this is at the core of the govt mission.

        Rather than argue in abstract, it would be nice to see if someone has studied trends in claim denials comparing Medicare to private insurers.

      • Mike, there can be non-government non-profit insurance providers, who can whatever mission they want, including ‘serving the public good.’

  • Ben

    As a young, fit, healthy rationalist, I already assume that preventative medicine will result in better healthy than visiting the doctor.

    While we’ve spent a lot of time analyzing the flaws and lack of correlation between medical care and health, is there a similar study identifying the most effective forms of prevention?

    From what I can tell from death and illness statistics, by not smoking, maintaining a healthy weight, not drinking, not owning a firearm, maintaining a good driving record (debatable), using protection during sex, and not using sophisticated drugs, I prevent the causes of death responsible for 46% of the mortality rate in the USA.

    Is there a corresponding list of preventions vs. medical cost? For example, given the high cost of dental work, I imagine flossing saves thousands of dollars a year, but I have no first order evidence.

    • q

      > I prevent the causes of death responsible for 46% of the mortality rate in the USA.

      it does not follow that you will be saving any money.

      if you avoid the illnesses that kill you, you may end up living a very long time with very expensive lower intensity illnesses — COPD for example, or organ impairment requiring dialysis, etc.

  • Unnamed

    Interestingly, your criticism of econo-pundits engaging in punditry without using their econo-expertise is similar to one that Matt Yglesias made, spawning this post of yours.

    On Krugman’s claims about adverse selection and moral hazard, Brad DeLong adds a very rough sketch of how the argument is supposed to go. For a more detailed argument from Krugman about what’s wrong with the health care system and what to do about it, take a look at this article with Robin Wells. That’s what a minute of googling turned up for me.

  • I think there is also a correlation across countries between life expectancy and per captia medical expenses. I suggest if we all die a bit younger we will all save plenty on medical care. Maybe that’s another one for the pro-euthanasia debate.

    The main problem with health care and costs of health care is that we all die. It doesn;t matter how much we spend on each person. And suggest to previous commentor Ben, that there is a flaw with correlating healthy activies and causes of death. You could get enough data in the regression to show that you are so healthy you may never die!

    Anyway, as an economist, I actually believe the structure of health care is not about maximising social welfare by some objective measure, but having a system that the people believe is beneficial to society, even if objectively, superior systems exists.

    Well, that’s about all from me.

    • No, once we control for other standard factors, we see no correlation between med spending and health.

      • Alan

        I have probably missed a lot of the subtleties of this argument, but isn’t this what you would expect? Health spending increases with illness and morbidity, healthy people do not incur large costs. Much of healthcare spending in developed nations is incurred for very marginal improvements, the easy stuff (sanitation, vaccination, childbirth) has been largely covered, what is left is the difficult, high-cost, low benefit areas. This is exacerbated by “lifestyle” costs such as obesity, diabetes etc, which vastly increase costs for little or no gain in outcome over “healthy” people without such conditions. In developing nations I would expect health spending to correlate strongly with health outcomes, in “first-world” countries I think we are well towards the asymptotic part of the curve.

  • lxm

    Aren’t these market failures?

    1. Fragmented markets allow insurers an advantage. In many states there are only a few health insurers which allows the insurers to raise rates with little regard to competition.

    2. Not an even playing field. If you are an individual buying insurance outside of a group, you are pretty much screwed: you pay more for less. Individuals have no bargaining power with health insurers.

    3. Bad incentives. Health insurers have an interest in selling you insurance, not making or keeping you healthy. Therefore, they do their best to deny your claims. Yet, people are forced to deal with health insurance because they cannot afford the catastrophic cost of care for a serious illness.

    If these are, in fact, market failures, they are discussed all over the place.

    This may or may not be a market failure:

    4. Inability to explicitly limit care when that care is of minimal utility, particularly in end of life care. Should you spend millions of dollars to keep me alive for a few more weeks?

  • I just added a point 11 to the post, which addresses issues raised by many comments.

  • gnat

    Do we want to add product or service transparency? Medical choices are extremely complex. What I want to buy is an effective healthcare outcome weighing costs and benefits. What I pay for is doctor hours generally for each specialist (each with its own overhead). The market tends to be stove pipe by type of specialist. To the extent they have market power doesn’t this create double marginalization.

    Quality rankings: Virginia has begun publishing hospital quality measures (e.g., infections from IV) which certainly should have an impact on consumer choice as it has on on restaurant inspections.

    Medical incomes are extraordinarily high even with insurance company cost controls. Moreover, incomes seem to be positively correlated with the number of medical specialists in a metro area. I understand that a new class: “doctor’s assistants” are being used to serve medicare-medicaid patients in some offices. The expansion of these types of providers may lower the overall cost particularly for routine care. This may allow medical offices to compete on price and quality.

  • Pingback: Overcoming Bias : Government Insurance Advantage?()

  • Doug S.

    You know what one of the biggest market failures in the U.S. health care market is?

    Individuals don’t get to choose their insurer. Their employer chooses their insurer for them. And employers have little incentive to choose the insurer their employees would prefer.

    Agent failure is a well-known type of market failure, isn’t it?

  • Guy Thomas

    2. Insurance adverse selection – If those who privately know their risks are lower buy less insurance, too little insurance gets bought.

    This is standard economics, but simple arithmetical argument shows it may often be wrong. The objective function of a utilitarian public policymaker should (arguably) be the risk-weighted quantity of insurance bought. Given this objective, some adverse selection may actually increase coverage (when coverage is correctly measured, ex-post not ex-ante). Some “adverse” selection may not be “adverse” at all. Google “loss coverage as a public policy objective”, or look at these papers

  • Pingback: It’s Not Over « The American Catholic()

  • Pingback: Southern Appeal » It’s Not Over()

  • Pingback: Is US gun control an important issue? | 80,000 Hours()