Explaining fragmentation

First, let me add my voice to the chorus of “thank you’s” to Frank and Glenn for organizing this forum.

Second, I must admit that, even though I specialize in health law, I am overwhelmed by the complexity of the discussion on fragmentation (let alone the fragmentation itself). Just in this forum there has been discussion of fragmentation of financing and then fragmentation of delivery. The proposed solutions to these (variously: single payer, accountable care orgs, and medical homes) seem also quite complicated, often requiring their own set of complicated regulations and controls (such as certification). On top of this there are concerns about irrational consumers (which raises the question of whether providers or regulators are any more rational) and public health issues (e.g., externalities, fiscal and epidemiological). What I really want – and I suspect others do too – is a simple way to think about the problem and thus about solutions. The problem is that I haven’t seen one. And I suspect others do not have one. Moreover, people who do advertise simple solutions (single payer, free markets) don’t have a way to transition to those solutions from where we are now. And so it is not surprising to me that our health care system is fragmented.

But I do not want to end on such a dour note, so let me offer some discrete and hopefully constructive thoughts on the problem.

1) There is a hilarious (and scary) youtube video, available at http://www.youtube.com/watch?v=5J67xJKpB6c, that imagines what air travel would be like if airlines worked like health care. I think it is meant to make us cringe at how dysfunctional health care is. But instead it got me wondering, how did airlines make air travel so easy? Or how did automakers and auto insurance companies make owning a car so easy? My first instinct is to say, free competition. But airlines for a long time were regulated. And so are car makers and insurance companies. But regulation cannot be the answer either. Airlines deregulated in the late 1970s, early 1980s. Automakers were never seriously regulated (except in the way that drug company products are regulated for safety). I think the lesson is that competition is not incompatible with integration and relatively efficient delivery of complex goods, but regulation may be required to address some safety problems. Interestingly, note that beyond regulation of how much you have to be paid when you are bumped from a flight or the (lax) regulation of auto insurance rates by state insurance commissioners, there is not much regulation of internal pricing, as there is in health care insurance and delivery.

2) I suspect that health insurance coverage begets regulation, and thus complexity. There are two reasons. First, moral hazard on the part of patients and on the part of docs. This moral hazard creates demand for regulation of patients and providers. This regulation contributes to complexity. The question is why this complexity is greater in the health insurance context than the auto insurance context. If the answer is that health care is a more complicated product to monitor, we’re unlikely to have much simplification of the health care system. (We can hide it within organizations through integration, but it will still be there, in the form of internal firm rules.) Second, insurance solves the problem of health consumption risk by putting people in a common risk pool. But this pool creates externalities across insureds. This is solved either by low risk types exiting, which is the familiar adverse selection problem that unravels insurance. Or – and this is important – if we solve adverse selection by things like insurance mandates, high risk types impose financial costs on low risk types. This is either because high risk types are inherently risky (think genetic ailments) or because they engage in risky behaviors (think smoking). If it’s genetic, the problem with insurance is that it is redistributive. And that redistribution is not compensated. If it is behavior, there will be demand for regulation of risk behavior. All of this adds complexity. (Note: the usual retort is the people do not know their risk level. But that is plainly false. A lot of us have private information that we’re more or less at risk than others. I don’t know what fraction of variance in risk is private information, but enough that there is adverse selection and demand for regulation of others’ risk behaviors.)

My point is not to say health insurance is bad. It is to say that health insurance may require complexity. And when there are multiple actors involved, e.g., private insurer and government, or single payer with different branches of government, complexity yields fragmentation.

3) I think the ACA gets an A for coverage and C or D for fragmentation. I tried to teach the ACA last spring and found it incredibly complicated and open ended. E.g., there are tons of experiments but no clear mechanism for picking the ones that will work and making sure they are uniformly implemented. E.g., there are often fixes for problems that the ACA itself creates, such as subsidies for medical education to offset cuts in physician reimbursement. E.g., the bill tackles all sorts of problems without any sense of coordination among them. There are provisions for expanding preventative care, for improving nursing homes, for fixing health insurance markets, etc. But these all affect each other.

If the ACA is that complicated to teach, I strongly suspect it will worsen fragmentation rather than reduce it. I wonder if life would have been easier if the ACA simply expanded Medicaid and gave some tax breaks and left the rest of the system alone. Certainly it would have made for a much smaller and less complicated piece of legislation. Worse, if you support single payer, I am not sure the bill makes it easier to get there. It creates a whole new set of vested interests that would prefer the new status quo to a reform that simplifies the world.

4) I wonder if it would be too crazy to suggest that we should have used the states to perform a federally controlled experiment with different models of health care financing and delivery. E.g., the US government could set up 3 models. In one, which we can call the health insurance exchange and coverage subsidies model, we have a Clinton/Obama type model of state exchanges + subsidies for poorer folks to purchase insurance. In another, which we call the McCain model, we allow free markets across states (or at least states that follow the McCain model) in insurance + give some subsidies to the poor. And in the third, we have a single payer along the lines of Canada. Let states choose one of the three models (or perhaps allow them to stick with the status quo). (The feds would pay for expanded coverage and all the plans except the status quo address the main uninsured problem.) Then let the experiment run for 10-15 years. Whichever works best, we extend to the rest of the country. If there is no obvious winner, then let states continue to choose the best option for them.

I know this proposal is unrealistic. But once we condition on solving the coverage problem, I am honestly not sure which system yields more efficient production of health. I have read enough empirical work in health care to strongly suspect that no one else does either. In fact, I would not be surprised if there are multiple equally reasonable solutions to the problem of efficient delivery of health. Because the solutions can be judged along a large number of dimensions, no one solution may strictly dominate the others.

You may also like...

6 Responses

  1. Frank Pasquale says:

    Thanks for that very interesting set of proposals. A few thoughts:

    1) On the air travel/health care analogy, Altman et al. had this interesting article in HA a few years ago:


    “By 2025 the need for general hospitals to cross-subsidize will greatly increase, but their ability to do so will be diminished. U.S. hospitals could begin to resemble U.S. airlines: severely cutting costs, eliminating services, and suffering financial instability.”

    I am pretty surprised that the makers of the video would want to say that airlines should be a model for any industry, let alone health care. Air travel has become one of the most stressful, hated experiences out there. And for someone as tall as me, it’s just not an option in may cases: I can’t fit into a standard coach seat without a lot of discomfort, and I can’t afford first class.

    2) As for the controlled experiment: I really like the idea, but I worry about a “race to the bottom.”

  2. Anup Malani says:

    1) I disagree that airlines should not be a model. They are a fantastic model. Very low death rate (safety). You can reach a ton of places (heterogeneity). Sure, not all people will be happy. But no product does that. And if you don’t like airlines, try cars/car insurance. Fantastic options. Really high and improving quality. Also important is the fact that there is much less waste in those sectors compared to health care.

    2) If race to the bottom leads to bad insurance, we’ll see worse insurance in the free market states. THat’s the point: see what works and what doesn’t. But I am skeptical. I don’t think there is great evidence for race to the top or to the bottom in corp law and we’ve had free markets for corp law for decades. With insurance there is additional risk of adverse selection, but we also have regulatory solutions to address that. Moreover, here may be sufficient risk aversion (or tax benefits) to get even low risk types to insure.

  3. Frank says:

    1) Given Joel Waldfogel’s work in The Tyranny of the Market, I think you’d have to acknowledge that a), at least for someone as tall as I am, the airlines are not a fantastic model, and b), this is a common problem in many markets. You say it’s inevitable that “not all people will be happy,” but that’s a much more significant problem in health care than in, say, soft drinks, leisure trips, or sneakers.

    Any work that tries to apply economic models developed in industries outside the health sector, directly to the health sector, risks missing the unique aspects of the experience of sickness and lack of health. (Ani Satz has suggested these in her contribution to the symposium.) That’s one reason I recommend Geoffrey Hodgson’s work “Towards an alternative economics of health care” to economists who do not focus primarily on health care. As Hodgson states,

    “Leading mainstream health economists suggest that health care has special features that make it different from other domains of application, posing restrictions on the appropriateness of some neoclassical assumptions. . . . [T]he literature points to the presence in health care of externalities, information asymmetries, uncertainty, supplier-induced demand, and derived demand. But while all these features are important, they are neither universal in health care systems nor unique to them.”

    “I try to identify the peculiarities of health care systems by building on the neglected but vital concept of need. By contrast, mainstream economics starts from the subjective satisfaction or utility of the individual. Modern mainstream economics rejects or ignores the concept of need, but many leading economists from Adam Smith to Alfred Marshall have acknowledged objective needs as well as subjective satisfactions.”

    2) “Worse insurance” is “what works” for many large corporate interests.

  4. Anup Malani says:

    1) a) Hetergeneity is a problem when there are scale economies (so centralized producers) and lots of hetergeneity. But less so when there is decentralized production (docs and hospitals). Perhaps health care demands are much more heterogenous, but the problems that most people point to are not about hetergeneity, but mainstream treatments people do not get.

    b) For every Waldfogel book, there are a 10, perhaps 100, articles pointing out how how well markets work or unintended consequences of regulations. Citing Waldfogel or Friedman is not the point. The point is, does more or less competition work in health care and what is the apporpriate level and form of regs. I readily admit that more regs may be the right approach, but I am fairly confident that neither you or I know for sure. I would like to experiment.

    c) This is the most important: even if health markets have flaws, the question is whether governments do better. There are lots of places where they do much worse: post office, defense procurement, prisons, infrastructure maintenance, managing pensions, etc. It is not sufficient to say health care markets have problems, you must show the govt does it better. For some areas, clearly the govt can help, like addressing adverse selection or redistributing wealth so there is an even distribution of health. I am also optimistic about insurance exchanges. But beyond that things get iffy. For example, do we know if government rules against referral fees are better than vertical integration?

    2) Again, that is a closer to cliche than an argument. There is evidence that corps care about the fringe benefits they offer workers: it’s part of the wage competition. Suboptimal insurance (for the worker) yields a lower wage. (Note: I didn’t say less insurance, I said suboptimal.)

  5. Richard Saver says:

    Anup/Frank: I will sidestep the airlines vs. health care debate, but you may be interested to see that others share Frank’s pain


    To return to Anup’s general comments, I think he aptly points to the problem of how to “transition to … solutions from where we are now.” And I further agree that the Affordable Care Act gets higher marks for coverage and lower marks for fragmentation. There was a degree of path dependence to start with. Plus, a political strategic choice by Democrat leaders to sell health care reform as not really disturbing the status quo for those who already have coverage and are happy with it. While the politics are understandable, this has resulted in a statute that, as Anup observes, has provisions that are not necessarily coordinated and that leaves an awful lot open ended. Because of the statute’s complexity, I am dreading having to teach it for the first time next semester (Anup and others: any teaching tips?) If it is so difficult to even teach, little wonder it has become hard to sell to digruntled voters, even those who stand to benefit from it.

  6. Frank says:

    Thanks for the CNN article, Richard!

    As for the “transition to … solutions from where we are now,” I think we face a pretty stark choice: we can either double down on reinforcing private insurance and private profit motives in a system that is already an outlier in the developed world in terms of its reliance on these kinds of institutions and incentives. Or we can learn more from Europe, Taiwan, etc., as Kieke Okma, Tim Jost, T.R. Reid, Callahan & Wasunna, and others have suggested.

    For me, comparative health policy demonstrates the US system to be wanting, particularly in terms of cost-control, but also in many other dimensions. (Not surprisingly, there are similar deficiencies in broadband internet access in the US compared to many other countries with a more dirigiste approach.)

    I have little doubt that we’ll continue to hear that the market has yet to work its magic; that if only we free it more completely, it can bring us dramatic improvements in cost-control, access, and quality. As long as we have a mixed system, we’ll have endless ideological debates attributing bad outcomes to either market or regulatory forces.

    On a final conciliatory note: I do think that further experiments and pilot programs are important, and I applaud the ACA for setting up many of these. Anup, your point on the referral fee laws is well-taken, and it will be interesting to see how the HHS “Secretary may exercise waiver authority or create new exceptions and safe-harbors related to the physician self-referral law” regarding ACO’s, as this workshop discussed: