Monday, August 30, 2010

The Gospel According to Ken Arrow

Cross-posted from Critical Condition on National Review Online.

Austin Frakt helpfully points us to several posts by Princeton economist Uwe Reinhardt on the topic of Ken Arrow and his influential paper on health care economics, “Uncertainty and the Welfare Economics of Medical Care,” published in 1963 in the American Economic Review. Arrow’s paper, which endorses the view that “the laissez-faire solution for medicine is intolerable,” is widely considered to have founded the field of health care economics, and continues to be celebrated, especially by those to the left of center. If one were to assemble a New Testament of liberal health care policy, Arrow’s paper would be the Gospel According to Matthew. But what is truly remarkable about the gospel according to Kenneth, from today’s perspective, is its antique unsophistication about modern financial concepts and the modern regulatory state.

Arrow’s paper argues that the delivery of health care deviates in fundamental ways from a normal free market, and, therefore, that government intervention is necessary to correct for these deviations. It therefore can’t be considered surprising that modern liberals, who believe that the free market leads to unjust and suboptimal outcomes, celebrate the work of an economist who argues that a free market for health care…leads to unjust and suboptimal outcomes.

All of us, right, left, and center, are tempted to gaze uncritically upon those works by others that reinforce our own worldview, while ignoring contrary perspectives. No one exhibits that tendency more than Paul Krugman, Reinhardt’s fellow Princetonian, who once wrote that “ever since Kenneth Arrow's seminal paper, [economists have known] that the standard competitive market model just doesn't work for health act all wide-eyed and innocent about these problems at this late date is either remarkably ignorant or simply disingenuous.” (Then again, pretty much everything Krugman knows proves that those who disagree with him are ignorant or disingenuous.)

So let’s examine Arrow’s catalogue of health care’s market distortions, by category (for those who have read my National Affairs piece on the subject, this will be old hat):
  1. Unpredictability. Arrow points out that people’s needs for health care are unpredictable, unlike other basic expenses like food and clothing. But while we can skip the occasional meal or sale at Old Navy, our need for health care can be far more urgently necessary.
  2. Barriers to entry. Arrow notes that you can’t just set up shop on the side of a road and practice medicine: you must have a license to be a physician, and gaining that license requires years of expensive schooling and training. As a result of this constraint on the supply of physicians, there is a constraint on the supply of medical services.
  3. The importance of trust. Trust is a key component of the doctor-patient relationship; if a surgeon makes a serious mistake during an operation, for example, the patient may die or become permanently disabled. The patient must trust that the surgeon knows what he’s doing, and can’t test-drive the surgery beforehand.
  4. Asymmetrical information. Doctors usually know far more about medicine than do their patients. Therefore, the consumer of medical services (the patient) is at a serious disadvantage relative to the seller (the doctor). Patients are therefore vulnerable to exploitation. In addition, third-party payors of medical bills, such as insurers or the government, are that much more removed from the particulars of a given case, and unable to effectively supervise medical practice.
  5. Idiosyncrasies of payment. Unusually, patients pay for health care after, not before, it is received (that is, if they pay for health care at all). Because patients don’t see the bill until after the non-refundable service has been consumed, and because patients are given little information about price and cost, patients and payors are rarely able to shop around for a medical service based on price and value. Compounding this problem is the fact that patients rarely pay for their care directly.
Arrow isn’t wrong to point out these distortions. Where he is wrong is in believing that they are unusual, and that government intrusion is required to correct them. Indeed, the phenomena that Arrow identifies are widespread throughout the economy, and are made worse by government policy. Let’s take them one by one:

Unpredictably is hardly a remarkable phenomenon. In the half-century since Arrow’s paper came out, a large industry has emerged to address it. Extended warranties, overdraft protection, traveler’s insurance, etc. etc. are all products designed to smooth out the unpredictability of life. Most of these products have managed to function well with minimal government intervention.

Barriers to entry remain a problem in medicine, but again, this is a problem that is hardly unfamiliar to anyone who has ever tried to start a car company or an airline. We require licensure of lawyers, but so far as I know, no one is fretting about market distortions in the legal industry. Indeed, our economy is suffused with, and possibly suffocating from, the over-credentialization of everything.

It is true that trust is important in medicine. But so, too, is it important in many other commercial transactions. We trust that airline pilots, and airline mechanics, are competent. We trust that the manufacturers of car brakes have thoroughly vetted their products. Markets are capable of assigning value to such things, as are lawsuits.

There certainly are asymmetries of information in medicine, but again, is this unique to medicine? How many people, when taking their cars to the shop, are able to argue with the mechanic about the need to repair the transmission? How many of us rely on a financial advisor, because we feel incompetent to make investment decisions entirely on our own?

Indeed, nearly every economic transaction involves some asymmetry between the buyer and the seller. The ancient Latin aphorism, “caveat emptor” (let the buyer beware), has been enshrined in property law for centuries as a way of reminding people that sellers know more about the defects of what they are selling than do buyers. Nonetheless, we hold buyers responsible for understanding what they are buying, so long as the sellers haven’t engaged in fraudulent misrepresentation.

The internet has done much to democratize medical information. I speak to physicians all the time who say that their patients know more about new drugs than they do. A parent whose child has Lyme disease is likely to know as much, if not more, about available treatments than is your garden-variety family practitioner. That isn’t to say that going to medical school doesn’t matter—but it is to say that there are ways in which the gaps in information are narrowing. Electronic medical records, and on-line postings of medical prices and costs, will do even more.

Finally, we get to the most important of Arrow’s concerns: that the way we pay for medical services, especially via third-party health insurance, cannot function as an efficient market. On this point he would get wide agreement. But Arrow’s favored solution—government-sponsored health insurance—makes this problem worse by further removing patients from the price and value of the care they receive.

The real solution is to increase the degree to which patients pay directly for non-catastrophic health care, so that health insurance can function as insurance does in other fields: as a way to protect people from catastrophic financial and medical loss. A related, but equally important, reform would be to increase the degree to which patients buy insurance for themselves, as opposed to having it purchased on their behalf by employers or the government. A true individual market for health insurance would help rectify many of the distortions in our current system.

Ken Arrow and his disciples in government have done much to exacerbate, rather than correct for, the inefficiencies in our health care system. Obamacare is a prime example of this. One can only hope that we eventually learn from Arrow’s mistakes, and finally move in the right direction.


  1. Hmm. Long comment did not come through. So, shorter this time.

    Barriers. The barrier is not the license, it is the time needed to train. Long lag times.

    Asymmetry. In almost every other major purchase, you can test or use what you will buy. You can walk around a house, sometimes even rent it before buying. Test drive a car. Test a cell phone. Not so with your CABG. Even more importantly, people do not understand the statistics behind the explanations for treatments. All of this is influenced by the emotional state of the patient. These big decisions are usually made after getting some bad news.. There are time and geographic restraints. Your cancer care might be cheaper 70 miles away, but who will drive you home while you puke?

    On your last points we have some agreement. I dearly wish we had gotten non-employee linked insurance. I am unsure if this will lower costs, but it is worth a try. I would also like more exposure to costs, something like they do in France where all costs must be displayed.

    Query- My corporation has gradually been going to higher deductibles, co-pays and requiring employee contributions as we consistently look at double digit increases. It seems to me that we, like much of the rest of the country, are heading towards high deductible plans that are often touted as a means to cut costs. Is there any evidence that this is slowing spending?


  2. Hi Steve,

    Test-driving is an issue, but one that can be addressed by having publicly available ways of measuring quality of outcomes for individual physicians and institutions.

    As to your query: Yes, there is evidence that the growth in spending in high-deductible plans is lower than it is for conventional plans. Here's one that I mention in my most recent post:

  3. Thx for the link. Looks like it needs to be followed a bit more. Has anyone studied provider response, in the economic sense? What will stop providers from cost shifting to their higher cost procedures? If we know we are going to make less money on our carpal tunnels as a low fee procedure, why not make up for it on total joints or CABGs? Are cost savings being achieved, I couldnt tell from the article, by lower utilization, lowering reimbursements for individual procedures or both?

  4. "Is there any evidence that this is slowing spending?"

    Yes. Still can be dismissed as "anecdotal" but evidence is accumulating. No, I don't have any links for you.

    But that type of reduction in "spending" is not sufficient. It is related to reductions in utilization, not to the cost of delivering medical care in the first place.

    The meat of the matter is - why are a doctor's office costs so high? Why are hospital costs so high? Why does it cost so blinkin much to deliver medical care in the first place?

    When those questions are answered, we can then address this one: Is there something driving those high costs that people are willing to do without?

    After more than 40 years stumbling around in the medical-cost wilderness, we are still nowhere near the promised land, at least in part because our so-called leaders have not addressed these fundamental questions in a meaningful way. OK, medical care is a tougher problem than finding one's way out of a desert. But still . . .

    In a dream last week, God appeared to me and I asked him "will we find the solution to the medical cost problem?" And God said "Certainly . . . but, not in my lifetime."

  5. Hi Steve,

    On the issue of utilization: the AAA study shows that utilization is higher in some categories and lower in others (as you would expect). For example, utilization of prescription drugs was higher, but within that bucket, the utilization of generic drugs was higher and of branded drugs lower. As to pricing changes from providers, I don't think that has been studied, but it is unlikely to have happened so early in the life cycle of CDHPs -- CDHPs aren't enough of the market at this time to drive such changes. The changes aren't driven by lower reimbursements.

  6. Are employer sponsored health insurance plans demanded by law? That is, once a company reaches a certain size in terms of labor pool and/or monetary resources, is it obligated by the federal government to provide health care to its employees? It makes a great deal more economic sense to allow individuals to shop for and choose their own insurance plans, but if employer sponsored insurance plans are a product of the free market already, do individually chosen plans really stand a chance?

  7. Hi DN, welcome to the blog. Yes, employer-sponsored health plans are demanded by law.

    Our current system is an outgrowth of WWII-era wage controls, in which employers were constrained in what they could pay as salary, but not what they could pay in benefits (such as health insurance). This system was later enshrined in the tax code: while you pay income and payroll taxes on your salary, you don't on your employer-provided health insurance. If you buy health insurance on your own, however, you do so with after-tax dollars. This creates the single greatest inefficiency in our health care system.

    In the 2008 campaign, McCain campaigned on eliminating this tax subsidy, but Obama objected, arguing that it would be a tax increase.

    Obamacare takes things a step further: if you are an employer and you employ more than 50 people, you are *required* to provide health insurance for all of your employees, or you will be fined a huge sum (such that you basically have no choice but to provide health insurance). Naturally, this will serve as a hindrance to employment because it increases the cost of labor. One of the many truly idiotic aspects of our new health care law.