« China fact of the day -- life insurance for bird flu | Main | Review of File-Sharing Papers »
Why doesn't the health care market work so well?
I see a few candidate hypotheses:
1. Adverse selection in insurance markets. In this view, outside of the corporate employment context, mostly the unhealthy -- the "lemons" -- show up to buy health insurance.
2. Poor information about the cost and benefit of different medical procedures and providers. I am puzzled why we don't have better institutions for evaluating providers and spreading this information to consumers. Part of the problem is legal, part of the problem is measurement (doctors could dump near-hopeless patients to get better ratings), and part of the problem is that many people don't want to know how good (read: bad) their doctor is.
3. Time consistency. Ex ante, we are most worried about catastrophic health risk. Ex post, most of the overinvestment in health care comes for victims of catastrophic health risk. No set of institutions can square these dual perspectives satisfactorily.
4. Showing that you care. This is Robin Hanson's hypothesis. You spend money senselessly on health care, mostly to show your wife you love her and the kids.
Believers in national health insurance tend to emphasize number one, which might be alleviated by forced mass pooling. In contrast, I am skeptical that adverse selection is the significant problem.
It is less clear that national health insurance could improve performance on number two. People would remain underinformed. Government might have less incentive to rip them off, but the implication would be that government provision is "lazier" in general. Not a comforting thought.
National health insurance does address numbers three and four, albeit in backhanded fashion. Ex ante, people feel protected and the program is popular. Ex post, such systems spend less money on the last six months of life than does the U.S. system. The relevant denial of treatment is often invisible rather than a stern hand pulling the plug.
The bottom line: Defenders of NHI place great stock on #1. If #1 were significant, we could, at least in principle, use national health care to both lower costs and improve treatment.
If #3 and #4 are the major problems, national health insurance provides benefits by restricting overinvestments in health care. This is consistent with Europeans living longer and spending less on health care. The U.S. maybe could replicate these benefits if a) we push out private insurance companies, b) we ration or abandon expensive procedures which don't extend lives very much, and c) we adopt healthier lifestyles.
I am skeptical of #1, and I am not ready to bet on a), b), or c), much less all three at once.
Posted by Tyler Cowen on November 22, 2005 at 06:48 AM in Medicine | Permalink
TrackBack
TrackBack URL for this entry:
http://www.typepad.com/t/trackback/3576/3695254
Listed below are links to weblogs that reference Why doesn't the health care market work so well?:
» Marginal Revolution: Why doesnt the health care market work so well? from Independent Sources
Econoblogger Marginal Revolution gives a nice rundown on why the health care market doesnt perform so well. One would expect that given an efficient market the healthcare ship would ultimately right itselfbut it hasnt. Why is th... [Read More]
Tracked on Nov 28, 2005 12:07:49 PM
» Talking to Liberals.... from The Lone Elm
Arnold Kling, a MIT PhD economist, has a fine article out (the first of a series) aimed at how liberals might communicate better. Part of the message is that they might think through the implications of their policy ideas. Kling [Read More]
Tracked on Jan 17, 2006 8:38:14 AM
Comments
The problem with (b) is that some fraction of our expensive procedures are steps towards better research and cheaper procedures for other, non-terminal people. Terminal cancer patients enrolled in cancer trials are one example. A larger effect may be in the value of repeated surgeries for the surgeon -- there is a lot of evidence that surgical outcomes are better when you get a surgeon/hospital who perform the same procedure very frequently.
Posted by: DK at Nov 22, 2005 8:19:16 AM
Adverse selection is a non-problem in a free market.
As The Economist pointed out recently, the insurance industry
is heavily regulated, so much so, I would argue, that it's absurd to
call it a market. State regulations prevent health insurers
from screening and pricing applicants according to the risk they
represent.
I am a healthy male, resident in NY, have never made a claim,
and my carrier just informed me that it's dropping my high deductible
plan ($5,000) effective next February. The company is offering me a
plan that pays from almost the first dollar of claims (it has a very
low co-pay) and jacking up my premium by 91%.
I asked the actuaries there if I've they'd write me a plan with a higher
deductible ($20-k, $50-k, or higher) in exchange for a lower premium.
(I bet Lloyd's would.) They got back yesterday saying they are strongly considering it, and probably will, but there are state-imposed regulatory hurdles they have to overcome and other issues.
Another advantage of a free market in health insurance is that Krugman
could then launch his uninformed attacks on other targets and stop pretending
that he knows something about insurance.
Posted by: Bill Stepp at Nov 22, 2005 8:45:02 AM
Ah, but once NHI was in place, then people buying add-on insurance to pay for expensive end-of-life care would no longer really count, would they? I suspect that those expenditures would then be put in the 'optional' rather than 'required' category (along with orthodontics, breast augmentation, etc).
Posted by: Slocum at Nov 22, 2005 8:52:17 AM
I'm skeptical that NHI would help much, since most health expenditures are already on the elderly, who are covered by Medicare already. There's no evidence that Medicare spending is more efficient than private spending, or matches the cheaper results of foreign NHI systems.
Posted by: John Thacker at Nov 22, 2005 9:15:06 AM
Unless b) and, to a lesser extent, c) come to pass, any health care "reform" is just arranging deck chairs on the Titanic.
One irony is that while us Americans are much more religious than Europeans, we're also far more frightened of death, therefore far more insistent on costly end-of-life procedures. It should be the other way around. We should be more willing to accept a peaceful, natural death because we believe that something better awaits us on the other side (unlike those heathen Euros, who believe that death brings oblivion).
Posted by: Peter at Nov 22, 2005 9:35:20 AM
I'm skeptical that NHI would help much, since most health expenditures are already on the elderly, who are covered by Medicare already.
If that's the case, then why the escalation of employee insurance costs for businesses (who don't employ or provide insurance for the elderly)?
Posted by: Slocum at Nov 22, 2005 9:49:53 AM
I'm with Bill Step here, with the addendum that while markets outperform government reliably, there is no guarantee that government/corporate hybrids will.
Posted by: michael e vassar at Nov 22, 2005 9:59:44 AM
I wish I could buy health insurance that paid for moderate-costing things, but excluded really expensive things that don't prolong/improve life much.
Posted by: Paul N at Nov 22, 2005 10:34:10 AM
I put the most stock in #2, regarding the availability of price and performance data on services and providers.
I would also add a #5 hypothesis: People use health insurance to pay for normal, expected, and repeating expenses rather than to insure against some risk (probability < 1). Why do we use insurance to pay for birth control and annual check-ups?
I believe the solution is in the HSA/HDHP. When I am holding the cash to pay for procedures, I am far more interested in shopping prices for medical services.
The HSA is efficient in other ways as well. The premium I pay is less than my previous employer paid for a group health plan.
Posted by: david at Nov 22, 2005 11:04:52 AM
It isn't regulation that causes the adverse selection, its preexisting conditions. People want to be protected if they fall into the 5% of people with nasty conditions; this isn't the same as the worthless end-of-life care.
Any evidence people are overconsumming birth control and annual check ups?
Posted by: yoyo at Nov 22, 2005 11:29:24 AM
For what it's worth, my personal experience reinforces #1. For years I was a computer consultant and maintained personal health insurance (for general principles, I thought everyone did). My buddy the insurance agent guided me on getting a policy ("don't put 'self-employed' that means you are a bum").
He also told me horror stories, including one about a client who had gotten a required physical exam, and the blood work came back with a "high normal" on the prostate cancer test. Despite being within the normal range, he was refused coverage. He retook the test, and scored completely normal. At this point the insurance company offered him coverage for four times the normal rate!
Anyway I kept my original Blue* policy for years, and ever year they increased my rate. I asked my buddy, why is this? He said, "They raise your policy on the assumption that something is wrong with you. If you are healthy you'll leave and get a new policy somehwere else. Only sick people stay."
I could continue, I went to a permanent position (with company health care) before retiring way early and searching for my own. Long-story-short, even though I am healthy I was refused everywhere until I argued in writing and got an independent policy.
They make you fight for coverage, make no mistake. And not only do you need to be healthy, you've got to have the resources and skills to navigate their system.
Posted by: odograph at Nov 22, 2005 11:32:52 AM
yoyo said "It isn't regulation that causes the adverse selection, its preexisting conditions."
I think it is, in the sense that the regulations define a battle. The regulators want to get everyone covered, and the insurance companies want to maximize profit. The way to do that, clearly, is to get the other company to take the lemons.
Posted by: odograph at Nov 22, 2005 11:36:37 AM
#2 is a different level of cause from the others. If we don't have better institutions, we need an explanation for that fact. The other three explanations are more fundamental.
Posted by: Robin at Nov 22, 2005 11:39:59 AM
"The problem with (b) is that some fraction of our expensive procedures are steps towards better research and cheaper procedures for other, non-terminal people. Terminal cancer patients enrolled in cancer trials are one example. A larger effect may be in the value of repeated surgeries for the surgeon"
True enough, but much of the expensive end-of-life care that's so costly does not involve experimental procedures or complicated surgery. It instead involves "warehousing" the senile elderly in nursing homes or keeping the brain-dead (technically) alive via life support. Care of this sort produces little or nothing in terms of research or training.
Posted by: Peter at Nov 22, 2005 11:51:44 AM
Odograph's experience points not to #1, but rather to something slitghtly different: The amount of expense insurance companies go to in order to deny claims. A profit-driven enterprise whose job is to pay out money is driven by rather contradictory motives, is it not?
In fact, ANY sort of enterprise whose job it is to pay out money is going to operate in a very skewed market, no?
Posted by: Tim at Nov 22, 2005 11:55:50 AM
I like your four candidate hypotheses, but I think it would be useful to split the question in two: why doesn't the health insurance market work well, and why doesn't the market for health services work well? Your #1, adverse selection, is certainly the main reason behind the first question. I'm not so concerned that government provision would be "lazier" when you're talking about provision of insurance, not health services. As is evident in many of these comments, the easiest way for private insurers to be less lazy is by risk-selection, not by providing services people want. Government could take over the insurance function, make insurance available to everyone, but buy services from private providers - the buying power would just be much more concentrated than it is currently. Opinions on whether this is a good or a bad thing are varied.
Your #2 addresses the health services market and is essentially moral hazard. People are not good health care consumers because the insurer pays, there is no good information on value for money (as you say, it's just really hard to measure), and simply because it's hard to be a good consumer when you're sick.
Posted by: Peter Hussey at Nov 22, 2005 12:01:57 PM
I'm curious why you left off:
5. Administrative costs. Billing and paperwork. The private system as currently conceived is byzantine and confusing. It costs a great deal of money just to navigate it. In many cases this is deliberate. For example, insurers aggressively audit claims looking for even small mistakes that would allow them to deny payment. Doctors and other providers then spend to defend themselves. I believe I read of a recent study in California which found billing and paperwork to account for at least 20% of costs in the private health care system. In contrast, Medicare is widely reported to have only a 2% or 3% overhead. Now, I suppose you could argue that by spending less on auditing, Medicare is wasting money and is thus less efficient overall, but I'd like to see the evidence.
6. Credit risk. It seems to me that we have something approaching de facto universal coverage already. It's just dysfunctional. Emergency rooms, at least, are prohibited from denying care. Many people with large medical bills end up declaring bankruptcy. These costs are obviously recovered by raising the overall price of healthcare, and thus paid by insurers and the rest of us. Furthermore, receiving expensive emergency care at the last minute when complications arise from perhaps minor conditions does not strike me as efficient. It seems likely that many of these could be treated earlier (more cheaply) or prevented entirely if more people could afford to see a doctor regularly.
Posted by: Jack Lecou at Nov 22, 2005 1:43:30 PM
Couple observations:
1. With adverse selection, it's important to distinguish between the direct losses caused to insurers by adverse selection and the loss to all policy holders caused by the actions insurers must take to protect themselves from adverse selection. In general, adverse selection in the market for medical insurance is "not a big problem" only because insurers engage in very expensive risk classification and underwriting practices to avoid it, which in turn impose costs upon all policy-holders. Thus when insurers succeed in "solving" the adverse selection problem, in the sense that they succeed in charging most policy-holder something like actuarially fair premiums, is not really "solving" the adverse selection problem, insofar as the measures they take create significant transaction costs, impose deadweight losses, etc. If you look at the standard sort of breakdown of health care spending in the United States and Canada, one of the numbers that really leaps off the page is the level of spending on insurance overhead, which is about 4 or 5 times higher in the United States. There's other stuff going on, but that's basically where you see the adverse selection problem (or more precisely, it's where you see the cost of adopting a less efficient solution to the adverse selection problem).
2. I've always thought the best way to get a qualitative feel for why markets in health care are troubled is to take a dearly beloved pet to the vet. There's (typically) no insurance to cloud the issues, you're just straight-up paying for medical care. The vet looks you in the eye and says, your dog has cancer, I recommend that his leg be amputated. What do you say? Gee doc, how much will that cost? Naw, I think I'll just let him die. Most people don't even have the courage to ask the price in advance of service (for fear of how it makes them look). The vet does whatever he wants, names a price, you pay up, end of transaction. Now imagine it's your grandmother.
3. As for end-of-life issues, and in particular, how to approach the delicate task of cutting people off when the quality-of-life improvements no longer justify the cost (and given the fact that no one is ever paying out of pocket), I think there is enormous wisdom in Ronald Dworkin's piece, "Justice and the High Cost of Health Care" (reprinted in his Sovereign Virtue). What is elegant is the way he proposes using the insurance mechanism as a non-utilitarian rationale for cutting people off (basing entitlements on the level of premiums people are willing to pay ex ante).
Posted by: Joseph Heath at Nov 22, 2005 1:45:43 PM
Add me to the list of those who say that the question of preexisting conditions must be addressed. I recently shopped for private health insurance (when it appeared that my former employer would abandon 25 years of "When you retire, you can stay in the company group plan on the same terms as employees" promises). The one group plan which I could join that covered preexisting conditions had premiums twice as high as any other plan. The broker I was working with said that the HSA/HDHP she represented would not increase the premium based on such conditions, but would specifically exclude the cost of treating them from the coverage. As others have mentioned, we can deal with the common cases, it's the uncommon ones -- eg, someone who requires dialysis that costs more than they earn in a year -- that create problems for the market system.
I believe that we will eventually end up with a national insurance plan, and its creation will be driven by the group aged 55-64. As more conditions can be detected, and more and more expensive treatments become possible, the private insurance firms will eventually price that group out of coverage entirely. It is worth remembering that Medicare was not created because someone thought it would be a nice thing to do -- by the mid-1960s, it was nearly impossible for those aged 65+ to buy private insurance.
Posted by: Michael Cain at Nov 22, 2005 1:59:28 PM
"I believe that we will eventually end up with a national insurance plan, and its creation will be driven by the group aged 55-64. As more conditions can be detected, and more and more expensive treatments become possible, the private insurance firms will eventually price that group out of coverage entirely."
What is strange is that people in the 55-64 age group are more and more likely to retire, sometimes even before getting into that group, despite the difficulty they often face in getting health insurance. Wouldn't it be worth their while to keep working and enjoy the benefit of employer-paid insurance? Not to mention the waste of human capital that too-early retirement often represents.
Posted by: Peter at Nov 22, 2005 2:35:45 PM
We really need to work on improving our health care. I hope we can soon resolve this issure and all have universal health.
Posted by: Blue Cross of California at Nov 22, 2005 3:03:54 PM
Geez. Reading this post, you'd think that nationalized healthcare was some kind of cutting edge concept that has never been tested. Well, it's not. As Tyler well knows, just about every first-world country has nationalized healthcare. It works for them. Perhap their systems are not perfect, but they're all better than ours.
Posted by: Battlepanda at Nov 22, 2005 3:16:17 PM
"Not to mention the waste of human capital that too-early retirement often represents."
That's an interesting consideration, and one that I sometimes think about in my early retirement ... I tell people I do "newtonian work" as I hike and bike.
Do you think that by not seeking and holding a programming (or similar) job I hold down the economy, or do I just let it be filled by someone who needs it more?
BTW, you do benefit from my blog comment insights ;-)
Posted by: odograph at Nov 22, 2005 3:29:44 PM
Seems to me that it doesn't have to be a choice between public and private. We all want to have health care when we need it- it doesn't seem to harsh to legislate that we should have to have insurance. Then it could be legislated that insurers can't charge higher premiums on the basis of pre-existing conditions. (Though they might lower premiums for persons who make choices that reduce their risk- excercise, regular checkups, not smoke, etc- but that's maybe to much detail at this point.) The idea of insurance is to spread risk over a larger group of people, not to charge those who actually need health care the actual cost of their health care. Insurance companies will lose money on some customers, and most customers will never receive back the majority of what they pay in premiums. That's the whole idea.
The government could set up a not-for-profit scheme that would have to be self supporting from premiums, not taxes, that would compete with all comers. If the government service was lousy, private companies would pop up and offer a better service that people would select. As several people have already mentioned, the admin costs of an NHI are fairly low, the government fund could continue to operate even with very low numbers, to prevent profit gauging etc.
Individuals without the means to afford basic insurance would receive a (tax-funded) voucher that could be used with any provider, public or private.
Anyways, it doesn't need to be a choice between public and private, the two systems can operate as a check on each other, if the correct regulations and no more are in place.
Posted by: Daniel at Nov 22, 2005 6:03:22 PM
People all the time let their cats and dogs die because their medical care is too expensive. In my social class spending that kind of money on a pet is considered ridiculous (and I personally consider it immoral).
Posted by: Noumenon at Nov 22, 2005 10:33:01 PM
