Arnold Kling  

Austan Goolsbee on Health Care

What's Wrong with REVENGE?... Ballots Are Better than Bullet...

He writes

If you set up a market-based health system, allowing insurance companies to pick and choose who and what they will cover, you give them overwhelming incentives to dump, deny, avoid and neglect the sick people.

But the data do not show this. Brad Herring and Mark Pauly have shown that insurance companies engage in risk pooling, not adverse selection. It comes as a great disappointment to many economists, but adverse selection is not the core problem of our health insurance system.

Goolsbee gets much else wrong.

He says,

A single-payer system here would have to also include some truly major rearrangment of the tort system to bring those costs down.

Lawsuits, like adverse selection, are over-rated as a cause of high medical costs. We would like to believe that there is an enormous free lunch out there from tort reform. There might be some modest savings, but not enough to make a dent in overall health costs.

Goolsbee goes on,

we should start by fixing the most glaring problems of our system without ripping it up and starting over. We could use pooling to move away from the dump-and-deny insurance we have now. We could reward doctors for doing a good job, the way they do in the United Kingdom. We could focus more on preventing sickness, the way they do in Cuba, to reduce the number of illnesses.

In the terminology of this essay, Goolbsee represents the Democratic wonks, as opposed to the revolutionaries. But the wonks do not have much to offer.

Preventive care is like motherhood and apple pie, but we don't have any hard evidence that we can use preventive care to save money. I would argue that some types of preventive care, such as cancer screening, tend to have a very high cost per life saved.

To "reward doctors for doing a good job," you have to know what a good job is and you have to be able to measure it from far away. This is extremely difficult. Imagine trying to run a system in Washington to pay professors for "doing a good job."

Forced-pooling health insurance is not a solution, because "dump and deny" is not the main problem. A bigger problem with the individual health insurance market is that there are 50 state regulatory fiefdoms, and insurance companies are not allowed to market products across state lines. The biggest problem is that most people think that employer-provided health insurance is "free." So when they do not get insurance from an employer, they cannot bring themselves to pay for what other people get for "free."

The biggest problem of all is that nobody, here or in other countries, wants real insurance. Instead, everybody's idea of health coverage is to be insulated from costs.

As I pointed out in the essay, the wonks want to keep the existing system going, with a few patches here and there. If the Democrats take the wonk route, then they had better lower everyone's expectations about what they will accomplish. They certainly are not going to slow the rate at which health care spending is gobbling up tax revenues. They certainly are not going to slow the rate at which employer-provided health insurance is unraveling under the pressure of swelling costs. They certainly are not going to address the coming bankruptcy of Medicare. They are not going to come to grips with America's extravagant use of medical procedures with high costs and low benefits.

The wonks do not wish to confront Americans concerning our beliefs about health care--the belief that someone else should pay for it, the belief that cost-benefit analysis is inappropriate, the belief that someone who is suffering or dying should be spared no effort in terms of treatment, and so on. If the beliefs persist, then so will all of the major problems. There is no clever wonkish way to get around them.

Comments and Sharing

TRACKBACKS (3 to date)
TrackBack URL:
The author at Asymmetrical Information in a related article titled How adverse is adverse selection? writes:
    The problem of adverse selection is generally acknowleged to plague the market for health insurance, which is why many people think we need the government to finance it. For those who are not familiar with the concept, adverse selection is what happens... [Tracked on July 2, 2007 10:43 AM]
COMMENTS (12 to date)
Karl Smith writes:


One question I have, which you have probably discussed elsewhere but I missed, is why people are so attracted free healthcare.

Of course, people want things for free but why is health care singled out. Historical accident seem hard to swallow since, the cry for universal health care is well, so universal. Not only here in the US but abroad.

Is there some externality that we are missing? Perhaps, what people really want is not to have to see sick people?

Perhaps seeing a sick person makes us feel sick. Therefore, left to their own device everyone will under-insure because they are ignoring the emotionally painful effects their sickness has on everyone else?


Lauren writes:

Hi, Karl.

Robin Hanson did a podcast recently on EconTalk where he discussed this very question. One answer he suggests that is consistent with the evidence is that, because we are terrified of and don't want to think about death, there are some things we prefer to overinsure or lay off responsibility for. Another answer is that it ultimately stems from evolutionary selection for signalling loyalty.

You can find the podcast at Hanson on Health. Skip to the halfway point if you want to get right to the answers I've summarized, but I recommend listening to the whole thing.

Kurt9 writes:

I think the reason why people expect "free" health care even though they do not expect "free" care maintainance when the car breaks down is partly because of the huge costs associated with medical treatment. Car repair can set you back a few thousand dollars. Home repair can set you back a few tens of thousands of dollars, but this is quite rare. Medical treatment, on the other hand, can be several hundred thousand dollars.

I think the other reason why people do not consider the risk insurance model to be appropriate to health care is because, unlike the car or home, you do eventually need health care because of the aging process. Most people replace their cars before they really wear out. Most people move or build a new home before the old one wears out.

The problem with medicine is that you cannot (yet) replace or regenerate your body in the same manner as a car or a house when it slowly becomes non-functional due to the aging process. Since current medicine does not treat aging itself, everyone faces the prospect of huge bills for "band-aid" medical treatments that does not treat aging itself as you get older.

If we had something like SENS (assuming that it is not too expensive), then conventional medicine would be needed only to treat trauma. Once the need for medicine is needed only for trauma, I expect the risk insurance model to be quite popular for medical treatment.

Karl Smith writes:


Thank you, that was fascinating.

The signaling hypothesis is cute but ultimately I am it falls a little flat with me.

I actually, like the idea that people "don't like to think about death" though I would posit the more general hypothesis that

Decision making is stressful and the level of stress is proportionate to the difficulty of the decision.

To borrow a page from Bryan, people then rationally adopt unfounded views such as

1) My attempts to revive my loved one are working

2) My doctor knows a lot because he is friendly

3) I don't need to buy this procedure because I am going to be fine anyway. (Which is why less medicine is purchased at the point of delivery than in advance)

They also get upset when you attempt to disturb this view because it will require them taking on costs they don't wish to take on.

What really confounds matters, however, is that irrationality is also an effective treatment. That is there are placebo effects.

It could also be that we invest a lot in treatment because believing that we are getting the best treatment is an effective treatment.

Also, a key phrase is that "We are more afraid of thinking about death itself"

Well, when your loved ones are dying in front of you, you are forced to think about death. Your willingness to pay, to think that death is less likely may be much higher than your willingness to pay to prevent that death.

Thus, finding out that you hospital has a bad could actually have negative value.

Gary Rogers writes:

I love Karl's question. What is it about health care that makes it different?

I would suggest first that governments world wide do subsidize other "basic necessities" including fuel, food, education, wages and rent. When these are subsidized, they become a target for negotiation just like we are doing right now with health care. Since subsidies cause increased costs and allocation problems, there are always issues to be resolved. Where food is subsidized, the issues are with food. In our case it is health care.

Second, the way health care has developed in the United States, the subsidies are neither equal nor fair. As I understand the history, health care was added as an employee benefit during the wage and price controls of WWII. Medicare and Medicade were added in the 60's to bring another large group under subsidized health care. Since both of these are closed markets with considerable negotiating power, the resulting cost increases are passed on to the remaining segment of the market that pays cash for health care and, in this case, has comparatively little negotiating power. This creates an unfair subsidy that is easily attacked by political opportunists.

I would also point out that our government subsidizes education, health care, housing and food in that order. If you look at cost, problems and political issues there is a significant correlation. So, in summary, here is how I would summarize my beliefs about subsidies:

The more something is subsidized, the more the subsidy appears to be necessary.

The more something is subsidized, the more expensive it becomes for everyone even though the costs are hidden by paying indirectly through the government.

If you are going to subsidize, do it equally for everyone. (The argument for nationalized health care)

If something is never subsidized, everyone is better off and there are very few complaints. (The best solution)

Biomed Tim writes:

"Lawsuits, like adverse selection, are over-rated as a cause of high medical costs. We would like to believe that there is an enormous free lunch out there from tort reform. There might be some modest savings, but not enough to make a dent in overall health costs."

I am suspicious of this claim. The straight cost from malpractice suits may be modest, but I would guess that the opportunity cost of practicing "defensive medicine" is quite high. Physicians frequently order extra tests in order to protect themselves from legal scrutiny; those are resources and time that could have been spent elsewhere.

Does anyone have any idea where I can find some hard data regarding the cost of defensive medicine?

Gary Rogers writes:

I agree that lawsuits are probably overrated as a potential savings. I am not a doctor, but it would be reasonable to find out how much malpractice insurance costs and spread that amount across total billing for the same period to get an estimate of the total impact. I also suspect that when the patient is making the decisions he or she is less inclined to sue over their own decision than one that is made by a buraucracy as experienced with socialized health care.

Ashley writes:

Biomed Tim--

You might try starting with the Kessler-McClellan study, but I don't know for sure if that is what you are looking for. If nothing else, it might give you ideas of where to look for the data you want.

dick white writes:

I agree with Arnold's description of the health care problem but I don't understand how a
knowledgeable economist like Austan Goolsbee would so misinterpret the data to reach the conclusion he has. Any thoughts?

bingo writes:


I agree that the actual, measured costs (insurance, settlements, jury awards) of medical malpractice are modest in relation to the $2 trillion spent on medical services in the United States. However, it has been estimated in many places that anywhere from 15-25% of ALL medical expenses are the direct result of defensive medicine. If that is so there are $300-500 BILLION being spent each year on medically unnecessary care.

Our present system of payment for medical services does not provide any disincentive to the patient, the purchaser of the service, to ask for every last possible medical care item for a particular problem. In addition, no amount of care that can be ordered by a doctor is enough care if additional care will decrease the liklihood of being sued for malpractice, or if having been sued from losing the suit. A diagnosis is never secure enough and a teatment is never cure enough. There is nothing to decrease demand for either the consumer (patient) or the provider (doctor).

Unlimited demand equals infinite cost.

Radical tort reform will remove the incentive for the provider to order medical care that you have termed "premium care". Unlimited demand from the provider can then be managed; outcomes data and statisical probability of desease presence can then be applied to care if the need to "treat the chart" is no longer there.

Much of your work and most of the proposals circulating speak to the control of care requested by the patient--patients will request less "premium" or extra care once the incremental benefit of said care is quite small--if they are no longer insulated from the cost of that care. But there are two demand points in our system, and removing the need to practice defensive medicine will, indeed, decrease provider demand and free up a substantial sum.

bingo writes:


Do you review the comments on your essays? Do you participate in the conversation?

George writes:

"We could focus more on preventing sickness, the way they do in Cuba, to reduce the number of illnesses."

For example, Cuba is way ahead of us in preventing obesity-related illness. Unless you're a Party leader.

Comments for this entry have been closed
Return to top