Bryan Caplan  

Krugman Blames Adverse Selection Problem on Government Regulation

PRINT
Arlo Guthrie Says TANSTAAFL... Government Health Insurance an...
Not long ago, Krugman chided economists like Mankiw for overlooking moral hazard and adverse selection, the supposedly insuperable textbook barriers to free-market health insurance.  I objected:
As long as rates are legally allowed to reflect risk, there is a lot of evidence that selection is actually advantageous rather than adverse.  Despite its massive citation count, Arrow's paper is deeply wrong.  But even if Arrow were right about the empirical importance of adverse selection, government regulation doesn't do much about it.  In fact, it often creates adverse selection problems that wouldn't exist by imposing regulations that make it harder for rates to reflect risk.
I'm struck, then, that Krugman now seems to have covertly accepted my point!
The reason we have restrictions on interstate sales of health insurance is that a number of states regulate insurers. In particular, some states have a form of community rating, which basically says that insurers can't deny you coverage or charge extremely high premiums if you have a preexisting condition. And community rating will be unsustainable if individuals can buy insurance from out of state; insurance companies in states that don't have community rating will cherry-pick the healthy, good risk people, leaving the community rating states with only the highest-cost people.
Instead of bemoaning insurers' inability to distinguish low- from high-risk customers, Krugman now admits that they're quite able to do so.  By his own account, adverse selection arises because the natural free-market response - higher rates for higher risk - is illegal.

Alas, Krugman can't just admit his mistake.  He has to toss in a new error:
And what are the results of competition there? The answer is that insurers compete by doing their best to deny coverage to anyone who might actually need medical care.
Nonsense.  Insurers compete to charge people rates that reflect the cost of their care.  If you can reasonably be expected to need 50% more care, they'll be happy to sell it to you for 50% more.  Take a look at life insurance.  They don't "deny coverage" to anyone over the age of 40.  But they do charge higher premiums.  Why is that the end of the world?


Comments and Sharing





COMMENTS (21 to date)
RD writes:

You forgot the latest from Krugman:

"And I thought I was just saying what everyone knows. Put it down to the wonkish fallacy: when you spend a lot of time immersed in a subject, you tend to forget that others aren’t."

http://krugman.blogs.nytimes.com/2009/07/27/markets-and-healthcare-a-startling-reaction/

Don't expect a response, you must be FAR outside the mainstream to not see it his way.

Kurbla writes:

You were right before, and Krugman is right now.

Premiums converge toward actual price of the medical treatment. Perfect insurance agency would recognize with 100% accuracy that I'll have disease X, and that price of the treatment would be 4000 dollars, and they will charge me with premium of 4040 dollars. Closer to the perfection, less sense insurance agency has.

From leftist point of view, plain solidarity and not risk management is really needed here.

hacs writes:

The point is how to solve that, because simply to let insurers charge higher premiums doesn't assure admittance to the system. The solution has to have at minimum two characteristics: to insure everybody and to reduce the costs. Anything without those aspects is free-market or governmental trash.

Max writes:

Well, this accuracy can be achieved for almost any easy treatment (cold etc.), because the costs of an aspirin or antibiotica is easily known.
When uncertainity comes in, it is with more dangerous and complicated procedures where you can't judge the final outcome, because costs rise along the way. Also, you can't determine when someone will get those kind of diseases, so there is also a time-variance and a statistical uncertainty.

But this gives a fairly good lead on where an insurance is necessary and where not.

stephen writes:

Max

Thats why actuaries learn about stochastic processes, aggregate loss models, empirical and parametric probability distributions, time series, statistical inference, price inflation......

Anyway, the point is that it is more then possible to calculate accurate expected losses do to health care costs for a large population. The issue with health insurance isn't the black swan effect. The problem is ex post laws mandating procedures not priced into the models and price caps on premiums making it impossible to cover the losses in a way we would expect a reasonable market to be able to do.

Thomas DeMeo writes:

"By his own account, adverse selection arises because the natural free-market response - higher rates for higher risk - is illegal."

He didn't say that.

He said "...community rating will be unsustainable if individuals can buy insurance from out of state". Because states have different regs, higher rates for higher risk would be exposed in the market.

It would become legal.

Adverse selection doesn't encourage competition, it discourages it.

So- health insurance companies have to cover certain things. There is still significant space to compete on quality and on efficiency, and it never happens. No one breaks from the pack because of adverse selection.

gnat writes:

It is really quite simple Bryan:

The 25 year old Bryan is unwilling to insure the 50 year old Bryan, and given the choice insurance companies would rather insure the 25 year old Bryan. A 50 year old Bryan may have a 15% risk of heart desease yet A 50 year old Bryan with heart desease is uninsurable.

J writes:

The 25 year old Bryan is unwilling to insure the 50 year old Bryan, and given the choice insurance companies would rather insure the 25 year old Bryan. A 50 year old Bryan may have a 15% risk of heart desease yet A 50 year old Bryan with heart desease is uninsurable.

Nonsense. In the real world, if the insurer can reasonably measure the risk, and charge 50 year old Bryan for it, they will. The problem is the real world is prohibited to them. They can't charge a higher rate for higher risk.

The solution then is to eliminate community rating, knock down state restrictions on out-of-state purchasing of insurance (I find it odd that I can purchase a car from Tennessee or a bottle of wine from California, but not an insurance policy from Wisconsin), and eliminate state mandated coverage. The thing I noticed when I received my first employer provided insurance is that even though I was a single male with no girlfriend, I had to have maternity coverage. Odd, no??

Thomas DeMeo writes:

J- If you take the position you do, you have to deal with the problem it causes.

You look at 50 year old Bryan and his 15% potential risk of heart disease. In real life, that means 15% of 50 year olds actually have heart disease. When any of them need to change policies, perhaps due to a layoff, the market prices accordingly for the certainty of large medical bills. They can't pay. What do you do?

Niccolo writes:

Transition away from employer mandated health insurance and into more consumer driven insurance by finally getting rid of the old 1930's style tax system.

And gnat,

The entire concept of the "uninsurable" rests on the presumption of government regulations. In reality, no one is "uninsurable." An owner of a company is usually willing to cover anyone, for a price, that is.


So it all comes down, once again to costs. Forget the idea of increasing competition between insurers by adding a public plan because it won't address the real issues - cost of delivery systems.

Tom Crispin writes:

DeMeo:

What to do? You rob Peter to pay Paul, or in this case, Arnold to pay Bryan.

gnat writes:

Niccolo

My solution would explicitly recognize that the employer's share comes out of the employee income and have mandatory contribution to health insurance with minimums (e.g., HSA plus catastrophic coverage). That solves the Bryan-25- doesn't-like-Bryan-50 problem. The firm acts as administor collecting premiums from the employee and the mandatory insurance requirement moves with the individual. This provides a demand for insurance as distinguished from prepaid healthcare, and may allow an insurance market to development, provided the cost of delivery is predictable. For the later, you have to fix delivery. Today doctors are specialists, are the only participant with the information and can charge whatever they want.

Michael Kolczynski writes:

You're challenging the arguments of a dishonest person. It's a shame, but that's just what he is, dishonest.

J writes:

J- If you take the position you do, you have to deal with the problem it causes.

You look at 50 year old Bryan and his 15% potential risk of heart disease. In real life, that means 15% of 50 year olds actually have heart disease. When any of them need to change policies, perhaps due to a layoff, the market prices accordingly for the certainty of large medical bills. They can't pay. What do you do?


I don't have to do anything. Or at least I have no moral obligation toward Bryan to do anything, it doesn't necessarily mean I won't. I think the question you meant to ask is: What does Bryan do? And I'm not Bryan so I have no idea. But he has to make a choice just like we all do. I think where you run off the rails is that you assume that insurance is somehow fundamentally different than other products and other markets. I agree there are differences, but not fundamental ones. If you get government out of the way, and allow individuals to negotiate and contract with third parties (severing the employer-insurance link, something I failed to mention in my previous post)I'm pretty sure Bryan would have something available to him. But of course we're never going to find out if some individuals won't let others try.

Max Borders writes:

Krugman also seems to be confused about the difference between community rating and guaranteed issue. Doesn't community rating mean you charge young, healthy people the same as old, fat, sick people? Maybe he's saying guaranteed issue is a form of community rating. But guaranteed issue makes insurance not insurance at all, but prepaid healthcare. In any case, I think it's pretty clear Krugman is out of his league on this issue. It would be better simply to subsidize poor people to get insurance and get folks with pre-existing conditions into high risk pools--if you're going to NOT be free market.

Colin K writes:

We don't expect life insurance companies to sell million-dollar policies to people with metastatic cancer.

In a financial sense, life insurance and health insurance are not all that different. One pays out before you die and the other pays out after. Both have to factor in uncertainty about when and whether a payout event would occur.

IMHO health insurance should be structured more like a flexible-payout annuity where the consumer agrees to a payment schedule in exchange for a lifetime benefit of up to $X. The insurance company would have only an advisory role in deciding how you spent that money. If you have $500k of benefit and want a drug that costs $100k but improves your chances of survival by 5% over the $10k drug, well, you can have it.

J writes:

Doesn't community rating mean you charge young, healthy people the same as old, fat, sick people?

Max:

I think community rating means that you charge the same premium for everyone in some specified group, where the group is composed of both low-risk and high-risk buyers.

Guaranteed issue only means that you can't be turned down for coverage because of your health.

matt writes:

I think some people are confused, thinking about the 50 and 25 year old Brian. Imagine the 50 year-old has heart disease he no longer need insurance he needs money to pay for his health care. Think about health insurance vs disability insurance, in the later you pay and pay, and when you get disabled you stop paying, and get paid. Health "insurance" doesn't work like that. Once you get sick what are you being insured against, a different sickness. I have seen, on this very blog the idea that if you get a sickness your insurance pays out a certain amount of money which you can use for treatment as you see fit. That isn't a perfect solution but at least it addresses the problem.

People need to be insured, but they will eventually be sick. If you charge them for insurance 100% for the care they will use while sick its not insurance, its paying for care. The goal of national health care should be to nudge people into paying for their lifetime care needs over a long time horizon, and to redistribute health based on luck. We cannot expect people to do this on their own. Community rating and mandatory insurance are needed, but there is no reason the government must be involved in any part of the administration of health care.

Boonton writes:

J

Nonsense. In the real world, if the insurer can reasonably measure the risk, and charge 50 year old Bryan for it, they will. The problem is the real world is prohibited to them. They can't charge a higher rate for higher risk.

And then this ceases to be insurance. If knowledge became perfect then premiums would simply be whatever your medical expenses are plus some profit for the 'insurance' company to manage the paperwork. What you have is essentially a Christmas club type account to help you save up for known medical expenses.

Colin
IMHO health insurance should be structured more like a flexible-payout annuity where the consumer agrees to a payment schedule in exchange for a lifetime benefit of up to $X. The insurance company would have only an advisory role in deciding how you spent that money. If you have $500k of benefit and want a drug that costs $100k but improves your chances of survival by 5% over the $10k drug, well, you can have it.

A problem is that unlike life insurance, we as a society have decided you get healthcare regardless of your ability to pay. Say you have a lifetime policy of $500K. You spend $499K chasing a cure for cancer and you get it! Now you're 40 years old and you have a new lease on life. But wait:

1. You have a lifetime of additional premiums to pay to the insurance company because you certainly didn't put in $500K before you got early cancer. How does the insurance company force you to make payments?

2. You now have a lifetime more with no coverage since you exhausted your policy....yet you must keep paying premiums. If at age 45 your have a sudden heart attack in the supermarket are people not going to call 911? Is the hospital not going to treat you?

Boonton writes:

J

The thing I noticed when I received my first employer provided insurance is that even though I was a single male with no girlfriend, I had to have maternity coverage. Odd, no??

Let's think about this. If you're getting charged for real for maternity coverage then women must be getting charged less. After all the portion of premiums you are paying for maternity coverage will never come back to you, right?

But almost certainly the premiums are probably the same for either men or women. So what does it mean that you have 'maternity coverage'? It means absolutely nothing. If your doctor submits a claim against the 'maternity coverage' you have we both know its going right to the fraud department.

So what is really happening is that you (and everyone else) is putting money into a big pot that covers a bunch of medical expenses. some of them you know you won't need (maternity), some you think you probably won't need (Ebola treatment) and others you can't be so sure (cancer, diabetes etc.). On the other hand, some things you may need (prostate checks, viagra) your female co-workers will not have much use for.

This is a not quite like risk-based insurance...it's a bit like going out to lunch with your co-workers and splitting the check evenly because its too much work to figure out what everyone's cost really is. What you're doing is less about risk management and more about cost sharing.

J writes:

Let's think about this. If you're getting charged for real for maternity coverage then women must be getting charged less. After all the portion of premiums you are paying for maternity coverage will never come back to you, right?

But almost certainly the premiums are probably the same for either men or women. So what does it mean that you have 'maternity coverage'? It means absolutely nothing. If your doctor submits a claim against the 'maternity coverage' you have we both know its going right to the fraud department.

So what is really happening is that you (and everyone else) is putting money into a big pot that covers a bunch of medical expenses. some of them you know you won't need (maternity), some you think you probably won't need (Ebola treatment) and others you can't be so sure (cancer, diabetes etc.). On the other hand, some things you may need (prostate checks, viagra) your female co-workers will not have much use for.

This is a not quite like risk-based insurance...it's a bit like going out to lunch with your co-workers and splitting the check evenly because its too much work to figure out what everyone's cost really is. What you're doing is less about risk management and more about cost sharing.

Actually what it says is that the state has mandated all insurance policies sold in the small group market in the State of Michigan have maternity coverage. And if it is about cost sharing, do you mind picking up my mortgage next month, I'm going to be a bit short? Thanks :)

Comments for this entry have been closed
Return to top