David R. Henderson  

Adverse Selection by Law

PRINT
Boone Pickens Illustrates Beck... The Quants...

Finally, Krugman admits the problem with a ban on pre-existing conditions clauses in health insurance.

From Paul Krugman's column today:

Suppose, for example, that Congress took the advice of those who want to ban insurance discrimination on the basis of medical history, and stopped there. What would happen next? The answer, as any health care economist will tell you, is that if Congress didn't simultaneously require that healthy people buy insurance, there would be a "death spiral": healthier Americans would choose not to buy insurance, leading to high premiums for those who remain, driving out more people, and so on.

He could have generalized and said, "as any economist who has thought about insurance will tell you." This knowledge goes well beyond health economists.

This is what I called, in a 1994 op/ed, "adverse selection by law."

What's Krugman's solution? An individual mandate. But what he doesn't tell you in his article that urges the House to pass the Senate bill is that the penalty for not getting health insurance is only 2% of income. So if someone making $30,000 wants to avoid getting health insurance, he would pay an annual tax of only $600. And this, Krugman apparently thinks, is enough to cause that person to buy health insurance even though he can wait until he gets sick (remember: insurance "discrimination" based on health is banned) to buy it. Does he really think that $600 or $800 or $1,000 is enough of a penalty? I doubt it.

So what Krugman is really calling for, wittingly or un, is a "death spiral" for health insurance.


Comments and Sharing





COMMENTS (19 to date)
Leo writes:

If only the bill's proponents had thought to include some sort of an additional incentive for low income people to purchase insurance -- a subsidy of some sort perhaps.

Oh, well, can't think of everything.

RL writes:

Alternatively, Krugman may want to get a tax established, and let Congress modify the level of tax later. All of a sudden, to prevent the death spiral, Congress finds it MUST tax people at 5% or 10% or more of their income...granted, something most people would feel too high, but the principle of taxing non-insurance buyers was already established.

David C writes:

My guess would be the cheapest type of insurance plan most people can get (one with a high deductible and co-pay) is probably approximately $600 a year for an individual. For a lot of people, it'd be cheaper to buy the cheapest insurance available. They'll have a strong incentive to upgrade their insurance after they get sick, but that will push the premiums of the most expensive plans up while leaving the cheaper insurance plans the same. Eventually, a lot of insurance companies might refuse to offer the more expensive plans because it's impossible to make money off of them. This would lower health care costs overall as more and more people would be forced to stick with the cheapest plans out there. Using this logic, a low tax penalty is a good thing.

Nathan writes:

Is adverse selection really as bad as naive theory would suggest? Many health problems are self-inflicted through poor lifestyle choices. My gut tells me that people who drink, smoke, overeat and do other things to endanger their own health are not the type to buy health insurance just because it's a relative bargain to them, whereas healthy, clean-living, risk-adverse individuals are exactly the people who insure out of a sense of personal and family responsibility, even if the rates they get are sub-optimal.

Of course, if we take insurance non-discrimination laws to the extreme--by, say, letting people buy "insurance" after they've already been injured or diagnosed with cancer--we still run into problems. But if we just consider general health history and still exclude specific pre-existing conditions, I tend to think that those who would be the biggest drains on insurance companies are the exact sort of people who have little interest in buying insurance.

(Note that the above is not an endorsement of such laws, nor does it solve the problem of what we do with the uninsured-by-choice when they do get sick.)

ThomasL writes:

This is a very tricky problem. It strikes me as something not dissimilar to a fixed exchange rate on a bi-metal standard.

It is highly unlikely that the fixed ratio reflects the market reality, so people seasaw back and forth between the metals depending on the real exchange rate between the two, taking advantage of the artificial exchange rate to make a quick profit.

The mandate+penalty and the cost of insurance seem like much the same thing, really. The end result is going to be like one of those desk-toy wave things, with people flowing back and forth depending on the current balance point of the cost of the penalty and the cost of insurance.

gamut writes:

David C:
- "..but that will push the premiums of the most expensive plans up". Correct.
- "..while leaving the cheaper insurance plans the same...". INCORRECT.
- "Eventually, a lot of insurance companies might refuse to offer the more expensive plans because it's impossible to make money off of them...". INCORRECT.

I suggest drawing a little pie chart of share of revenue and anticipated costs to help clear this up your head since you've got the consequences back-asswards.

Right answer:
- Cost of all premiums goes up because real costs of coverage exceed revenue (since cheap plan members don't pay enough in to cover themselves).
- Insurance companies probably stop offering cheap plans because they wind up just being a sham. They collect a mere 600 bucks a year and pay out many times that when people get sick and decide to 'upgrade' last minute.

In fact, the cost of all plans would HAVE to rise to a sufficient level that even relatively cheap plans, now much more expensive, would cover the costs of even sudden at-gun-point upgraded benefits, otherwise the whole thing would collapse.


Radford Neal writes:

- Insurance companies probably stop offering cheap plans because they wind up just being a sham. They collect a mere 600 bucks a year and pay out many times that when people get sick and decide to 'upgrade' last minute.

This doesn't make sense. As I understand the proposal, they wouldn't be restricted to "upgrading" only with the same company that originally sold them the cheap plan. They could upgrade to a plan from another company. So there's no reason for a company to not offer cheap plans, as long as they would make a profit on them if the customer stuck with that plan. In fact, a company offering only cheap plans would make even more money if most customers upgrade to a fancier plan at another company once they got sick.

gamut writes:

That is insane. If this were the case, two classes of companies would develop: one class that would offer the cheapest plans they could while still covering members, and then offering such harsh service so as to push members into 'upgrades' at competitors -- who would by necessity have very expensive policies to both deter new arrivals and stay in business while still being forced to accept them. I can't see such a mechanism as feasible since those same people probably wouldn't have money to upgrade anyway.

Also, another reason why this is nuts is that the bill also requires minimum levels of coverage above those presently offered; As far as I can tell, 600 bucks a year doesn't even begin to pay for those since the mandated federal minimums are analogous to those offered by plans in the 5k range. This means that small plans will explicitly become unprofitable and be eliminated.

Anyway, this is making my head spin. This is why I tend to favour a simpler explanation of the proposed phenomenon. 10%+ increase in contractually obligated demand, 0% increase in prices, and.. 10%+ increase in supply? One or two of those numbers is definitely wrong since they are mutually exclusive. If it's the first, then the plan will fail it's objective, if it's the second, then there will some mighty expensive plans, and if it's three, then get ready for some serious line-ups.

Frankly, since the first number is kind inflexible as law requires it to be true, then the result will be some serious movement in the second and third. Healthcare is a sticky industry, and a 10% increase in demand in any industry is astronomical; see gasoline prices circa 2009 to see what happens when demand increases by an unexpected 1%. And people can actually choose not to drive, while most will not choose to skip the cancer treatment.

David C writes:

"who would by necessity have very expensive policies to both deter new arrivals and stay in business while still being forced to accept them. I can't see such a mechanism as feasible since those same people probably wouldn't have money to upgrade anyway." - gamut

Right, which is why all the expensive companies would go out of business or at least have their market share rapidly diminish. The best way to make money in such a system is to offer cheap plans that are slightly less than what people would pay in taxes. Any rate above the tax rate is going to start incurring big costs very quickly.

"As far as I can tell, 600 bucks a year doesn't even begin to pay for those since the mandated federal minimums are analogous to those offered by plans in the 5k range." - gamut

This may change the equation. I'm not sure what the minimums are, but I doubt they're in the 5k range for an individual. In fact, I'm certain you're overestimating without looking at what the rules are, but they still might prevent a $600 a year plan.

Matt writes:

I am twenty-five years old and I am in perfect health with no pre-existing conditions. I do not have health insurance and I do not want health insurance (at least for any reasonable cost).

If my penalty was $1000 a year and health insurance cost me $1000 a year, I would opt for the penalty, because I would only have to worry about it once a year when I am doing my Turbo Tax.

Also I think that it would be likely that somehow I could get out of it (like a deduciton for someone enrolled full time in school, or for military members).

gamut writes:

Not that a single post is proof, but Matt makes it unlikely that David C is correct. The best way to make money in a system where you're competing against a state-subsidised cheap option (the tax), is to offer a different product. More expensive, and offering more services. Competing against a monopolistic entity with no concern for losses is suicide. If supply could actually meet demand here (which it cannot at the present price), then the entry point for insurance would, by necessity, become just above the default tax amount, and probably above it by an average of the subsidies offered. However, there is no chance at all that supply would increase to meet such a jump in demand without a commensurate increase in fees (and insurance premiums). Like i said: health care is an inelastic good. It basically takes one of three things for you not to get a procedure done: either your insurance doesn't cover it (because it's too expensive), which would be illegal for many procedures, or you die waiting (or give up due to the wait). We have the later here in Canada because our prices are fixed by the government, you will get a mixture of the two.

David C writes:

gamut, Matt's decision is highly irrational. It's very easy to set up insurance so that it automatically deducts from your account (usually your paycheck through your employer), so one only has to worry about it when either the rates change (which can be expected to occur about once a year) or when you decide to change your plan. It requires a lot less effort than paying taxes. Meanwhile, choosing insurance instead of the tax gives people protection from a very expensive medical bill that would occur in case of an emergency. If everybody is as irrational as Matt, then our debate is pointless because there's no telling how people will react to the tax.

"Competing against a monopolistic entity with no concern for losses is suicide." - gamut

The reason for that is monopolies can lower prices whenever they want to knock an individual out of business. In this case, the price is fixed, so it doesn't matter that the tax has a monopoly.

"Like i said: health care is an inelastic good." - gamut

If health care were truly inelastic, then no change in government policy would have any effect on how much people spend on health care. The whole point of reform is that a great deal of health care spending is unnecessary and can be eliminated without any reductions in quality.

"It basically takes one of three things for you not to get a procedure done: either your insurance doesn't cover it" - gamut

There are other ways for insurance to save money. They can increase the deductible, they can increase the co-pay, or they can reduce their share of the co-insurance.

Lord writes:

Insurance is already in a death spiral. My expectation is we won't make progress until it collapses entirely.

Boonton writes:

Actually people do seem to act a bit irrational with health insurance. For example, a few mornings ago I heard a bit of an NPR interview with a woman who had an $8000 a year health plan from her union who was aghast at the idea that the 'Cadillac tax' might cause the value of the plan to be cut but for that savings to be made up for with wage increases. They asked her exactly what her medical use was over the last few years. Answer: She used the ER once over the last few years for a broken arm, other than that nothing. The insurance she is clinging too is hardly being utilized yet she would forgo actual cash increases in her wages to stop any cuts in her benefits.

To be fair I did this too. I carried dental insurance for a few years without using it. This year I got a bit smarter and cut off the vision insurance for $20 a month and opted instead to fund my health spending account enough to cover a pair of glasses and eye exam.

Anway, if people opt to get insurance the cost of insurance should go down and if people are a bit irrational they may opt for insurance rather than pay a scarey 'penalty'.

Also missing from your equation is a proper comparision. It's not a $600 a year penalty versus a (say) $3000 a year policy. Clearly if they brought the policy they'd have a benefit beyond avoiding a $600 penalty. They'd have health coverage. So its the $600 penalty versus whatever benefit they think they are deriving from not buying a policy now.

Let's say someone today deems the value of getting coverage is $2600 but the policy cost $3000. Clearly they would opt to go without coverage. But a $600 penalty would alter that equation. Now they are covered. The moral hazzard equation clearly does not require the penalty for violating the mandate to be equal or greater than the cost of a policy. Granted it may need to be greater than the current bill (or the mandated insurance may need to be lower). It might even be a net social benefit if that fellow figures he can always declare bankruptcy or hit up Medicaid or charity care if he got really sick and that was the reason he values the policy at only $2600

Boonton writes:

So here's the issue in more condensed form:

It's pretty obvious if you mandate coverage for everyone and ban discrimination you probably will lower costs. Insurance companies would no longer spend money trying to push the sick off their rolls onto someone elses. But it's not so easy to really mandate coverage so you implement a penalty.

If people are a bit irrational....wanting to get coverage even if their need for it actually justifies going without coverage or having little coverage, you actually lower costs again because people are buying insurance who won't be taking it out in the form of medical claims. In a world of infinitely intelligent 'Spock like' economic agents insurance would be impossible. Anyone seeking it has calculated his expenses will exceed the premiums therefore it would be irrational for any insurance company to sell it to him.

In the real world there are probably plenty of people 'on edge' in terms of coverage. They would like to have coverage but they judge its cost just less than its benefit. These people are probably relativley healthy. They don't believe they are perfectly healthy (if they did, they would value coverage much, much less than its cost). They also don't believe or know they are sick (if they did they would value coverage much more than its cost). AS a result a modest penalty might be just enough to tip the edge towards getting covered which would lower costs...hopefully helping everyone else out.

Now whats also interesting is that private insurance was/is not screaming against the current bill. Private insurance seemed much more concerned about the 'public option'. This hints that private insurance suspected the modest penalty probably would offset more or less the adverse selection costs that would be imposed by ending discrimination.

So working for this model you have:

* A bit of irrationality. People seem to feel $1 of coverage is worth more than an extra $1 sitting in a savings account for health emergancies.

* People who don't opt to buy coverage don't value coverage at $0. The difference of their value means a penalty much less than the cost of a policy may be sufficient enough to flip their decision making process.

* Toss in another bit of irrationality. People may go out of their way to avoid a $600 penalty simply for psychological reasons. Many of us do this with gas prices. How many times have you driven an extra block or make an awkward u-turn to get at that station that's a few pennies cheaper? With a ten gallon tank your heroic efforts saved you not even a quarter...yet you'll let the supermarket con you into buying candy bars at the checkout line blowing your savings away!

* Insurance companies are screaming about other ideas!

Boonton writes:

Also I'm not sure what merits the 'finally' that begins this post? Krugman has been clear for quite a while now that if you're going to ban insurance discrimination you have to follow up with a mandate otherwise you'll have an adverse selection problem. The only thing new is his use of the good analogy of a making a two legged versus three legged stool. Must economists be so catty to each other?

David R. Henderson writes:

Boonton,
You might be right that the "finally" is unjustified. I thought it was but I could be wrong. Could you provide the cite where he admits that earlier?
Best,
David

Boonton writes:

Try http://krugman.blogs.nytimes.com/2009/08/01/health-reform-made-simple/

Everything else is about making that core work. Individual mandates are a way to prevent gaming of the system by people who don’t sign up until they’re sick; employer mandates a way to hold down the on-budget costs by preventing a rush by employers to drop insurance; the public option a way to create effective competition and hold costs down further.

David R. Henderson writes:

Thanks, Boonton.

Comments for this entry have been closed
Return to top