« Risk vs. uncertainty | Main | Tyler Cowen begs for hate mail »
Adverse selection is NOT the problem
The adverse selection story is a wonderful example of McCloskey's argument that great rhetoric persuades even when it shouldn't. The market for lemons is simple enough for your friends to understand but profound enough for them to be impressed at your learning, so it's a hard story not to tell!
The facts of the matter, however, are that adverse selection is not an important part of the market for automobiles (trucks), or of auto, life insurance or health insurance (on the latter see below).
One reason adverse selection may not be that important in practice is because buyers and sellers use testing and certification to remove the most important information asymmetries. You can buy a decent used car, for example just get it inspected or certified. Only if such adjustments are illegal, or in some other way not allowed, will adverse selection become important.
Second, the asymmetry may run in favor of the sellers. Do I really know more about my own life expectancy than an insurance firm that has access to sophisticated actuarial models? And, assuming that I do have extra information is it all that important? After all "the race is not to the swift, nor the battle to the strong, neither yet bread to the wise... but time and chance happen to them all." Or, more prosaically, the signal is near irrelevant when the noise to signal ratio is high.
Third, propitious selection can be more important than adverse selection. What sort of person buys a lot of life insurance? Is it people who expect to die soon? Or is it the sort of person who is so worried about not leaving their family in trouble that not only do they buy life insurance they also buckle their safety belt and eat healthy? The price of life insurance falls the more you buy so evidently insurance companies believe it is the latter.
Everyone talks about adverse selection in the market for health insurance but in fact non-group policies in these markets are not relatively expensive and not hard to get. The national average annual premium for reasonably generous coverage for a single person is just $2,268.
Sure, that's a lot of money but the point is that it's not a lot relative to what an employed person and their employer would pay for similar coverage in the group market. There is no evidence for an adverse-selection death spiral in the market for health insurance. That's not surprising because non-group health insurance is medically underwitten (i.e. medical inspections just like car inspections). Most people are accepted a few are not. Only in states that require insurance companies to accept all or most buyers are rates high relative to the group market (rates in New Jersey, an outlier, are almost three times as high as the national average.)
There are problems in the health insurance market, including a lack of long term insurance, job lock and the inequity of affordability, but adverse selection is not one of them.
Thanks to Bryan Caplan, Robin Hanson, Tyler Cowen, Tim Harford, and Ray Lehmann for discussion.
Addendum: Comments are open.
Posted by Alex Tabarrok on December 13, 2005 at 07:16 AM in Economics | Permalink
TrackBack
TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341c66b253ef00e5508355cc8833
Listed below are links to weblogs that reference Adverse selection is NOT the problem:
» Adverse news for adverse selection from Asymmetrical Information
Alex Tabarrok has a marvellous post up on adverse selection. As he puts it, the central insight of adverse selection ". . . is simple enough for your friends to understand but profound enough for them to be impressed at your learning. so it's a hard st... [Read More]
Tracked on Dec 13, 2005 12:51:44 PM
» Adverse news for adverse selection from Asymmetrical Information
Alex Tabarrok has a marvellous post up on adverse selection. As he puts it, the central insight of adverse selection ". . . is simple enough for your friends to understand but profound enough for them to be impressed at your learning. so it's a hard st... [Read More]
Tracked on Dec 17, 2005 9:57:54 AM
» Texas Plans To Really Screw You, Hardstyle. from The One-Handed Economist
I'm not much for local politics, or for updating my blog regularly, but a headline in today's Express-News made me angry. Superpowers notwithstanding, I shall blog. Texas is considering, as part of a tax-reform package, a measure that would charge... [Read More]
Tracked on Apr 7, 2006 3:44:38 PM
» Adverse selection is so the problem from Whimsley
Alex Tabarrok at Marginal Revolution claims that Adverse selection is NOT the problem. This is not surprising given his enthusiasm for Markets Everywhere. Most of the comments address the empirical aspects of one particular case, which is the insurance... [Read More]
Tracked on Apr 23, 2006 6:31:38 PM
» Adverse news for adverse selection from Asymmetrical Information
Alex Tabarrok has a marvellous post up on adverse selection. As he puts it, the central insight of adverse selection ". . . is simple enough for your friends to understand but profound enough for them to be impressed at your learning. so it's a hard st... [Read More]
Tracked on Apr 23, 2006 9:10:19 PM
Comments
"There is no evidence for an adverse-selection death spiral in the market for health insurance. That's not surprising because non-group health insurance is medically underwitten (i.e. medical inspections just like car inspections). Most people are accepted a few are not."
I'm not really sure what you're trying to show. It is correct that "adverse selection death spirals" are not a problem in the health insurance marketplace because, as you note, insurance companies carefully avoid insuring sick people to avoid going broke by paying their health care bills. But I don't think that when people complain about the problem of "adverse selection" in the health insurance marketplace they are concerned with the welfare of the insurance companies; rather, they are worried about the sick people who can't get coverage. I may be missing your point here so I'd appreciate any clarification you can offer.
Posted by: alkali at Dec 13, 2005 9:53:34 AM
Non-group health insurance for a single person may be relatively affordable in most states, but family coverage tends to be very expensive.
Posted by: Peter at Dec 13, 2005 10:02:49 AM
To answer alkali's question I am mostly speaking to economists who use the adverse selection argument to suggest that there is in a technical sense a "market failure" in health insurance markets. It's not a market failure, however, if bad cars don't sell for much. Nor is it a market failure if people who are born with medical problems have to pay a lot for health insurance. These are issues of distribution not efficiency. Issues of distribution are important, but we need to correctly understand the problem we are dealing with to have an appropriate solution.
Peter, the national average premium for family coverage is $4,424 - again not expensive relative to what you would pay in the group market.
Posted by: Alex Tabarrok at Dec 13, 2005 10:16:40 AM
You fail to understand the difference between individual and group coverage. The HIPAA regulations regarding preexisting conditions, the COBRA requirements, the lack of medical underwriting, the renewal clauses. These are incredibly valuable options and comparable coverage is not available in the individual market. Thus, individuals, who are deemed basically healthy (medical underwriting via tests) are required to pay (on average) the same as groups (no medical underwriting and valuable protections). Your own data thus points to a vast disparity in the two markets.
Posted by: elliottg at Dec 13, 2005 10:35:03 AM
My gut reaction is your on crack. I have never heard of anybody at anytime ever say that individual coverage is close to the cost of group coverage. I've looked into individual coverage during a period of self-employment, the cost was factors more than the $50 I pay now (my company is cheap, so I assume they don't pay more than 50, probably 0).
At any rate, if there was a problem, it wouldn't be markets, which I think is your point. Health insurance is probably the most regulated industry in the economy.
Posted by: cb at Dec 13, 2005 11:03:50 AM
Argh!! I just clicked through for your source and found it was the AHIP. I should have done that before I wasted bits responding at any level; serves me right. If you use BS sources then your arguments will sound like BS - a variation of the garbage in, garbage out maxim.
Posted by: elliottg at Dec 13, 2005 11:09:09 AM
[Everyone talks about adverse selection in the market for health insurance but in fact non-group policies in these markets are not relatively expensive and not hard to get. The national average annual premium for reasonably generous coverage for a single person is just $2,268.]
I have a real problem with people using statements about pricing as evidence one way or the other on adverse selection. It's meaningless. If you can price for a risk, it's not a case of asymmetric information. What we need to hear about is the excess or co-pay; how does this differ between pooled and individual policies? Also that average is a somewhat misleading number because it's the average over underwritten policies in cases when the denials rate was as high as 20%. If you cut out 20% of cases, of course the average is going to be lower; we would need to know the cost of providing insurance to the cases that were not underwritten to get a meainingful measure because not insuring a medical risk doesn't make it go away.
And Alex's "Second" above is a really bad argument about adverse selection. The overall signal/noise ratio may be low, but it will be higher in some cases than others and the high cases are the ones that create the problem.
Posted by: dsquared at Dec 13, 2005 11:25:01 AM
elliottg, can you explain why the AHIP data is garbage? And even if there are prices differences between group and individual premiums, couldn't this reflect the fact that the government subsidizes many kinds of group premiums?
Posted by: Javier at Dec 13, 2005 11:27:54 AM
Alex's point is a good one, but easy to misunderstand.
The main problem with the "health insurance" system is one of distribution,
not insurance. We want to be able to efficiently redistribute from healthy
people to sick people. Large group employer policies do this to some degree,
but in general the free market is not designed to provide redistribution.
Imagine we had technology to perfectly predict everyones future health. Then
there would be no hidden information, no risk, and no need for classic "insurance."
But society would still want a way to give resources to the sick people.
I wish Alex would tell us his opinion on why New Jersey has so much higher rates than states without community rating.
Posted by: ed at Dec 13, 2005 11:51:21 AM
I don't understand why everyone is so hostile to Alex's very good points. In insurance markets, I find that when people complain about adverse selection or "uninsurable risks", they are really demanding free insurance. Just because a premium is costly does not mean the market has failed. Plenty of allegedly uninsurable risks (such as flooding, hurricanes, sickness, etc.) can in fact be priced efficiently by the market. Consumers just balk in a childlike fashion when they are confronted with the true cost of their lifestyle decisions and/or desires. Most allegations of market failure in insurance are in fact demands for a subsidy. As Alex points out, this is a question relating to distributional fairness, not market failure. Too many people believe "affordable health insurance" means a right to unlimited free health care. Simply pretending that something shouldn't be costly doesn't mean that it won't be in the real world.
Posted by: Will at Dec 13, 2005 12:19:09 PM
I wish Alex would tell us his opinion on why New Jersey has so much higher rates than states without community rating.
Alex, the above quote points something out, albeit indirectly, adverse selection can be a problem when laws prohibit the market from working.
Take "community rating," for example. Depending on how fully implemented the system is, community rating forces insurers to sell to all people for the same price, regardless of health condition and other factors. Surely adverse selection is a problem in this case. People have an enormous incentive to buy more health insurance as their health worsens, and the insurance companies cannot price people differently. Since the insurance companies price insurance the same for everyone, and people can buy insurance at any time, it's always worth it for sick people to buy insurance, but not worth it for healthy people to subsidize the sick people. If you're healthy, it makes sense to wait until you get sick.
Adverse selection definitely happens, but it happens most strongly in situations where various consumer protection laws prevent sellers from acting on publically available information but buyers obviously can. There are situations where regulated markets are worse than both freer markets or socialized ones.
Posted by: John Thacker at Dec 13, 2005 1:05:33 PM
AHIP is simply a trade lobbying group. I have yet to read a report of theirs that does not support their lobbying positions. That would be fine except that most of the reports that I have read clear through have logical errors or facts that directly contradict other, more disinterested (in a financial sense) research. I admit that I did not read this particular report clear through since it's conclusions seemed intuitively wrong to me. Had it been someone other than AHIP, I would have read it to challenge my preconcieved notion, but experience with previous AHIP literature made me unwilling to go any further.
To Alex and Will, I fail to see how the "correct" pricing of insurance which leaves many unable to afford it and at risk for worse health outcomes (at a high cost to society)as anything but a market failure. You might argue that my view essentially dictates a universal health scheme, that's correct. I don't understand how wanting to stay healthy is a "lifestyle" choice nor do I believe that this desire is associated at all with "unlimited free health". Even for those which the market works (I have had medical insurance 100% of my life) I have never seen a plan that entitles people to unlimited care.
Finally, the evidence of the individual market and the group market charging similar rates (assuming AHIP did good research here) when the group plans have important, costly, demonstrably valuable options means that the individual market (which draws from a healthier pool of people once you consider medical underwriting) is suffering distortion.
If you wanted to show a lack of market failure then you could easily do a study for COBRA insured. Here the problem of adverse selection is obvious, weel-known, and priced into the group policy premium. This means that healthy individuals who want health insurance should be able to easily find cheaper healthcare in the individual market compared to the COBRA rate offered. The studies I've seen show that people who forego COBRA choose no health insurance. The number of people who voluntarily give up COBRA in favor on an individual policy is negligible. If there was no market failure then it should be obvious that healthy individuals would fare better on their own.
Posted by: elliottg at Dec 13, 2005 1:11:14 PM
Alex is right in diagnosing a common fallacy: the problem of adverse selection (i.e. the efficiency loss) in health insurance is that low-risk individuals get priced out of the market, not that high risks have to pay too much. Similarly, the problem in used cars is that people with good used cars hold on to them, rather than sell them, because they can't get a good price.
However, it is not useful simply to point to insurance rates for individuals (vs. group rates) as evidence that there is no adverse selection problem. Far more telling is the fact that so many individuals (young and in relatively good health) don't buy it, at the supposedly bargain price of $2,268. I don't know much about their utility functions (and I suspect Alex doesn't either), so my inclination is to regard this pattern of market preference as significant. The more plausible conclusion is that the group rate is too high for many as well, but individuals who get health insurance through their employer either have no choice but to participate, don't see the true cost of their policies, can't opt out with full reimbursement (esp. regarding tax treatment of the benefit), etc.
With this in mind, consider the following argument: The fact that any solution to the problem of the uninsured seems to involve massive cross-subsidization (as one finds in group plans) is itself an argument in favor of social insurance. At least then the cross-subsidization has a principled and consistent basis, rather than being ad hoc and morally arbitrary.
Finally, to get back to used cars for a second -- there have been a number of major initiatives in the past few decades, by manufacturers, dealers, and in consumer protection law, aimed precisely at correcting adverse selection problems in this market. The practice of offering warranties on used cars, which many dealers now do, is an example of a relatively recent innovation, aimed at correcting these problems. Note also that the price of used cars has steadily risen in the past decade as well (at least where I live). So one must be cautious in saying adverse selection is "not the problem" in this market. It's not the problem anymore, but only because the market has been tweaked and regulated in various ways to fix the problem. It would become the problem again very quickly if these kludges were removed. Furthermore, because the problems involve information asymmetries, things have the potential to change very quickly. The most recent problem I had with my car involved a software glitch. It was solved by bringing it into the dealer for a firmware upgrade. My suspicion is that these sorts of problems would probably go undetected in a traditional car inspection.
Posted by: Joseph Heath at Dec 13, 2005 1:20:49 PM
The practice of offering warranties on used cars, which many dealers now do, is an example of a relatively recent innovation, aimed at correcting these problems.
Yes, and it definitely can work. Note, however, that if "consumer protection" laws force used car dealers to offer the same warranties on all used cars for the same price, adverse selection can still occur.
Posted by: John Thacker at Dec 13, 2005 1:23:45 PM
Isn't the fact that insurance companies (and credit institutions etc...) have put in place tools to reduce information asymetry a proof of the existence of the adverse selection?
These tools allow insurers or creditors the ability to better gage a person's risk, and thus better price their offer. Its a cost the insurers have to bear because there is not much one can do to convincingly signal one's true risk level. In addition, the incentives in this respect work against the insurer: the "inseree" will try to present the best profile possible.
Posted by: Philippe at Dec 13, 2005 3:32:23 PM
A comment to Ed above. Insurance in states with community rating (e.g., New Jersey and New York) costs more than states without it (e.g., California and Tennessee) because insurers are unable to price the risks they underwrite appropriately. NJ and NY also have guaranteed issue, which means they are forced to accept all applicants and can't exclude unacceptable risks.
As for your comment about redistribution of resources to sick people, with the exception of charity, "we" don't do this, the state does--by using coercion.
Posted by: Bill Stepp at Dec 13, 2005 3:59:53 PM
"I don't understand how wanting to stay healthy is a "lifestyle" choice..."
Because hardly anyone does even the basic things that are generally considered to be the components of a healthy lifestyle. Such things include regular exercise, maintaining a healthy weight, eating mostly healthy food, and avoiding constant stress. A study I read at
http://www.healthfinder.gov/news/newsstory.asp?docID=525339
estimated only 3% of the US population practices such healthy habits routinely.
Wanting to stay healthy and actually doing what it takes are not the same thing.
Posted by: Unknown at Dec 13, 2005 4:48:11 PM
Here's a data point, for what it is worth.
I am self-employed and have an HSA with a 2,500 deductible and a one million dollar lifetime cap on coverage. This costs me $85 a month or $1020 a year - well below the $2268 average cited. I'm healthy and 37 and I live in Iowa. If I were employed I'd want the same thing instead of the ridiculous $300-400 per month my past employers have negotiated (of which I paid $20-$80 of that) with a group plan provider that provided 100% coverage except for little co-pays. I'd rather have the extra $200-300 in my pocket for other things.
Posted by: Unknown at Dec 13, 2005 4:52:54 PM
I hope you stay well unknown, but that, to me, sounds like pretty lousy coverage at a relatively high price and that's without knowing your copays and drug coverage. Also, Iowa health costs are less than the national average so you benefit from that as well.
Posted by: elliottg at Dec 13, 2005 5:18:51 PM
It's major medical coverage elliotq - there are no copays. I pay for everything below the annual deductible. It makes perfect sense for me. $1020 a year (which is tax deductible) vs. $4800 a year that I and my prior employers paid for my co-pay only coverage. Which coverage is lousy?! I pay for all medicine out of pocket. Since I only consume allergy medicine a few times a year, I don't spend more than about $200 on pharms which I buy with tax-free money from my HSA.
Since it is an HSA ( so I can put $2500 a year into a tax sheltered account and use it for any medical costs. If I don't spend the money on medical costs, I can use it (but taxed) for anything else including letting it grow just like it was in an IRA. So I have a strong financial incentive to stay as healthy as I can.
Posted by: Unknown at Dec 13, 2005 5:29:26 PM
Remember, just because you gamble and win, doesn't mean it was a good bet.
Posted by: ellliottg at Dec 13, 2005 5:32:16 PM
What ARE you talking about?! Do you understand the benefits of the HSA? Is your standard assumption that everyone is going to incur catastropic medical costs year in and year out? Most people do not and that's what makes the super-care insurance that is standard at most companies overkill.
Posted by: Unknown at Dec 13, 2005 5:34:11 PM
Do you also understand that the $4800 a year (to continue my case) for coverage (individual coverage btw) that employers pay comes out of the employee's compensation? The marginal difference between the cost of an HSA and the cost of standard big company insurance is wasted money that could be put to more productive use elsewhere.
Posted by: Unknown at Dec 13, 2005 5:36:38 PM
The absolute first thing I did after being diagnosed with lymphatic cancer in 1996 was drive to my place of employment, go to the Human Resources department, and increase my employer-provided life insurance to the maximum allowed.
Posted by: Steve Sailer at Dec 13, 2005 8:15:00 PM
To me Alex's argument fails. The low figure cited for individual insurance is for a LIMITED TIME, usually one year. After which you policy is reevaluated and your premiums skyrocket. This doesn't happen in a large group policy where there is still always a certain percentage of healthy people to share the costs.
Posted by: David Y. at Dec 13, 2005 9:28:17 PM
