Adverse Selection and Risk Aversion

Adverse Selection is often cited by proponents of socialized health care as a reason why private health insurance cannot work. Although the people who do so usually use the term incorrectly and get its implications backwards, as David Friedman catches Austan Goolsbee doing here, it is true that adverse selection does pose a problem for insurance markets (see Friedman's linked post for an explanation if you're not familiar with it).

This problem is not specific to health insurance. Other types of insurance such as life and fire insurance are also prone to adverse selection, yet we have functional markets in these. Alex Tabarrok gave several possible reasons for this here.

One factor I haven't seen addressed elsewhere is risk aversion. Even though buying health insurance is a negative-expected-value proposition for me, it still has positive utility because of risk aversion. That is, I prefer a certain loss of $2000 per year in premiums to a small but nonzero risk of losing my life's savings due to a catastrophic medical expense, even though the former option will cost me more on average.

So while adverse selection does create a disincentive for me to purchase insurance, risk aversion negates this effect by creating an incentive for me to buy insurance even when its expected value is negative.

Share this

Adverse selection is great

Adverse selection is great when you're the adverse selector: it gives a warm fuzzy feeling... like stepping in an all-you-can-eat buffet with an empty stomach. 

Health Coverage Land Mines

Health insurance in the US is a strange bird because for the most part it is supposed to be paid for by employers and the government. If you can get employee’s insurance, you can get health insurance no matter how sick you are (there are waiting periods for preexisting conditions), if you are well enough to work. Many people work for certain companies just so they can have health insurance for themselves or their dependents. Another way to get health insurance is to be older than 64 and have Medicare. A third way is to be entirely destitute and get Medicaid. A fourth way if you have some special diseases such as needing dialysis, the government pays for it. A fifth way to get medical care is to have no insurance or money and show up at an emergency department acutely ill. All the payors bargain fiercely with doctors and hospitals to pay less then the usual and customary fees. Providers still maintain usual and customary fees for these services. You will notice I have left out those who purchase their own policies and those who pay cash or just set up a monthly payment scheme with the hospital. Those who can purchase their own policy can actually get coverage cheaper than it costs the employer who has to buy policies that cover all risk. The only trouble is the private policies exclude many people who have even slight risk factors or preexisting conditions. The last option is to pay cash. This seemingly sensible option is a serious trap for the unwary, if you have any assets. This is because there is a vast difference between the usual and customary charge (list price) and the payment providers will accept from the insurers or the government. For hospital charges the list price is usually three to five times the Medicare allowable. If you ever have a chance to examine a hospital bill and compare it to what the hospital actually gets you will see what I mean. What if you are uninsured, have assets and can’t get insurance? Don’t even go near a hospital. The above information concerning the interaction between providers and payors especially the amazing discounts providers give to insurers is seldom taken into account in these discussions. Dave

Adverse Selection

Yes, what hospitals bill the uninsured far exceeds what they pay private insurers, medicare and medicaid for the same treatment.

However, as an uninsured customer who uses the emergency room for your care, you have access to the medicare allowances for any treatment listed by CPT code, though you might have to file a FOIA to get it the list from the state agency.

Then you can ask the hospital, doc, lab, etc., that provided you the service to provide the CPT code for each item billed.

Then you can demand that they charge you no more than they charge medicare, etc, under the theory of "quantum meruit [as much as is deserved]," which allows them fair payment for their services. What is fair is determined ultimately in court, where you can present evidence that they charge similarly situated persons much less than they are trying to charge you; i.e., they don't deserve what they are billing you.

The result is that the hospital, doc, lab, etc, will write off all the charges, or at least deeply discount them, since they feel it would be embarrassing and severely precedent-setting for them to have to lose to you to whom they are giving a now public royal ream-job.

I have done this twice for friends, who were able to have bills of $1700 and $1800 entirely written off!

Inflated Hospital Fees

Yes, I believe there has been successful litigation to force the outrageous fees down. The last time I checked the local hospital was giving a 40% discount for cash payments for hospital care. This is nowhere near the discount insurance companies or Medicare get, however. Another tactic if you have an elective procedure is to talk to hospital administration and get bids for the procedures from various hospitals. They would really be tickled to get a mere portion of the vastly inflated list prices. The question is, why do they inflate these fees so much to begin with?


No competition

The question is, why do they inflate these fees so much to begin with?

To be quick, I bet it's an almost complete lack of price competition.

Maybe another factor behind

Maybe another factor behind excessive fee inflation is that hospitals know how HMOs and Medicare/Medicaid representatives make their purchasing decisions; i.e. if I am a hospital price setter and I know that HMO/gov't bureaucrats generally follow an easily discoverable algorithm that determines whether or not a given medical procedure is a "good deal", or whether they should bargain harder or look elsewhere, then I, as a hospital price setter, will raise my list price accordingly, knowing that I can then take, say, 50% off the list price and the bureaucrat will likely accept the deal. I notice this sort of thing happening at clothing stores like Macy's all the time.


i.e., what appears to be price competition ultimately isn't because hospitals can game the system (since the system isn't really a market).

Btw, my experience with dental care is much different. Dentists have been very willing to cut deals for people who aren't insured and are willing to pay in full rather than on some form of credit.

Hospital fees

A couple of reasons why hospital fees are so high:
1. As Dave alluded to, insurance carriers have PPO arrangements with hospitals and doctors, wherein the care provider is willing to offer services at a discount to the carrier in exchange for being in the insurer's network (thus presumably getting more business). Over time, though, these discounts drift. If a procedure costs $100, and the insurer negotiates a 20% discount, some providers will increase the cost of the procedure to $120 so that they are still getting roughly the same payment. Since most insurers' payments are based on "reasonable and customary" charges, if enough providers do this, this increased price becomes the norm. This causes the insurer or PPO network to negotiate steeper discounts, and the cycle repeats.
2. The threat of litigation can have a large effect on the prices of services. Medical waivers that are intended to protect the hospital against adverse (but not negligent) outcomes may not be honored by the patient. More services are ordered by the attendant physician in order to protect against increasingly unlikely scenarios because if there is something that isn't checked, it opens the doctor and the hospital up to litigation. Additional procedures add to the total bill. The individual procedures themselves must also be marked up so that the doctor or hospital can afford to pay the malpractice insurance necessary to stay in business.

Regarding risk aversion

As the actuarial exams taught me, risk aversion is the reason why insurance is a viable product. Assuming you have a utility curve with negative convexity and an initial wealth position W, let's suppose you have a loss X, where X is variable. The utility of the initial position minus the expected loss is going to be greater than the expected utility of the initial position minus the variable loss (so, U(W - E[X]) > E[U(W - X)]). In other words, you're better off taking the certain loss at the expected value than you are taking your chances with the loss.
For people with greater risk aversion, the utility curve is more negatively convex (i.e., steeper), and they get a greater relative benefit from insurance. If the initial wealth position is high, the utility curve is going to be flatter, and there is a lower relative benefit from insurance. The greater the expected loss and the greater the variability of the loss, the greater the benefit of insurance.

You've got it backwards, 

You've got it backwards, people are risk averse, and a utility function is one way to show it.

( And more concave doesn't mean steeper... )

( And really there are some utility curve representing a person's preference, and a subset of those can represent preference for lottery as a probability weighted linear combination of utility for certain outcomes. )

You've got it backwards,

You've got it backwards, people are risk averse, and a utility function is one way to show it.

I don't understand your objection. No one was claiming that utility functions exist so therefore people are risk averse. Crash asked us to assume that a person has a certain shape of utility curve, which is the same thing as asking us to assume the person is risk averse.

The utility curve is the map, the risk aversion is the territory represented by the map. Saying "assume a map has features X and Y" is not claiming that the territory in question has features X and Y because the map has the same features. I don't understand why you would think Crash, or anyone else, believes that.

Crash ends up demonstrating

Crash ends up demonstrating that we may be willing to pay an amount greater than an expected cost to avoid it - that is the definition of risk aversion. He uses a utility curve to demonstrate it. He has it backwards because the utility curve was merely constructed to reflect this observation.

Yes, shortly after I posted

Yes, shortly after I posted my comment, I realized that more negatively convex does not equal steeper.

As for my use of utility curves, I was just trying to make a comment about the usefulness of insurance as a product. Depending on the individual's preferences and wealth level, and the amount and variability of the loss, insurance may or may not be a beneficial product.

Again, I don't understand

Again, I don't understand your objection. Using a utility curve to demonstrate why risk aversion makes insurance rational is in no way incompatable with your claim that "the utility curve was merely constructed to reflect this observation." Are you claiming that Crash or anyone else believes (or communicated) that people can be risk averse solely because of manipulations to a utility curve? I've seen nothing in Crash's post or anywhere else to support that notion.

The map does not determine the territory. And no one believes that it does. People who don't like maps--Austrians, for example, though I am not claiming here that you are an Austrian--often accuse map users of erroneously believing maps determine or create the territory. But this is ridiculous - no one believes this. If you don't like maps, just say so, but don't accuse map users of confusing maps with the the territory they are designed to represent unless you have actual evidence of this confusion.

Are you claiming that

Are you claiming that Crash or anyone else believes (or communicated) that people can be risk averse solely because of manipulations to a utility curve.


The map does not determine the territory. And no one believes that it
does. People who don't like maps--Austrians, for example, though I am
not claiming here that you are an Austrian--often accuse map users of
erroneously believing maps determine or create the territory. But this
is ridiculous - no one believes this.

Yup. But it's not about that.

What's your definition of risk aversion ? I define risk aversion as the fact that one is willing to pay more than an expected cost to avoid it. We can represent that with concave utility curve. What's the point of saying a concave utility curve produces risk aversion? Checking that 'it works' ?