Adverse Selection: Why Malpractice Insurance Costs So Much
Today?s topic is adverse selection: an economic market failure in which a party gets exactly the wrong outcome as they intended to get. The reason they get the wrong outcome is because of the problem of asymmetric information, which you can read about here. Adverse selection has many important policy implications, and we are going to explore the topic through a very simplified model of medical malpractice insurance. This model is going to show that rates are driven up for doctors, not because of unscrupulous lawyers, money-grubbing HMOs, or any fault of the system. Rather, costs are increased because under asymmetric information, adverse selection occurs and causes the market to perform more poorly than it would if equal information were present in the market. Let?s begin with an example.
Suppose there are two doctors in the world, Bryan and Raul. Bryan is a good doctor, and, therefore has a low probability of making a mistake and getting sued. Raul is a poor physician, and has a higher probability of being sued. Suppose they both choose to purchase malpractice insurance against the chance that they will be sued. Suppose, the insurance companies know that Bryan is good and Raul is bad. In that case, the insurance company will offer a rate to each that reflects their probability of being sued. If Raul has a 40% chance of being sued in a year, then he will have to pay $0.40 per dollar of coverage he buys. Bryan, on the other hand, has only a 20% chance of being sued in a year, he must, therefore, pay $0.20 per dollar of coverage?half of the rate Raul pays. In this market, both doctors are offered actuarially fair insurance, and they both purchase full insurance that exactly offsets the amount they would expect to lose from a lawsuit. In addition, because the insurance company offers separate premiums to Bryan and Raul that exactly match their risk (and, therefore, expected payout), they are able to break even: they get $0.60 in premiums, and, on average, pay out $0.60 cents per year.
Let us now introduce asymmetric information into the model. Suppose now that Bryan knows he is a good doctor, and Raul knows he is a bad doctor. On the other hand, the insurance company does not know who is a good doctor and who is a bad doctor. All the insurance company knows is that, on average, doctors are likely to be sued 30% of the time: 1/2*40%+1/2*20%=30%. Remember, from our discussion on asymmetric information, what is likely to happen in this case. First, look at Bryan, the good doctor. He knows that his fair insurance price is $0.20 on the dollar, but in this case, he will have to pay $0.30 on the dollar. As a consequence, he buys less than full insurance depending on his risk aversion. The point is, not only does Bryan see high insurance premiums, he buys less than full insurance because at those prices, it?s not worth it. The story does not end here. Look at Raul; he sees insurance as cheap, he should be paying $0.40 for fair insurance, and he pays only $.30. In this case, he would like to buy more than full insurance. Most, insurance companies do not allow policies that payout more than you lose, so he is capped at relatively cheap (to him) full insurance. Look, however, what has happened to the insurance company. They get $0.30 in premiums from Raul, and less than $0.30 in premiums from Bryan. The expected pay out is however, larger than $0.60 per dollar now, because bad doctors are purchasing more insurance than good doctors.
The cycle does not stop here, either. Suppose that instead of two doctors, there are lots of doctors at all sorts of levels of ability?and likelihood of getting sued. Since the insurer is now offering better than fair insurance, they raise rates to allow them to insurance the break-even rate. But as they do this, the best doctors see this is a poor value and buy less insurance, and the worse doctors see this as a good value and buy more. The insurance company potentially loses more money and raises rates to get back to the fair rate. This cycle could go on until bad doctors have driven good doctors from the market and insurance rates are sky high. In an extreme example, insurance does not exist at all because the only rate is one that is too expensive for anyone, even if everyone actually wants to buy insurance and would be profitable to insure in the equal information model.
Obviously, this does not happen in such an extreme fashion. Tools such as deductibles, and the ability of the insurance company to determine some level of differing risk between doctors (often predicated on specialty or past lawsuit history), allow insurance companies to offer different rates to different risks. This story does occur in less extreme examples, however. Highly regulated states often experience higher malpractice premiums. In these states, it is often difficult to find doctors of certain high-risk specialties, because, in order to practice in those states, doctors require higher wages. These higher wages, however, are prevented by regulation. The result is that for some specialties, the general level of talent in the physician market is lower.
This is one side to a very complicated story, insofar as it relates to healthcare and malpractice premiums. The basic story, however, holds true: insurance premiums can rise without any frivolous lawsuits, and certainly without over zealous lawyers just trying to make a buck. In fact all that our story required was that there was a potential for lawsuits, and that there were different levels of ability among physicians?both realistic assumptions. One final note: the story here is not enough information is not available, but that unequal information is available. I urge you to work out the situation where nobody knows their actual probability of being sued, just the average probability of being sued. You should arrive at the same result as the full information model.
April 11, 2004 |
Posted in: Economics |
Author: Charles | 
Print
One Response
How to Solve the Health Care Crisis - UtilityMinimization - PubPolicy.com: Thoughts on Policy and Economics - June 6, 2008
[...] we talked about adverse selection and asymmetric information in terms of insurance markets. The main concept is that nobody knows [...]
Please continue discussion on the forum: link
Leave a Reply
You must be logged in to post a comment.