Wednesday, February 17, 2010

Self-Promotion: Pro and Con on the Individual Healthcare Mandate

I was featured in Monday's L.A. Times "pro and con" article on the individual healthcare mandate. Here's a key bit:
The point of the individual mandate is to balance the risk pool, but that's not really what insurance is supposed to do. With car insurance, the idea is not that you want good drivers to pay for accidents caused by bad drivers. Instead, you want there to be pools of people with similar risk: bad drivers and good drivers who pay different premiums.
I notice that a couple of commenters have taken issue with this argument, essentially saying "OF COURSE insurance is supposed to balance risk!" They're mistaken. But this is a point worth dwelling on, because it's easy to get confused.

Everyone knows that insurance does "balance risk" in a certain sense. It transfers wealth from those who have been fortunate (e.g., didn't get in a car accident, didn't have a house fire) to those less fortunate (e.g., did get in a car accident, did have a house fire). The premiums paid by everyone get paid out to a small group ex post. But ex ante, every policyholder could have experienced either outcome. To make this worthwhile for all buyers, they have to be pooled with other buyers who have similar risk. Otherwise, some buyers will find that it's just not worth it: their risk is too low to justify the high premiums.

This is very different from trying to "balance risk" in the sense of balancing low-risk policyholders with high-risk policyholders and charging them similar premiums. This approach creates a cross-subsidy from some policyholders to others ex ante. That is, even before we know which people will experience an unfortunate outcome, some policyholders are getting a much better deal than others. That's why some buyers drop out of the insurance pool -- because they're getting a lousy deal. Hence the demand for the individual mandate: it's needed to rope in the people who would rather opt out.

As a matter of terminology, the former should be called "risk pooling" and the latter "risk balancing"; using the same term for both just creates confusion. And as a matter of historical and economic fact, the insurance business arose primarily to do risk pooling. The idea that insurance should engage in risk-balancing is almost exclusive to healthcare, and it's assuredly the result of decades of government efforts to force insurance to accomplish policy goals that were no part of its origin.


Brandon Bertelsen said...

Switzerland has a system where everyone is forced to pay for mandatory insurance. ( Balancing risk, risk pooling - whatever you want to call it - the net effect is to make healthcare less expensive for everyone as it only represents 10.8% of GDP in comparison to almost 16% for USA. (or may that litmus test of cost is a bit to broad)

jimbino said...

I would be helpful to point out that, whether in risk pooling or risk balancing, it is stupid for the average non-risk-averse person to participate. The rational risk-averse person would only willingly participate if he held private information of his higher-than average risk of loss.

While the payout for roulette is around 96%, the payout for flood insurance is around 60%, for health and car insurance around 50%, and for title insurance around 2%, making insurance one of the worst games of chance the average person could participate in.

Glen Whitman said...

Jimbino -- I don't think that's exactly right.

You're correct that non-risk-averse people won't participate -- but true risk lovers are actually pretty darn rare, so it's not clear how relevant this is. The overwhelming majority of people are risk averse.

Then you say, "The rational risk-averse person would only willingly participate if he held private information of his higher-than average risk of loss." This is not necessarily true. The essence of risk aversion is a willingness to accept a lower expected value (that is, pay a premium) in return for the avoidance of risk. So, for instance, suppose you have a population of identical (risk-averse) people facing an identical risk. It would be perfectly possible to construct an insurance policy that all of these people would accept, even if the insurance policy's price were higher than the expected loss.

You are correct, however, that some risk-averse people will drop out of the insurance pool if their risk-aversion is mild enough. And as the insurance premium rises relative to the true expected loss, more and more risk averse people will drop out.

jimbino said...

And nobody takes into consideration the problem of Amerikan expatriates and even tourists, who would presumably have to carry useless insurance while they are traveling or residing in a foreign country, where not even Amerikan drugs can be bought or imported with Amerikan insurance proceeds.