How adverse selection destroys insurance markets

I have not seen an adequate explanation of adverse selection, a key issue in the debate over Obamacare, so I am offering one here. Consider a simple scenario: one insurance company, Acme, with three potential customers. Alice, who is very healthy, has only $30 in medical expenses in a typical year. (I’m making the numbers small in the interest of simplicity; if it makes you feel better, add some extra zeroes to every number.) Bob, who is of average health, has $90 in medical expenses in a typical year. Caroline, who was born with a genetic defect, has $300 in medical expenses in a typical year. To account for the unpredictability of medical expenses, here’s one additional twist: one of these unlucky souls will have an accident that leads to an additional $210 in medical bills.

While each person knows their typical medical expenses, and knows they face a one-third chance of an additional $210 in expenses, all Acme insurance knows is that the total costs for all three are $630 ($30 + $90 + $300 + $210). That’s $210 per person; Acme decides to charge $220 per person (the extra $10 goes for administrative costs) for a policy that covers all medical costs. Is this a smart decision?

No, it isn’t. Take a look at the financial landscape from each person’s perspective:

	Minimum possible costs	Maximum possible costs	Insurance premium
Alice	$30	$240	$220
Bob	$90	$300	$220
Caroline	$300	$510	$220

Insurance looks like a fairly good deal for Bob and Caroline, but Alice is skeptical. Insurance would be a whopping $190 higher than her minimum expenses, and $120 higher than her expected costs of $100 ($30 plus one-third of $210). In the worst scenario, she only saves $20. Alice decides not to get insurance. This is the tragedy of the insurance business: the most profitable customers have the least motivation to buy insurance.

With Bob and Caroline as customers, Acme collects $440 ($220 from each), but can expect to pay out a total of $530 ($90 + $300 + two-thirds of $210), for a loss of $90.

Acme must set a higher price for insurance. If it knew Bob’s and Caroline’s costs, it could, for argument, set a price of $270 per person, 22% higher than before. Take a look at the adjusted financial landscape:

	Minimum possible costs	Maximum possible costs	Insurance premium
Alice	$30	$240	$270
Bob	$90	$300	$270
Caroline	$300	$510	$270

It’s an even worse deal for Alice, but now Bob is also skeptical. Insurance would be $180 higher than his minimum expenses, and is $110 higher than his expected costs of $160 ($90 plus one-third of $210). In the worst scenario, he only saves $30.

Bob decides to skip insurance too, and Acme can expect to lose $100 on Caroline’s policy. What’s happening here is called the adverse selection death spiral. For all intents and purposes the market for insurance has disappeared, and Caroline is left to deal on her own with the financial consequences of her genetic defect. Absent external forces (such as subsidies or government regulation) Adam Smith’s invisible hand cannot establish a price for insurance at which the market will clear. The way this market is defined, Acme cannot offer insurance at any price.