Screening: Sorting Without Seeing

You wake up on a Monday morning in October 1973 to find a memo on your desk. You have just been promoted to head of underwriting at Helvetia Versicherung, a mid-sized insurance firm headquartered in Zurich. Your predecessor retired abruptly. The task left on your desk is straightforward on paper and terrifying in practice: design a single health insurance contract for the coming year's open enrollment period.

You know almost nothing about the people who will buy this contract. Some are 28-year-old cyclists who have not visited a doctor in three years. Some are 54-year-old smokers managing early-stage diabetes. Some are managing chronic kidney disease. Some are, by any actuarial measure, extraordinarily cheap to cover. Others will cost the firm ten or twenty times what they pay in premiums.

You cannot ask them which group they belong to. Privacy laws constrain you. Social norms constrain you. And even if you could ask, you know they would lie, or at least shade the truth in their favor. The person filling out your enrollment form has every incentive to present themselves as healthier than they are. You are, in the most literal sense, offering a contract to a room full of strangers whose faces you cannot see.

So what do you charge?

If you price the contract based on the average cost of covering everyone in the pool, you arrive at a number that feels reasonable in aggregate. But here is the problem you cannot shake: the healthy 28-year-old looks at your premium and does the math. She is paying far more than her expected medical costs. She walks away. The 54-year-old with diabetes does the same math and realizes he is getting a remarkable deal. He signs immediately.

This is the moment the floor starts to tilt beneath you. Your pool of enrollees is no longer a representative cross-section of the population. It is skewed, invisibly and automatically, toward the people who cost the most to cover. Your average-cost pricing has attracted exactly the wrong average.

This problem has a name. Economists call it adverse selection: the tendency for asymmetric information to cause markets to attract disproportionately the participants who are most costly or risky for the uninformed party. The word "adverse" is doing real work here. The selection is not random. It is systematically, predictably bad for you.

You are not dealing with bad luck. You are dealing with a structural feature of any market where one side knows more than the other. And in 1973 Zurich, sitting at your new desk, you have no obvious way out.