What is adverse selection?

Adverse selection occurs before a deal, when hidden differences between customers draw in the riskiest ones. It matters because a price based on the average customer attracts worse-than-average ones and can leave a market with only the highest-risk participants.

How it works

An insurer sets a price based on the average risk of its customers. Low-risk customers find the price too high and drop out or never join, while high-risk customers find it a bargain. The group that stays is riskier than the average the insurer had assumed, so claims exceed the price.

The insurer raises the price to cover the higher claims, which drives out the next lowest-risk group, and the cycle repeats. This feedback can end with a market for only the worst risks. Remedies are to price by observable risk, require everyone to join, or gather more information before the deal.

Example

Say a plan covers 1,000 people, half with an expected yearly cost of 200 and half with an expected cost of 800. The average is 500, so the premium is set at 500. The low-risk group pays 500 for a cost of 200 and many leave, say 400 of the 500. Now the group is 100 low-risk and 500 high-risk, with an average cost of about 700. The premium must rise to 700, and more of the low-risk group leave. The figures are an illustration.

How to model it in Machinations

Create Pools for the low-risk and high-risk members, a Register for the premium calculated as the average claim cost of current members, and a Drain from each Pool whose rate depends on the premium through a state connection. Add a Pool for the insurer's balance. Play for several steps and chart membership and premium to see the spiral, then add a Lever for a required membership and compare.