Insurance is good when one gets hit by a meteor.
The Health Care Cost Institute (HCCI) team uses their impressive data resources to make the point that a lot of people are hit by a meteor every year:
Substantial turnover among top 5% of healthcare spenders: 3 out of 5 in any year were not top spenders in prior year: https://t.co/xTM4TJP4Re #HCCIdata @WJohnsonomics @N_Brennan @IamSRodriguez @HealthCostInst pic.twitter.com/UPLcd2JSlt
— NEJM Catalyst (@nejmcatalyst) February 1, 2018
Roughly 40% of the Top 5% of the health care spending distribution in their data set showed up as part of the Top 5% in both 2014 and 2015. Here the optimization problem for health insurance is not the premium but a constrained minimization problem of total costs which is premiums plus out of pocket maximum spending.
The more interesting and insurancy part of the graph is the 60% of the high spenders in 2014 drop out of the Top 5% the following year and they are replaced. Their following year claims costs are not particularly high ($4,942). The people who replace them have slightly lower but similar claims costs in 2014 before getting hit by a medical meteor in 2015.
That is the value of insurance for most people. Meteors will hit.