Abstract
The distribution of health care payments to insurance plans has substantialconsequences for social policy. Risk adjustment formulas predict spending inhealth insurance markets in order to provide fair benefits and health carecoverage for all enrollees, regardless of their health status. Unfortunately,current risk adjustment formulas are known to undercompensate payments tohealth insurers for specific groups of enrollees (by underpredicting theirspending). Much of the existing algorithmic fairness literature for groupfairness to date has focused on classifiers and binary outcomes. To improverisk adjustment formulas for undercompensated groups, we expand on conceptsfrom the statistics, computer science, and health economics literature todevelop new fair regression methods for continuous outcomes by buildingfairness considerations directly into the objective function. We additionallypropose a novel measure of fairness while asserting that a suite of metrics isnecessary in order to evaluate risk adjustment formulas more fully. Our dataapplication using the IBM MarketScan Research Databases and simulation studiesdemonstrate that these new fair regression methods may lead to massiveimprovements in group fairness with only small reductions in overall fit.