From “Distributive Injustice(s) in American Health Care, ” by Clark Havighurst and Barak Richman:
[I]t should certainly be a cause for concern if consumption patterns vary greatly and positively with income—rather than with health needs alone—in situations where everyone pays the same premium for the same health coverage. This appears[*] to be the case in many U.S. health plans, since higher-income employees seem to make greater use of their coverage, demanding and receiving more and costlier services at plan expense than their lower-income coworkers. […] [T]he tax subsidy is the ultimate source of the problem. By causing health coverage to be purchased in heterogeneous employment groups (including individuals with disparate, income-correlated preferences and consumption patterns), it creates conditions in which lower-income premium payers may be paying—unwittingly—costs incurred by their more demanding, affluent, and influential coworkers. […]
Our concern, however, is not that health care is rationed or distributed unequally but the likelihood that conditioning eligibility for insurer payments on patients’ willingness to make certain out-of-pocket payments causes lower income participants in employee health plans to get disproportionately fewer benefits than their more affluent coworkers receive in return for equivalent premiums. […] Likewise, as employers pursue the increasingly popular strategy of funding health savings accounts and enrolling their workers in high-deductible health plans, it is possible that greater emphasis on cost sharing to contain moral hazard will cause insurers’ premium pools to be allocated even more disproportionately to the care of the affluent. […]
[A]n economist might suggest that employers unconsciously adjust the amount of wages they are willing to pay to different classes of worker to reflect the class’s propensity to utilize employer-financed health benefits—in which case it might be incorrect to hypothesize that lower-income workers actually bear costs incurred by higher-income, higher-utilizing participants in the same plan. […] [But, t]he notion that there is no regressivity depends on heroic assumptions about employee and employer perceptions, rationality, and the smoothness of the market’s operation. Thus, workers’ decisions about which jobs to take turn on many factors besides the implicit value of particular health coverage. Furthermore, employers probably think only rarely in terms of total compensation packages, perhaps even administering employee benefits and cash compensation in separate cost centers.
* The authors admit that they are unaware of a lot of direct evidence this happens.
h/t Nicholas Bagley
