If you were to ask 100 randomly-selected health wonks, probably 92 would say that the tax exclusion for employer-sponsored insurance is bad policy. This is one of those issues on which liberal and conservative health economists agree. Liberal and conservative politicians agree with each other, too, though not in the way hoped for by economists. Policies to limit ESI tax exclusion seem politically radioactive, as indicated by the savage reception to candidate John McCain’s 2008 health proposals, and as indicated by the delayed implementation of the “Cadillac tax” on high-value employer plans enacted under ACA.
Politics aside, most experts would argue that the tax exclusion is regressive, that it subsidizes over-insurance for highly-compensated people, and that such hidden subsidies make health care costs further opaque, and therefore undermine cost-control policies. TIE’s own Austin Frakt observes that many of us receive greater tax subsidies from the ESI exclusion than is provided to the typical Medicaid recipient. My own taxes would be about $7,000 higher if I were fully taxed on my generous insurance plan. This massive tax subsidy to a tenured professor seems hard to justify on either equity or efficiency grounds.
Joseph White of Case Western challenges the conventional policy-wonk wisdom in a recent essay in the Journal of Health Politics, Policy, and Law: The Tax Exclusion for Employer-Sponsored Insurance Is Not Regressive—But What Is It?
White makes several seemingly-obvious but generally-overlooked points. Although the dollar value of the tax exclusion is higher for high-income people who pay higher marginal tax rates, health insurance comprises a much higher percentage of income for low-wage workers. This is particularly so given the mundane reality that health benefits are provided in more egalitarian fashion across (full-time) workers than dollar-wages and salaries are. (My own employer, like Professor White’s, actually charges lower premiums to less-compensated employees.) So the tax exclusion is one mechanism to transfer resources, albeit imperfectly towards people with families and others in greater need of health services. If we believe that age discrimination is a significant problem in the American economy, tax subsidies for older workers aren’t the worst thing, either.
One strand of White’s argument becomes obvious with a simple example. Suppose GE provides $10,000 coverage to an executive earning $300,000 annually, and a mid-level worker who earns only $40,000. It’s quite possible that the ESI exclusion is worth $3800 to the executive and only $1500 to her less affluent counterpart. Does this mean that the exclusion is regressive? It’s not so obvious. After all, $1500 is a much larger proportion of $40,000 (3.75%) than $3800 is of $300,000 (1.27%).
For historical reasons, the United States has followed the problematic path of employer-sponsored health coverage for most employed citizens. It’s hard to imagine that we would institute this package of policies—including the ESI exclusion—if we could truly start anew. But we’re not starting anew. As Paul Starr and others rightly noted, providing health coverage through large employers brought genuine advantages along the way: Providing administrative economies and a reasonable internal risk-pool, providing some implicit redistribution to lower-wage workers, too. The challenges facing ACA’s federal and state marketplaces underscore the political and operational difficulties of standing-up viable alternatives to ESI, which would necessarily involve substantial government regulations and subsidies.
I’m not fully sold by White’s argument. Surely there is some way to limit excessive tax expenditures to families like my own. We don’t need the subsidy for these extremely generous plans. And it does bother me that the huge tax subsidy is so hidden, allowing prosperous employed workers to believe that other people—certainly not we ourselves—are beneficiaries of the American social welfare state.
Without a doubt, though, the economic and policy issues are less clear-cut than my standard economics lectures present. Everything in health policy is so darned complicated and path-dependent.