On health insurance subsidies, taxes, and distortions

By my own assessment, as well as that of some followers on Twitter, Casey Mulligan’s latest post on the Economix blog is not easy to parse. He makes a good point about implicit taxes in the Affordable Care Act that’s worth understanding, so I will elaborate. He also omits a related point that is no less worthy, which I will get to.

There could hardly be a health policy wonk that doesn’t know that (1) workers with access to affordable (defined here) employer-sponsored insurance (ESI) are not eligible for premium tax credits or cost-sharing subsidies for exchange-based plans; and (2) exchange-based premium tax credits and cost-sharing subsidies phase out with higher income (eliminated at 400% and 250% of the federal poverty level, respectively). Mulligan thinks the labor market effects of the former have not been fully appreciated.

The Affordable Care Act contains at least two economically distinct taxes on labor market activity. Even the experts on the law have failed to recognize all of them. […]

[T]he law’s advocates have yet to acknowledge the new implicit employment tax, let alone estimate the number of people who will face it.

It’s pretty obvious that both are work disincentives, since only by working can one potentially obtain access to affordable ESI. And, by working, one’s income rises, leading to a lower exchange subsidy if one qualifies for one. To clarify that these are two, distinct taxes on labor, Mulligan imagines the a world in which exchange-based insurance subsidies didn’t phase out.

[E]ven if the health insurance subsidies in the Affordable Care Act had been a specific dollar amount that was not phased out with household income, the law would still act as a tax on employment because most workers [those with affordable ESI offers] could not get the assistance during the months they were at work.

Mulligan is correct. However, I would dispute that this is something experts and advocates have failed to acknowledge. I recall many discussions of the differential subsidization of health insurance for two individuals with the same income, one with access to affordable ESI coverage and one without. I concede that finding those discussions now, several years after they occurred, is not easy. Perhaps Mulligan could not. I also tried and have, so far, failed.

What Mulligan perhaps means, however, is that until recently there has not been much work on quantifying the labor market effects of the fact that most workers are not exchange-subsidy eligible. He cites his own work, that of the CBO, and a paper by David Gamage that does so. I defer to his knowledge of the literature on this point and refer you to his posts for the links.

So, Mulligan’s point that there are two separate work disincentives bundled in exchange subsidies may not have been clear to some readers, but it is a fine one, as far as it goes. What he did not mention in his post,* however, is the other, major subsidization of health insurance for the working-age population—the tax exclusion of ESI coverage—the system it supports (the only guaranteed issue, community rated one for the non-elderly, pre-ACA), and how both those will be affected by the ACA.

That the cost of ESI coverage is not taxed is a huge work incentive. Moreover, in contrast to exchange subsidies, the tax exclusion is an implicit subsidy that grows with income. Of course there are huge labor market effects associated with it and the community rated, guaranteed issue, employer-based system it supports. Among them, pre-ACA, is the fact that some people could not as easily obtain affordable coverage on the individual market (due to underwriting and/or lack of subsidization) and had to keep working just for the ESI benefits. This “job lock” story is also acknowledged by the CBO, as it should be.

The ACA will affect this work incentive in at least two ways. It reforms the individual market so that consumers can’t be denied or charged extra for coverage just because they’re sicker. Also, beginning in 2018, it will impose a “Cadillac tax” on high-premium ESI plans that will offset the ESI tax exclusion, albeit in an imprecise and suboptimal way.

Like Mulligan, I’m no fan of the unequal and distortionary ways in which various explicit and implicit taxes act as work incentives or disincentives. What I would like to see, however, is an evenhanded treatment of them. Noting that the exchange subsidies impose an anti-work, distortionary force without acknowledging that the tax exclusion for community-rated, guaranteed-issue ESI (for those who can obtain it), as well as the pre-ACA individual market, impose pro-work, distortionary forces gives a misleading impression of what the ACA will do, and what the CBO and many others have expressed.

I’ll end with a final point: Work is good. But it is not the only good. If it were, we’d all be advocates for heaping on even more benefits for workers. We’d eliminate Social Security and Medicare. We’d heavily tax non-work income (capital gains). And, we’d certainly not object to very generous tax treatment and preferential market structure of ESI.

One thing most economists and policy wonks agree on is that the ESI tax exclusion, though a benefit to workers, is too distortionary and should go. Most economists are, therefore, willing to accept—even welcome—the removal of an incentive for work. Of course some people will choose to work less if it is removed. The question we should ask is: do we really want people to work just because it’s the only way they can obtain affordable health care insurance? All quantification of labor market distortions aside, that sounds like a significant reduction in liberty to me.

* He is well aware of it, and I trust it is in his work. It certainly is in that which he cites, the CBO’s in particular.

UPDATE: Further clarification of a key point in this post is here.


Hidden information below


Email Address*