• How will the Cadillac tax shift low-income workers to Medicaid?

    In his first post here at the Incidental Economist, Don Taylor (welcome Don!) suggests that capping the tax exclusion for health insurance premiums would be better policy than the Cadillac tax that is part of the Affordable Care Act.  He gives two reasons.  One is that a cap would be more progressive tax policy—it would tax low-income workers at a lower rate than high-income workers.  I agree that this is correct and would be desirable.  Don’s second reason is that he thinks a cap would shift fewer low-income workers to Medicaid than the excise tax.  This shift is a potential problem that I’ve discussed in recent posts here and here.  Unfortunately, I don’t think Don is right that a cap would be better.

    As we review in our recent working paper, the standard economic theory of employer-sponsored health insurance says that employers choose to divide total compensation between wages and insurance benefits based on the average marginal tax rate of their workforce, weighted by the relative marginal utilities of wages and health insurance for each worker.  If firms choose generosity of coverage and premium levels based on tax rates averaged across workers, then even a tax increase targeted exclusively at high-income workers will result in less generous coverage and higher employee contributions to the firm’s health plan required from all workers.  The incentive for low-income workers to shift to Medicaid lies in these changes to the firm’s plan; it doesn’t arise directly from the tax.

    As I said in my original post on this subject, we don’t have a philosophical objection to having more people covered by Medicaid, but we think this shift should be accounted for in the official budget scoring and as far as we can tell it hasn’t been.


    Comments closed
    • Steve
      Just finished Don’s post, and I am still confused.

      The low wage worker jumps ship because his/her tax bite goes up, let’s say $1000-$1500. NOw buys in HIE or gets Medicaid.

      Company pays a penalty of 2-3K as employee has elected to go elsewhere.

      Does this make sense? As inflation takes effect, at some point yes, but for the sake of discussion, we are talking nearer term.

      Also, still have not read your paper in full, but can you give me a real world example whereby, using Don’s chart, someone will earn $10K and obtain $19K worth of employer coverage?

      The numbers dont add up, and I am not seeing how a company would configure an inefficient position such as this (? hire P/T instead, no?).

      Wont happen at a fast food chain, or on the migrant farm circuit–maybe in mail room at Fortune 5000 company?


    • Austin
      Missing my point….

      You pay someone a very low salary, Don’s hypothetical 10K. Don also assigns and employer outlay of $19K for ESI. I am asking why? This has nothing to do with tax hit, but behavior of firm (keep in mind I am not an economist, just play one on TV).

      Why would a company want this drag? Hypothetically, they could outsource this cheap labor and avoid the health care costs, pre-caddy tax; or hire 2 P/T workers at 20 hours/week each, with a 5K stipend, etc.

      Just not seeing this through the right lens–I am missing something here?


      • @Brad F – It’s not a drag on the firm. Remember what matters to the employee is total compensation. How that is distributed between health care and wage is a matter of taste, though it is highly influenced by tax policy, which distorts the balance in favor of health care.

        You seem to think a firm can just do anything. But it isn’t the only actor here. There are other firms that are in competition for labor. Do you think firms are profit maximizers or not? If so, they’ve made the best balance that exists given the constraints. If not, you’re free to call them (or their agents) irrational and stupid, but that’s a theory that doesn’t produce any testable hypotheses.

    • Brad
      sorry slow on answer….pushing a competing renewal grant in due today…brief thoughts…more later.

      The biggest questions if we do change tax treatment of employer paid insurance (whether via high cost tax or reform tax code) is (1) what will firms do? (2) what will employees want firms to do? I think the firm is the least concerned, because a dollar in wages or in insurance premiums are both deductible against revenue as business expense. The subsidy now flows to the employee. Huge question is how much employees would want firm to change if all the sudden their premiums are taxable as income, and then how much power employees would have to get said changes. I proposed an experiment to a Duke HR person whereby we would randomly offer employees the option of less in premium support and more in wages to see what happened. He just laughed at me….

      • Guys (Don, Brad) – My paper with Steve, about which he blogged, is all about this question. It uses firm-level tax price as a key explanatory variable of firm offer, as well as private and public coverage. It references a whole bunch of other relevant literature on this question. So, we know something about what employers will do. If Steve’s posts aren’t sufficient, check out the paper.

    • Steve
      digging into the paper a bit more and I am trying to compare you guys findings with the CBO estimate of the final ACA bill . Table 2 of CBO report shows +16 M Medicaid, +24M into exchanges. with -4M from employer coverage and -5M from other private.

      On page 19 you say “four fifths of new Medicaid enrollment…would have been otherwise privately insured. So, you guys estimate much larger crowd out than did CBO so far as I can tell (CBO net -9 crowd out to medicaid & private combined). CBO does estimate crowd out, you are just saying not enough, correct? Any way you could ‘simulate’ table 2 from CBO table from your results? I am unsure. Interesting paper.

    • Don,
      You emphasize an important point that maybe we should hit a little harder: our Medicaid crowd-out estimates, while in the range of other published estimates, are probably a lot higher than what CBO used. As usual, it’s hard to tell exactly what CBO is doing, but they probably took an estimate from the literature and their results suggest that the estimate they used was on the low side (I also recall reading in one of their documents a 20% figure, but I can’t dig it up right now).
      Our results are based on analysis of a sample of working-age adults, which is different from most of the crowd-out literature. This is important because that’s the population targeted by the ACA Medicaid expansion. Our crowd-out estimate (about 80%) is plausible to us because our sample consists entirely of workers and their families– access to private insurance including employer-sponsored plans is much higher than it would be in a sample that included families with no connection to an employer.
      By excluding those families from our sample we lose the ability to compare directly to CBO projections, but we didn’t think it made sense to model the two groups together (and we wanted to compare to other papers with similar sample restrictions).
      The bottom line is that we think there are probably two problems with CBO’s projections. The crowd-out estimate is probably much too low and there doesn’t seem to be any accounting for the shift induced by the tax.
      Thanks for the thoughtful read.