• Cost shifting from the uninsured

    At the WSJ, John Cogan, Glenn Hubbard, and Daniel Kessler weigh in:

    [T]he administration and Congress have asserted that people with private insurance pay for care for the uninsured through “cost shifting”—higher prices charged by doctors and hospitals to recover losses from uncompensated care. […]

    Our review of the research has found that there is no credible evidence of a cost shift of any substantial consequence, either within state boundaries or across state lines. […]

    A study conducted by George Mason University Prof. Jack Hadley and John Holahan, Teresa Coughlin and Dawn Miller of the Urban Institute, and published in the journal Health Affairs in 2008, found that so-called cost shifting raises private health insurance premiums by a negligible amount. The study’s authors conclude: “Private insurance premiums are at most 1.7 percent higher because of the shifting of the costs of the uninsured to private insurance.” For the typical insurance plan, this amounts to approximately $80 per year.

    The Health Affairs study is supported by another recent peer- reviewed study that focused exclusively on physicians. That 2007 study, authored by Massachusetts Institute of Technology economists Jonathan Gruber and David Rodriguez and published in the Journal of Health Economics, found no evidence that doctors charged insured patients higher fees to cover the cost of caring for the uninsured.

    The authors go on to say that the large cost shifting estimates come from a Families USA report. Still, the body of evidence is very thin. This is all consistent with what I wrote in my forthcoming Milbank Quarterly paper on hospital cost shifting. The focus of that paper is on the cost shift due to shortfalls in Medicare and Medicaid payments, not the uninsured. But I did write this:

    Families USA (2005) estimated that private insurance premiums were about 10% higher in 2005 due to the use of health services by the uninsured. Kessler (2007) and Hadley et al. (2008) both find less than a 2% effect.

    So, I’m inclined to agree with the Cogan, Hubbard, and Kessler about the quote above, but not fully about this:

    If anything, the likely impact of the law will be to increase, not decrease, cost shifting. According to the Congressional Budget Office, around half of the people who are expected to become newly insured under the new law will be enrolled in Medicaid. But Medicaid payments to doctors and hospitals are so low that the program creates a cost shift of its own. In fact, a long line of academic research shows that low rates of Medicaid reimbursement translate into higher prices for the privately insured.

    There are two issues here. First, one has to consider where the new Medicaid enrollees come from. As I wrote in my paper, if they come from the uninsured population, then Medicaid payments may be an increase in compensation for care. However, it is likely that a sizable fraction will come from the privately insured population, a crowd-out effect about which you’ll see more about on this blog soon. Since Medicaid reimbursements are much lower than those from private insurers, that could play into cost shifting. Cogan, Hubbard, and Kessler have a point there.

    The second issue is, how much cost shifting is there from Medicaid? The bulk of the studies I reviewed in my paper focused on Medicare. Those that focused on Medicaid found even lower cost shifting rates from that program. Perhaps you recall that the most plausible Medicare cost shift rate was 21 cents on the dollar (a dollar in Medicare payment shortfall causes a 21 cent increase in private hospital prices). Medicaid estimates are much lower, but they have not been examined closely with recent data. (There is no evidence of cost shifting for physician services or pharmaceuticals. Hospitals are where it’s at, if anywhere.)

    Even if we assume the 21% Medicare cost shift rate applies to Medicaid too, that does not mean a 21% increase in private premiums. It’s a 21% increase in hospital payment rates. Premiums cover more than hospital care and the volume of hospital use has to be factored into the calculation to deduce the ultimate effect on premiums. No doubt there is an upward effect on premiums due to shortfalls in public payments, but it isn’t huge. The vast majority of lower public payments are translated into cost cutting, not cost shifting. That’s not necessarily a good thing either, which is just another illustration of how complicated the health system, and health policy, are.



    • –The 2008 paper in HA: They estimated ~120B or so in costs for these uninsured folks, and a good portion of the shortfall cane from charitable contributions from the community.

      While not a cost shift in the tradition of the usual hospital to MCO transfer, is it still not a cost shift? The answer is yes.

      –Hospitals, like you state, are different than commercial plans. However, deficit spends come out of general revenues and Part A, and it is sloshed around one big account. Hard to account for, but also, another shift.
      The danger here is trying to parse out the uninsured and the effect they have on private payers. In the end, because of our system’s inability to cost account properly and the comingling of Mcare, Mcaid, Commerical, and self pay buckets into one, we are losing site of the national and interlinked nature of our health care system. This is interstate, not local!

      Finally, less to policy and more to politics, the subtext of the oped was to add fuel to the eventual SCOTUS challenge and the commerce clause conundrum.

      That is a separate debate of course; the economic and cost shift dilemma is a challenge unto itself and worthy of discussion on its own merits– if not to settle some important questions on the policy front, but to help us decide where to allocate our dwindling resources as well. It is important to note the duality of the post as well.


    • From recent testimony of a hospital CFO before a state legislative subcommittee:
      – Bad debts and charity (mostly from the uninsured) add up to roughly 8% of gross revenue. For scale, Private insurance adjustments are 5%, Medicaid is 7% and Medicare 30%.
      – The hospital targets rate increases to procedures with high usage by commercial insureds.

      I’d say that qualifies as cost shifting, and while it isn’t as big as Medicare it’s far from nothing. As an aside, are other states talking to hospital officials on the record?


      Audio of the hearing (CFO’s testimony begins at 25:30, he says pretty explicitly that they cost shift at 29:00 minute mark)

      Exhibit accompanying his testimony. See page 5 for gross revenue breakdown.

      • @bjc – When stories — however true — replace data analysis, science is dead. We have data on these things. Scholars have studied it and produced estimates. No doubt there is variation. Yet, every hospital CEO will claim massive cost shifting, far above average. Are they all in Lake Woebegone?

        Also, and most importantly, price discrimination is not cost shifting. If you haven’t seen my posts on that point, search for them.

    • So where does the cost of caring for the uninsured go? Or is just not that much in the grand scheme of things?

      • Price discrimination (one group paying high prices than another) doesn’t imply causal cost shifting. There is subsidization without causation.

        • Oh my. I see where I’ve gone wrong now. I’ve been feeling particularly annoyed because I’ve believed that there is a percentage of insurance premiums being specifically set aside to pay for the uninsured–which, it seems is true. The error is in calling it “cost-shift” rather than “price discrimination”, Correct? (I just read the post covering this issue.)

          Which now simply makes this history/environmental science major laugh at you economists. You’re speaking “Scientist” while Families USA is speaking politics/heart/reality (i.e. “American”)–I’d put my money on you being able to find that their data is more or less correct once you’ve accounted for the errors in terminology.

          It’s rather scary how most arguments are dependent entirely on two people speaking different languages, but, alas, people look at ya funny when you pull out the dictionary every third word to make sure that you’re both actually saying what you think you’re saying.