• Price Differentials Are Not Evidence of Cost Shifting

    A recent Business Week article summarized an argument against a public option that is based on the unfounded claims of cost shifting made by insurers and hospitals. In it Jane Sasseen and Catherine Arnst write

    Proponents add that government competition would force private insurers to lower premiums, making coverage more affordable for all.

    But the business community keeps cranking out studies undercutting such arguments. The insurance industry trade group, America’s Health Insurance Plans (AHIP), compared the lower reimbursement rates for health care paid by public programs vs. private payers. The group claimed the difference reflects cost shifting, which added an estimated $1,512 to the average premiums paid by a family of four. “The existence of the private sector allows that shift,” says Karen Ignagni, the head of AHIP. “If you clamp down on one side of a balloon, the other side just gets bigger.”

    Ignagni’s balloon analogy is a false one, and AHIP’s cost shifting argument is faulty. It is based on a study that misapplies the term “cost shift” to price differentials.

    In December 2008 Milliman published a report with the unfortunate title “Hospital & Physician Cost Shift: Payment Level Comparison of Medicare, Medicaid, and Commercial Payers” (authored by W. Fox and J. Pickering). The report was produced at the request of America’s Health Insurance Plans (AHIP), the American Hospital Association (AHA), the Blue Cross Blue Shield Association (BCBS), and Premera Blue Cross. AHIP’s December 9, 2008 press release summarizes the report.

    You’d think with “cost shift” in the title the report would be about cost shifting. It isn’t. It is about the fact that public and private payers pay different rates to health care providers. The report shows that in the 2006-2007 period Medicare and Medicaid paid $88.8 billion less than private payers for hospital and physician services than they would have if public and private payers paid the same rates for health services.

    The report goes on to translate this payment differential into an amount attributable to health insurance premiums. The claim is that private health insurance premiums for a family of four are $1,512 higher per year than they otherwise would be in the absence of the payment differential identified.

    The Milliman report is widely misinterpreted, in part due its misleading language. Price differentials, which it describes, do not support the conclusions reached. Just because payers A and B pay different prices does not mean that if payer A paid more payer B would pay less. I’ve made this point elsewhere, as have Uwe Reinhardt and Richard Frank.

    Look at it this way: McDonald’s [sic] pays far less per pound of chicken (or beef) than you or I do. That’s a price differential based on market power. Being a far bigger purchaser McDonald’s has far more market power than you or I and commands lower prices. But if McDonald’s suddenly paid more, would we pay less? Do we believe chicken (or beef) costs have been shifted from McDonald’s to us? No. It is a price differential, not a cost shift.

    Another example: you take a second job on weekends and it pays less per hour than you make during the week. Does this mean your weekend job shifts costs to your weekday job? If your weekday job cuts your pay, will your weekend job pay more? No. It would be foolish to expect such a thing and strange for any employer to provide it.

    If price differentials are not cost shifting, what is? Economists’ definition is consistent with a plain language interpretation: a cost shift is a causal relationship between what is paid by A and paid by B. If changes in price paid by A cause changes in price paid by B then that is a cost shift.

    Scholars have looked for such cost shifting in health care and found very little evidence of it. In the rare cases it occurs, it does so at a minimal level. It is a myth that in health care lower prices paid by public payers are substantially offset by higher prices paid private payers. Yet this is precisely what the Milliman report claims (or strongly implies by its use of misleading language).

    The Milliman report did not come anywhere near looking for or finding such a causal relationship. It is not a report on cost shifting, despite what it says and the way it is used by its sponsors.

    Share
    Comments closed
     
    • Austin’s comment is right on the mark. In fact, the cost-shift argument confronts the health insurance industry with a dilemma.

      If it were, indeed, true that whenever government cuts payments to health care providers, private payers must pay providers more, then one may ask whether the private health insurance industry can at all constrain the growth in health spending.

      If insurers are patsies in one sphere, passively absorbing whatever cost shift comes their way from government, what makes us think they can resist and other cost shift coming from other quarters, e.g., the shift of costs caused by the medical arms race among hospitals?

      And if they admit that they cannot control costs shifts of any kind, what makes us think that it is wise to delegate cost control to them?

      Uwe Reinhardt

    • The Medicare Payment Advisory Commission (MedPAC) analyzed this point and they concluded that there is probably a causal relationship, just not the one everyone thinks (see my post on this at http://www.medicaidfirstaid.com/2009/08/chicken-or-egg-case-of-government.html ).

      A good graph that plots govt and private payment-to-cost rations is at http://www.medicaidfirstaid.com/2009/07/pottery-barn-rule-of-health-care.html .

      • @Brady Augustine – I fixed the URLs for you. Are you in agreement with MedPAC’s interpretation, which is consistent with what I’ve said? The debate might be clearer if we reserved the term “cost shift” for the dynamic, causal relationship between price of payer A and that of payer B and use “price differential” when that is what we mean. Other than folks who have an obvious conflict of interest I don’t know anyone who believes substantial cost shifting actually occurs, despite high price differentials.

    • Isn’t this apples and oranges?

      The Milliman report compares a “typical PPO plan” – which often has coverage of 90+% (the study doesn’t say what insurance plans they compared) By contrast, Medicare only pays 80% of the bill. The report is vague as to how it measures “reimbursement”. Is that total reimbursement, including medicare + patient responsibility? Or is it just the medicare portion? This is a critical detail, which isn’t stated.

      Isn’t is as simple as this – Medicare saves money because it pays 80% compared to a private plan which pays 90%.

      Also, don’t forget that plenty of the best doctors simply don’t accept Medicare. So this compares private insurance, which provides better coverage and access to more doctors to Medicare, which provides lower coverage (80% vs 90%) and access to fewer doctors. Of course it is cheaper!

    • Forgot my link about doctors that don’t accept Medicare. According to this article, at one hospital in NY, only 37 of 93 internists accept Medicare.

      Do you think the best doctors have the cheapest prices? Or do you think the best doctors have enough business that they just don’t need to deal with Medicare’s overly restrictive rates?

      In other words – in addition to a cost shift by moving to Medicare – you’ll also experience a reduction in choice and (equivalently) reduced quality.

      http://www.nytimes.com/2009/04/02/business/retirementspecial/02health.html

    • You really think that the reason McDonald’s pays less for chicken is because of “market power?” Really?

      I’d say the reason McDonald’s pays less for chicken is largely because of lower real costs. Chicken sold to McDonald’s doesn’t need the resources spent on distribution and marketing that are needed to get small quantities of fresh chicken to consumers. There is little or no packaging, or overhead, or spoilage, etc. etc. Contracting with producers probably also reduces transaction costs per unit and lowers risk for producers, thus lowering price again. These are all real savings and have nothing to do with lowering monopoly rents for producers.

      I expect the wholesale market for chicken is a quite competitive market with few monopoly rents. The grocery market is also fairly competitive, with low margins and lots of options. So if McDonald’s pays less, it’s not market power, it’s real cost savings.

      Can you cite any evidence for how important monopsony market power is in markets like this? I’d be sincerely interested.

    • @Austin Frakt: I agree with MedPAC’s analysis and your differentiation of differentials and shifts. Personally, I think the points you raise are not as well known as they should be. The phrases “hidden tax” and “cost-shifting” are often dropped into the health reform debate by knowledgeable persons at times of inattentiveness or convenience (including myself). ~BAA

    • Thanks for the response. And you’re right, it’s kind of beside the point as far as the topic of the post.

      I objected because I have been hearing lots of pundits casually mentioning the importance of monopsony market power ideas in a variety of contexts as if it were totally obvious, even where it may be completely misguided.

      I’m still not clear on what you believe causes differences in prices between public and private insurance. I’ll be interested in looking at some of your other posts.

      • @ed johnson – Public prices are set by a political process (though called “administrative pricing,” which is a misnomer). Private prices are set by negotiation, and are subject to forces of market power. The extent to which any price is sufficient depends in large part on cost. MedPAC’s report, to which I linked in an earlier post on this topic is good reading on this and I agree with it.

    • Something seems problematic here for me: the bargaining power that insurers have is not based on efficiency or scale; it’s based on the contract with subscribers to only go to certain hospitals in the plan, which is not a real efficiency and doesn’t create real value (their value is in cost sharing good treatment, not deciding treatment facilities). And importantly, having an insurance company deliver a patient to a hospital (versus them finding it on their own if they were uninsured) and getting billed is very different than how we purchase chickens. The hospital service is going to be delivered the same way to the patient no matter how they got there, as long as they pay comparable rates (because doctors prescribe and deliver treatments, not insurance companies).

      My take is that since there was never real value being created in how insurers channel hospital use by consumers, we’ve always been cost shifting — the uninsured have been picking up the tab for the “value” that insurers have supposedly been creating in this domain. The price differential is a result of many iterations of cost shifting and finding an equilibrium in which the uninsured still can afford to pay and the insured still feel like they’re getting a bargain.

      Better than chickens: let’s use the frequent shopper card at my grocery store. If I join the free program, I get lower rates; if I don’t (because I don’t think I’ll be using the services much, but in fact I do; cause heck, I really like milk), I end up paying the higher rates and subsidizing the people who have joined (if they are in fact getting a discount). The grocery store probably created this program for marketing purposes: to watch my spending habits. But here lies the problem: the supermarket isn’t creating real value for me (having the shopper card doesn’t change the way milk is produced or distributed, so there is no market efficiency here). Instead, the supermarket picks up the value of the system: they have personalized marketing data on my habits that I would normally want to protect, and they’ve gotten me to identify with their shopping brand (by having my carry around their card).

      (and even if I’m not subsidizing the cost of milk for those who have joined the program because their price is the true price of the product, I am contributing to the store’s profits, so my point is still useful here.. since the price differential facilitates the store’s business model)

      • @Birju Patel – I like your store card analogy. It is better than the chicken purchase one for the reasons you state. Let’s tweak it a bit to make it even better. Suppose my large employer has a purchasing program (in fact it actually does, as did my prior one). Through it I get access to certain durable goods (in particular cars) at pre-negotiated prices. Merchants, retailers, manufacturers participate because they get more reliable and higher volume. Employees participate for the lower prices. Meanwhile, those outside the program pay higher prices. That is a price differential.

        To what extent is it a cost shift? Well, what is a cost shift? A cost shift occurs when the lower price paid through the employer discount program causes the price others pay to be higher than it otherwise would be. It need not be dollar-for-dollar higher. Perhaps there is a cost shift of 10% or 20%. How would we find it? Strong evidence would be if variations in the discount price were related to variations in the non-discount price, all other things being equal.

        Move to health care. There have been studies that have looked for exactly that type of causal relationship. MedPAC, CBO, and I have looked at those studies. They either show no cost shifting in health care or a very low level of it, on the order of 4-17%.

        Meanwhile, the Miliman report gives the impression that the cost shift is dollar-for-dollar. That’s because they look at price differentials and call it a cost shift. That’s just not right.

        I challenge anyone to find a recent (within the last decade), peer-reviewed, body of work, not obviously conflicted through relationships with insurers or providers that shows a significant cost shift. If you find it, send it to me. I will read it. I will blog about it.

    • Yes, if a government takes over a sector the goods will not become more expensive. Really?

      Have you ever, ever, in the history of mankind ever recorded a single government good that was provided for cheaper than the public sector?

    • I guess when Mayo Clinic says that they can no longer cover the losses on Medicare patients in certain departments to the point that they are no longer accepting medicare patients in those departments that they are not cost shifting? So I get it. When providers make explicit statements that they are managing their portofolio of patients to manage cover their costs they are not cost shifting unless it is confirmed by some study?

      I think some “reality freebasing” is in order here.

    • This is such an old post, i do not know if this comment will be read–but i am going to try anyway.
      I understand the points about price differentials vs. cost shifting vs. market power. i have read most but not yet all of the posts here on cost shifting.
      I really like the analogy of the frequent shopper card–but no one has mentioned the fact that hospitals which accept Medicare not only accept their lower payments, but by law, are also required to accept all comers through the ER irregardless of the ability to pay. This is a significant difference between health care and other market sectors.
      Is the thinking that the effect of the uncompensated care losses would be expressed equally in Medicare and private insurance prices (or subject to the changes in market power as discussed above)?

      I will keep reading.