• Misunderstanding the CBO

    John Goodman* took issue with Paul Krugman’s interpretation of the following chart.

    I’ve posted a similar chart myself. About the chart above, Krugman wrote,

    If Medicare costs had risen as fast as private insurance premiums, it would cost around 40 percent more than it does. If private insurers had done as well as Medicare at controlling costs, insurance would be a lot cheaper.

    Goodman responded in a post titled “Krugman gets it wrong again.” In it, he wrote,

    The Congressional Budget Office has calculated excess cost growth in health care spending after accounting for GDP growth, changes in demographics and the age distribution and other factors. The CBO found that the excess cost growth in Medicare spending from 1975 to 2008 was 2.5 percentage points, on average, each year. Over the same period, the excess cost growth in Medicaid spending was 2 percentage points while the excess cost growth in all other medical spending was 1.8 percentage points. [Bold mine.]

    Goodman’s figures are from the June 2010 CBO Long-Term Budge Outlook. The CBO has been reporting similar figures for years, using the same methodology, which controls for demographic change. For example, one finds the same relationship between Medicare, Medicaid, and all other health spending in a November 2007 CBO report (Medicare with the highest growth, the “all other” group with the lowest). However, in that November 2007 report, the authors explained the methodology a little more thoroughly. On page 8 they wrote,

    Included in other health care spending are payments by private insurers, payments by people who lacked health insurance coverage, all other out-of-pocket payments by consumers, and health care spending by government programs other than Medicare and Medicaid.

    So, “all other” is not just private insurance. It also includes out of pocket spending, even by Medicare beneficiaries, and, crucially, the uninsured. No wonder the rate of increase in spending in that category is so low. It has been pushed down by the growing ranks of the uninsured. Medicare could achieve lower rates of growth too if it failed to cover everyone who is now eligible.

    It should be clear that this is a terrible set of figures upon which to base a conclusion that private spending has grown at a lower rate than that of public programs. The CBO agrees, writing,

    Consequently, the differences in excess cost growth between Medicare, Medicaid, and other health care spending should not be interpreted as meaning that Medicare or Medicaid is less able to control spending than private insurers.

    So, can we really be sure Krugman got it wrong, as Goodman asserted? Not by using the CBO figures Goodman cited we can’t.

    * Though it is Goodman’s blog, it was actually Devon Herrick who wrote the piece I cite. (Apologies for only realizing this now, nine days after I posted this.)

    • Thanks for providing the facts and figures

    • Austin: My reaction to Dr. Krugman’s post was similar to what Paul Levy has said on his runningahospital blog: http://runningahospital.blogspot.com/2011/06/economists-running-amuck.html . Namely “When there is a shortfall in Medicare and Medicaid revenues, the difference is made up by the rates paid by private insurers.”

      Having once worked for a large private health insurer I find this to be true. Contract are negotiated based on the needs of the hospitals and providers, taking into account the competition from other players in the market. When a hospital needs $X Million in revenue per year and collects a certain portion from Medicare and Medicaid, the remainder comes directly from private insurers and out of pocket payments from patients.

      Thus, when Medicare trends lower or pays less it should not be surprising that private insurers pay more … thoughts?

    • @love actuary

      Following your logic…Medicare and Medicaid should spend more?
      So private insurance is cheaper?

      Medicare and Medicaid are already subsidizing the Private Insurance business. Those programs came about and cover a need the private sector was and is still today failing at.

      Without Medicare and Medicaid we would have a huge number of uninsured…probably half the population. We would also have ridiculously higher private sector insurance costs and less hospitals to go to since the revenue wouldn’t be there to support them all.

      Even when you account for the amount of money Medicaid and Medicare cost the federal government and the states being put back in people’s pockets because you no longer need to tax people for those programs…people who were on medicare and medicaid wouldn’t be able to afford the private insurance. why? well, they are either poor, elderly, or both.

    • oh and before anybody says Medicare and Medicaid are crowding out the Private Insurance business…ummm…no.

      if that was the case we wouldn’t have needed health care reform. 33 million or more Americans not covered by medicare and medicaid and i didn’t see insurance companies lining up to earn their business.

      • The number of U.S. citizens who wanted and could not get health insurance was actually closer to 5 million.

        We did NOT need health care reform.

    • fausto: I’m thankful that Austin just listed a host of citations about cost shifting. His conclusion from reading the evidence* is that Wu demonstrates that ~ $0.20 of each dollar today of public-spending cuts (to Hospitals I believe) would be shifted to be private payers. Not 50% – 100%, but 20% is still sizable I think we would all agree.

      You imply that I am in favor of the extinction of medicare or medicaid but that’s not the case.

      Austin, thanks for your links on the noon post today!


    • I’m not a regular reader, so forgive me if I have missed other relevant comments. I am stricken by “love actuary” using the loaded and misleading phrase “a hospital NEEDS $X Million in revenue per year.” [emphasis mine] Let us not forget that this supposed need derives from a hierarchy of other “needs” starting with the need to grossly overcompensate doctors and specialists and administrators. In order to meet that “need”, the hospital “needs” to spend vast sums marketing itself to the public. To be extra marketable, of course, it will “need” to administer the latest treatments and medications, using the latest devices, regardless of efficacy or proof that the new really improves upon the old. A single payer system along the lines of the VA system magically eliminates all of these “needs” while keeping a focus on the most important literal need in the entire system: the need for affordable and effective healthcare for all citizens.

    • Thanks to you, Human Paraquet, for challenging the divine right to money that hospitals seem to take for granted.

      My only caution is that America has allowed hospitals to issue enormous amounts of debt.

      A serious crackdown on pricey hospitals might make us health care wonks feel good, but the collateral damage could be scary. Many cities and pension plans own hospital bonds, and a series of defaults might cause a very ugly cascade.

      Add to that the fact that hospitals are major employers in many cities.
      Hospitals have also sustained the construction industry.

      I think we have to nibble at the problem, in the sense that we try to curb the abuses of hospitals and let them contract gradually.

      Here are a few ways to do the above:

      1. Take the few state laws that limit hospital debt and go nationwide with them. In Illinois, NY, Connecticut, and Minnesota an uninsured person is not liable for more than the Medicare fee, and their debt payments cannot exceed 10% of their income. (That is a rough outline, each state is a little different.)

      Make these laws nationwide!

      2. Pay hospitals a flat per diem fee on all admissions, regardless of intensity or complexity.
      These are usually just code words to do ‘upcoding’ and get more money.

      3. Enact price controls on drugs with no substitutes. Some of the worst hospital bills are just pass-throughs of pharmaceutical price gouging.

      Bob Hertz
      The Health Care Crusade

    • Aaron, why didn’t you show the CBO graphic? It’s certainly more meaningful that the one that Krugman produced.

      In our post, we did not actually say Medicare is worse at controlling costs than private insurers. Paul Krugman, not John Goodman, is the one that is making the unsubstantiated claim. In particular he claims that Medicare saves money when in fact there is no empirical data to support that assertion.

      As for the CBO graphic, if “All Other” medical spending includes some out-of-pocket spending by Medicare enrollees, wouldn’t that indicate the per beneficiary Medicare spending is actually higher than the graphic illustrates?

      Finally, the difference in “Medicare” and “All Other” is substantial (1.8 versus 2.5). If nothing else, that should make everyone be especially suspicious of the unwarranted inference Krugman is making from CMS data.

      • Devon,

        Thank you for your comment.

        The key here is not that I’m “defending Krugman” but that I’m defending the proper interpretation of the CBO figures John Goodman reported. Though you may assert that Krugman’s figure does not prove Medicare controls costs better than private insurers, you may not assert that the CBO figures show the opposite. Therefore, you may not use CBO’s figures in a claim that Krugman is wrong. To make that claim is tantamount to saying that CBO has shown private insurers, in fact, do a better job at cost control. But, they do not show that. This isn’t me saying this, this is the CBO. I’ve quoted their words in describing their figures. Do you dispute CBO’s assessment of its own work?

        Fundamentally, this is not about “was Krugman right or wrong.” This is about whether we understand what the CBO is telling us or not. I think, by now, we do. And it is not what I think Goodman thought it was. Or, if he did understand the limitations of CBO’s figures, why did he not make it clear in his post? That he didn’t certainly left an opening to do so, which is all I’ve done. It’s a service I am sure those who value a fuller understanding of our health system appreciate.

        • The CBO data raises some interesting questions. We concede your point that the CBO data is insufficient to prove public coverage is more (or less) efficient than private coverage. However, despite its shortcomings, it still appears to be more relevant than comparing Medicare expenditures per enrollee to insurance premiums. If anything, this debate illustrates how something as complex as health care spending does not lend itself to overly simple comparisons.

          • It seems to me that Krugman’s graphs would be far more telling about the relative efficiencies of Medicare versus private insurance than the CBO figures used by Goodman, even though it doesn’t tell the complete story, It takes into account the change in the number of beneficiaries while the CBO figures do not

            Looking at the cost of medigap insurance, there is evidence that the comparison produced by Krugman may have some merit.
            Medicare benefits for part B are 80% of the allowed cost . The remaining 20%, on a per beneficiary, would increase at the same rate as the 80% paid by Medicare. The hospital insurance part has a stay limit beyond which the beneficiary has to pay. Part D has the doughnut hole, with a copay component. The math for parts A and D coinsurance is more complicated, but I have no reason to expect that it will show Medicare coinsurance is more costly per beneficiary than private insurance.

            In fact, we know in Massachusetts, private insurance has increased faster per beneficiary due to cost increases, than Medicare. Goodman has told us this, based on the state’s own figures.

            I think Goodman needs to concede that his claim, that the CBO graphic shows that “Krugman is Wrong Again” , is in fact wrong . Since it does not actually prove his point, and he should do it on his own blog.

    • Massachusetts is not characteristic of the country as a whole. Massachusetts, which is more of less the model for the PPACA, has some of the highest insurance premiums in the country. However, I fear Massachusetts is an example of what the PPACA will do to insurance premiums across the country.

      • Devon,
        You can look at the graphic that Frakt provided in his post on the subject.


        It shows private insurance increasing faster than Medicare for equivalent benefits in most years. It shows the annual rate of increase by year. Krugman’s chart is cumulative. It would be interesting to process the data and put both charts on the same basis, but I don’t have the skills.

        • Eric

          What I find intriguing about the graph is how the two lines more or less track each other but periodically switch positions. For instance, during the 1990s when HMOs increased market share the growth in premiums was low.


          I will give your query some thought.

          • Devon,
            I have been checking this thread for days now looking for your best data which shows that private plans are better at decreasing costs than government run plans. You promised to give it some thought. I am really curious to see it.

      • Devon,

        A serious question: what’s the best, yet simplest, evidence that shows private plans are better at controlling costs? I am not kidding when I say this is a serious question. Take your time in answering it.

        I ask because it seems very easy for the pro-government side to toss up very simple descriptive statistics that strongly suggest that Medicare is better and controlling costs. I know that the pro-market side has an argument as to why each one is not valid. I’m not challenging those arguments with this query. I’m asking for the best evidence on the other side, the chart that shows the market outperforming in the most accessible way, and yet that holds up to scrutiny (in contrast to the CBO figures that Goodman cited, just as an example of a bad example).

        Give us your best. I really want to see it. In all honesty, I’ve been looking for it myself. Where is it? Again, take your time.

        • You are very kind to give the free market advocate a lot of time. They have been making the same argument for years, without any real evidence to show which backs their beliefs. We must have experienced their best evidence by now.

    • As a veteran of the health insurance industry, I think that some of the comments above are off base in this respect:

      the commenters assume a connection between health insurance premiums and the ability of insurers to control costs.

      My own experience tells me that premiums are driven by demographics and selection.

      A group of insureds that is older and shrinking (like GM retirees) will have higher premiums no matter what an insurer does to ‘control costs.’
      A school district with mainly over 55-year olds due to seniority will have higher health costs no matter what. For that matter, a school district with many young women that covers pregnancy will have higher health costs no matter what.

      In a so-called free market, a person’s health premiums are set mainly the group they sign up with.

      Those of us who find this grotesque tend to favor broad-based taxes that cover everyone.

      Bob Hertz
      The Health Care Crusade

      • Bob,
        What we are studying is the percent annual increase in costs, not a comparison of actual costs.

    • Quote “s calculated excess cost growth in health care spending after accounting”…this is all JARGON
      why, oh why, can’t economists talk in English.
      spending over cpi, or spending that is greater then the general rate of inflation, or…..
      I also think the authors rebuttal of the the 2nd post, by love actuary, was snide and lazy (no link)

      Finally, the key logical point here – that non medicare cost are low cause they include poor people – is
      supported by the data; it is specualtion; it would be, so far as a i can tell, just as valid to say that the lower growth in private costs is due to the magic of the market…
      I don’t belive that, but the author of this blog hasn’t provided logical strong arguments

      • Yikes! My arguments are illogical? Really? The uninsured spend less on health care than they would if insured. Do you believe that? If so, there’s your logic. If not, why not? (And please be sure to cite research evidence, because all of it that I know shows that insurance causes more spending, not less.)

        • Since you don’t respond to my 1st two points, I take it that you concede..
          As to logic: As I read your blog, you did not offer evidence that the un insured spend less on healthcare. I agree with you that it seems like an eminently reasonable posistion, but feelings are not facts;
          as i read the blog *as posted* you did not put up data to support this, and demolish (i think ludicrous) idea that the private market is reducing costs.
          It is not that I dis agree with you, but that I think you didn’t do a complete job
          But it could be that I am at fault, and did not read the blog post carefully enough

    • “No wonder the rate of increase in spending in that category is so low. It has been pushed down by the growing ranks of the uninsured”

      That seems nonsensical from several angles. How can someone who isn’t benefiting from a policy be labeled a “beneficiary?” How can someone who isn’t buying health care have any costs in the first place? Why are private noninsurance payments not fair game for comparison with Medicare?

      I’m also not sure I believe, prima facie, that the 2010 methodology is necessarily the same.

    • It seems to me that Tall Dave is missing a few things. It has long been explained to those who are willing to understand why it is immaterial who WANTS to get health insurance versus who wants to avoid paying for it. The insurance mandate is a necessity if we are going to continue this absurd insurance-based medical industry paradigm. Most strong intellects prefer single payer, since it is both more cost effective and doesn’t require propping up highly immoral business models (i.e. profiting from denying coverage). Since the powers in charge in congress, from the thwarting minority in the senate to the Senator from Aetna, have said no dice on single payer, we must force those who “don’t wanna” to pay in and create a broader group for insurance coverage. Despite their pouting, one day they will be in the higher risk group and will benefit from young healthy people paying in too. (Note that most people don’t wanna pay taxes either, but that doesn’t remove their obligation to do it.) 

      The problem is that young, naive people and those who are just scraping by, in other words those generally at the low end of the financial spectrum, generally don’t have much disposable income. They don’t want to pay for insurance because they believe it is more expensive than any treatment they are likely to require. But that is simply a naive and ill-informed belief. I’ve known young healthy individuals to get badly injured doing thing older unhealthy folks never do (such as rock climbing). They didn’t think they were likely to be badly injured, but they were.  Only a fool or a willfully ignorant person thinks health insurance is only about illness and not accidents too. Accidents happen at all ages and health levels; the insured bear the cost of medical treatment for the uninsured who get injured and can’t pay for their treatment. Hence, the optimal system requires some skin in the game for everybody and, therefore, some spreading of the risk. 

    • *crickets*