In a May 15 Health Affairs Blog post, Jeff Goldsmith argues against creating a new Medicare-like public health insurance plan to compete with private plans. As part of his argument, Goldsmith asserts that Medicare has done a worse job of controlling costs than private insurers have done.
Goldsmith bases this assertion on a recent paper by Richard Kronick, citing figures that Kronick in turn quoted from a study by the Congressional Budget Office. The problem is that, as the CBO itself was careful to point out, the CBO study didn’t actually say that Medicare was worse than private insurers at controlling costs. In fact, the bulk of the available evidence indicates that Medicare has been better at controlling costs than private insurers.
What Goldsmith Argued
In Goldsmith’s words, “a recent study for the New America Foundation by Richard Kronick showed that ‘excess’ per capita growth rate of Medicare costs (that is, beyond inflation) was 20% higher from 1971 to 2005 than that of private insurance (2.4% per year vs. 2.0%).’”
The first confusion arises between Goldsmith’s report and what Kronick said. The Kronick numbers cited in the blog post were for 1975-2005, not 1971 – 2005. The numbers came from a CBO study published in November 2007, The Long-Term Outlook for Health Care Spending. Kronick summarized that “excess growth in Medicare spending per beneficiary was slightly greater than excess growth in private sector spending from 1975 to 2005,” and then noted the contrast between those CBO data and the fact that “other researchers have found that Medicare spending growth per beneficiary was slightly lower than spending growth per privately insured person over the 1970 to 2000 time period.”
Given that comparison, it is understandable that Goldsmith read the figures as referring, in his words, to a comparison between the growth of Medicare costs per capita and the growth of private insurance costs per enrollee. But that’s not exactly what Kronick said, although he did give the impression that his data addressed the comparison between the cost-control experiences of Medicare and private insurance.
What The CBO Said
This difference between private-sector spending per capita, on the one hand, and growth in private insurance costs, on the other, is more important than one very distinguished health policy analyst’s misreading of another’s article. In its report, the CBO explained:
“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 or Medicaid. 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” (p. 10).
This caveat was called for because, as both Goldsmith and Kronick acknowledge, other analyses have said that Medicare has had a cost-control advantage over private insurance. A recent example is MedPAC’s June 2008 Data Book on Healthcare Spending and the Medicare Program, Chart 1-7 on page 9. MedPAC reports that, “comparing spending for benefits that private insurance and Medicare have had in common — notably, excluding prescription drugs — Medicare’s per enrollee spending has grown at a rate that is 1 percentage point lower than that for private insurance over the 1970 to 2006 period.”
Two Different Comparisons: Medicare vs. Private Insurance, And Medicare vs. All Other Spending
This raises the question of whether we have any evidence as to whether a comparison to all spending other than Medicare and Medicaid would yield a different result than we’d get from a comparison between other insurance and the two big public programs. A logical place to look is in one of the sources cited in Kronick’s chapter: The Health Affairs Web Exclusive by Micah Hartman and coauthors from November 6, 2007. Exhibit 4 of that article (see below) breaks down personal health care spending per capita by various categories. That data can be used to calculate increases over the period from 1987-2004, which isn’t exactly the same as the CBO’s 1975–1990 and 1990–2005. But it is a policy-relevant time period, and it gives us a look at the relationship between private insurance spending and “all other” spending.
The most obvious comparison is between Medicare for the core Medicare population (age 65+) and private health insurance (PHI) for the core PHI population (ages 19-64). Over the 17-year period, costs per capita for Medicare 65+ rose 202%, and for PHI 19-64 by 246%.
Overall costs that were neither PHI, Medicare, nor Medicaid rose far more slowly than for either of those categories, which seems likely to explain a discrepancy between the standard comparisons of PHI and Medicare, and the CBO comparison of all non-Medicare, non-Medicaid spending with Medicare. For example, out-of-pocket spending for the 19-64 group increased by 80% during 1987-2004; “other private” spending for the same group increased by 156%; and “other public” increased by 90% — all much less than the private insurance figures.
Furthermore, the PHI figure in Hartman and colleagues’ work must understate increases in cost per PHI beneficiary, because it is not per beneficiary. It is per person within the age breakdown, and the proportion of people ages 19-64 with PHI fell substantially between 1987 and 2004. If we calculate from the estimates in the census data, the numbers yield a private insurance rate of 77.3% for ages 18-64 in 1987 and 70.7% in 2004. I do not recommend such raw calculations from tables; and the ages aren’t exactly the same; but it seems likely that the proportion of the population ages 19-64 with private insurance fell by about six percentage points between 1987 and 2004. If we adjusted the denominators for PHI costs in 1987 and 2004 for the portion of the 19-64 group that actually had private insurance, the cost increase for PHI would be about 270%.
The Evidence On Spending Per Person Still Favors Medicare
I am not supporting that precise adjustment. But the CBO was not reporting that Medicare controlled costs less successfully than private health insurers did; and, in fact, its analysis did not contradict the more standard estimates of Medicare’s cost-control advantage.
There may be good reasons to question the so-far-standard estimates of Medicare’s cost-control advantage over PHI. That could be a topic for further discussions, and I know Kronick has ideas. But the data he presented in his chapter of the New America Foundation report are not a sufficient challenge. He presented very striking evidence of larger increases in volume for some services for the Medicare population as opposed to the younger population. As he notes, some of this may be due to developments in technology that apply more clearly to diseases of the elderly, such as cataract surgery or joint replacement surgery (p. 16). He sees the volume increases as evidence that Medicare’s cost-control performance might not be as good, compared to PHI, as the previous estimates have suggested. A seemingly more logical conclusion from combining his utilization data with the MedPAC cost estimates would be that limiting prices and overhead is really important.
This last point, the importance of reducing overhead and getting the best possible control of prices, is perhaps the most important for current policy debate. Advocates of the “public plan” that Goldsmith was criticizing are almost the only ones reminding policymakers of the importance of prices and overhead costs. Kronick’s study is part of an overall report on Making Medicare Sustainable, which proposes versions of the current reorganization agenda for improving “value.” While many of these are attractive, it is worth remembering that paying a lower price for something makes it a better value.
Medicare, like the rest of U.S. health care payment, could be greatly improved, but the overall evidence on spending per person still favors Medicare. And the CBO’s warning about confusing “other health care spending” with the ability of private insurers to control costs appears to be warranted.