Medical Progress Report 5 | Comparing Public and Private Health Insurance

Comparing Public and Private Health Insurance: Would A Single-Payer System Save Enough to Cover the Uninsured?

The lower reported administrative costs for Medicare are unsurprising, in that Medicare spends substantially less on such functions as marketing, risk evaluation, claims scrutiny, and compliance with the regulatory requirements of the individual states. This does not mean that the higher reported administrative costs of private health insurance are “wasteful.” Instead, they serve the interests of consumers by reducing the extent to which insurance creates cross-subsidies among consumer classes; such cross-subsidies reduce the economic benefits of risk-pooling. Private administrative functions also impose discipline on the consumption of health-care resources, thus reducing upward pressure on insurance premiums.

In contrast to private insurance, single-payer systems must have the effect of creating and increasing cross-subsidies among patient and voter groups, because eligibility, tax burdens, and premiums are not based on health status, and the tax system prevents competition on the basis of price. Accordingly, the deeper question underlying the issue of relative administrative costs is far more fundamental: Should a health-insurance market be viewed as an institution with which risk-averse individuals and groups can pool risks efficiently? Or should it be viewed as a vehicle with which to redistribute wealth?

Note that the effort of individuals and groups to avoid the costs of subsidizing others is not a phenomenon limited to the private sector. Such competition for lower costs is a prominent feature of public finance as well, as various groups prefer—strongly—to enjoy increases in their preferred programs at the expense of others’ programs, and to shift the tax burdens necessary for public spending programs onto others. Neither private nor public health insurance in the context of allocating costs is a charitable endeavor.

No health-insurance system, whether private or public, can “cover” all individuals or all medical services because resources are limited always and everywhere. This means that both private and public health-insurance systems must impose limits on the consumption of health care: Some classes of services will be denied to patients, and some classes of patients will be denied given services. “Universal coverage” is an unattainable goal; private and public insurance programs are likely to use different criteria with which to impose those limits.

The empirical analysis presented in this study suggests strongly that the real economic cost of delivering health-insurance benefits under a single-payer system would be substantially greater—at a minimum, roughly double—than that under the current private system. Moreover, the administrative and other net costs of private health-insurance programs are very likely to be efficient in terms of satisfying the preferences of consumers. Such benefits of market institutions should not be discarded lightly.

It is clear that a social consensus—perhaps even near-unanimity—exists with respect to the proposition that to some substantial degree, health-care services ought to be made available to those who cannot afford to pay market prices. Efforts in the private sector to reduce the size and cost of cross-subsidies are efficient economically; but that says only that elimination of such efforts would not yield a free lunch in the form of lower administrative costs without adverse effects. This is not to say that subsidies for the consumption of health-care services necessarily are inefficient; again, such subsidies for, say, the poor are supported widely and thus may be efficient in terms of the preferences of consumers/voters even if delivered through the tax system rather than through private charity. The issue to be addressed is the relative virtues of alternative vehicles with which to deliver such subsidies if it is deemed appropriate to do so. Detailed examination of that question lies outside the scope of this study.

Single-payer systems inexorably must ration care and impose various types of price controls on providers, as the budget pressures attendant upon “free” (or low-cost) health care grows. A deregulated system not tied to employment, on the other hand, would resemble the markets for life insurance or long-term-care insurance: individuals would have powerful incentives to purchase such policies when young, paying actuarially fair premiums, with efficient risk-pooling in the insurance market yielding coverage for whatever level of health-care expenditures for which individual consumers are willing to pay. The problem of the poor can be addressed in such a system in several straightforward ways, among them the provision of vouchers for the purchase of private insurance plans. An alternative approach is the subsidization of private-sector competition in the provision of insurance services, an example of which is the Medicare Part D drug benefit. A reorientation of the current public debate toward such kinds of reform would be salutary.

ABOUT THE AUTHOR

Benjamin Zycher is a senior fellow at Manhattan Institute’s Center for Medical Progress. He is also a member of the advisory board of the quarterly journal Regulation and the advisory councils of Consumer Alert and USA for Innovation.

During the first two years of the Reagan Administration, Dr. Zycher was a senior staff economist at the President’s Council of Economic Advisers. He is also a former senior economist at the RAND Corporation, a former vice president for research at the Milken Institute, and a former member of the Board of Directors of the Western Economic Association International. Previously, he was an adjunct professor of economics at the University of California, Los Angeles and a former editor of the quarterly public-policy journal Jobs & Capital. He holds a Ph.D. in Economics from the University of California Los Angeles (1979) and a master’s degree in Public Policy from the University of California Berkeley (1974).

Dr. Zycher’s research focuses on the economic and political effects of regulation, government spending, taxation, and counterterrorism public expenditures. He has done considerable work on health-care policy and the economics of the pharmaceutical sector and on energy and environmental policy. Dr. Zycher has also examined long-term trends in economic performance and military capability, the use of trade policy in pursuit of foreign policy goals and measures of burdensharing within alliances. He is the author of “Defense Economics” and “OPEC” in The Concise Encyclopedia of Economics (forthcoming in 2007).

Acknowledgements

The author extends his thanks to Michael F. Cannon, Regina Herzlinger, Laurence J. Kotlikoff, and William A Niskanen for reviewing this report.

He would like to thank also the Ewing Marion Kauffman Foundation for its generous financial support.

The views and findings—as well as any errors—contained in this report are solely those of the author.

I. INTRODUCTION

The public debate over prospective reform of the U.S. health-care system has centered on several distinct issues, among them the implications of various reform proposals for costs, both in the aggregate and in terms of the relative burdens to be borne by such specific groups as patients, providers, and taxpayers. At the aggregate level, there remains an important debate over the extent to which given reforms can be predicted to yield efficiencies of one kind or another; in particular, some proponents of a “single-payer” health-care system, an example of which is Medicare in the U.S., argue that substantial cost efficiencies of various kinds would be realized through adoption of a single-payer system for the U.S.

“Administrative” costs are one particular example that has emerged as central in the public debate; some proponents of a single-payer health-care system for the U.S. argue that the administrative cost savings attendant upon such a policy change would be substantial, indeed, sufficiently large to finance health-care insurance coverage for all of the uninsured in the U.S. This study seeks to examine that general assertion, and so has as its central purpose a comparison of such administrative costs for Medicare and for private health insurance in the U.S., using definitions and the available data so as to allow for a full accounting for both systems.

In short, this study focuses most heavily on the following three questions: What are the respective reported administrative costs for private insurance and for Medicare under a full accounting? Would the reported administrative costs savings attendant upon adoption of a single-payer system of health insurance for the U.S. be sufficient to finance insurance coverage of those now without insurance? And: For both private health insurance programs and for Medicare, respectively, what is the total economic cost of delivering a dollar of health insurance benefits?

This study does not address the important distortions in the market for health-care services that government-financed single-payer health insurance, or subsidies for employer-provided private health insurance, inexorably must introduce. These distortions result from the weakening of price signals reflecting resource costs and service benefits, and from the one-size-fits-all homogeneity that single-payer systems inevitably must impose upon populations decidedly heterogeneous. At the same time, a single-payer system would break the link between employment and health insurance—although it is not the only way to do so—and thus would engender some benefits in the form of improved labor mobility. There exists a voluminous literature on such issues, some of which is referenced below; but this study does not address them except in passing when relevant to particular topics addressed in the discussion.

What Are “Administrative” Costs?

The cost of delivering insurance benefits, whether through private insurance or through such public programs as Medicare, essentially is of three types. First, there is the cost of purchasing medical goods and services (or reimbursing insurance beneficiaries for such purchases) from health-care providers. Second, there are the costs of operating the respective insurance systems, that is, of purchasing goods and services that are not “medical” in nature, but that are necessary for the delivery of insurance benefits. A simple example is the cost of management or “administration” narrowly defined, which clearly is not a medical service, but without which a system of insurance could not operate.

And, third, there may be additional costs (or, in economic language, “real resource” costs), whether incurred privately or publicly, necessary for the operation of a given insurance program and for the delivery of insurance benefits to beneficiaries. Such costs may not be “administrative,” whether defined narrowly or broadly; but they would be relevant even if they fail to appear explicitly in the accounting ledgers or budget of a given insurance program, and indeed even if they fail to appear explicitly in any formal budget at all. All that is required is that the given cost be unavoidable (or too costly to avoid) for the delivery of insurance benefits to beneficiaries. One example is the cost of legislative functions for the federal government, which do not appear in the respective budgets of federal spending programs, but which are necessary (at least to some degree) in order for those programs to be implemented and operated under the U.S. system of constitutional government.

For the latter two of these three cost categories, a term perhaps more useful than “administrative” costs might be “non-benefit” costs, that is, any necessary costs of delivering insurance benefits not incurred in the direct acquisition of medical goods and services, that is, medical benefits. Accordingly, “non-benefit” costs as a category include, but are broader than, “administrative” costs, a distinction important to keep in mind.

As discussed in detail below, it clearly is true that the administrative or non-benefit costs reported in the Medicare budget and other parts of the federal budget are lower than the measure of administrative costs used in this study for private health insurance. At the same time, we find the administrative costs of Medicare to be about twice as large as a proportion of total Medicare outlays as commonly asserted, because the administrative costs reported in the Medicare budget do not include the costs of other federal government administrative functions reported in other parts of the federal budget, for which it is reasonable to allocate some share to Medicare.

But that does not mean that a shift to a single-payer system for the U.S. would yield savings in administrative costs sufficient to provide insurance for the uninsured; the discussion below shows clearly that any such assertion is highly problematic. More broadly, a fuller cost accounting—of the total economic costs of a single-payer system—shows that adoption of such a system of health insurance for the U.S. would yield not savings in the total true economic cost of delivering health-insurance benefits, but a substantial increase in such costs instead.

The reason for this is straightforward: The federal government must impose taxes of various kinds to fund the Medicare program as well as other federal programs. Such taxes have adverse economic effects that are unreported, in substantial part because they are difficult to measure; these adverse economic effects are separate from the costs of delivering Medicare benefits that appear either in the Medicare budget or in other parts of the federal budget. This does not mean that government spending programs inherently are not worth their true cost; they may be, or they may be worth funding at lower levels. But it does mean that the true economic cost of a dollar of federal spending is greater than a dollar; and that the true total cost of delivering a dollar of Medicare benefits is greater than that dollar both because of the reported administrative costs of performing the relevant federal functions, and because of the unmeasured economic (non-benefit) cost of the federal taxation without which Medicare benefits could not be delivered. As discussed below, this unmeasured cost is relevant for that part of Medicare spending financed by taxes, excluding that part financed by premiums paid by Medicare beneficiaries.

Government insurance programs engage in such functions as advertising and other forms of marketing far less than is the case for competitive private programs. Government programs do not evaluate the risks (or expected future costs) presented by given applicants or groups. Government programs have far weaker incentives to scrutinize claims and to police fraudulent behavior.[1] And such federal programs as Medicare do not have to deal with the heterogeneous regulatory requirements imposed by the individual states. As also discussed below in detail, the fact that government insurance programs devote few or no resources to such functions does not mean that such private administrative activities are wasteful; but, nonetheless, for these and other reasons, it is to be expected that measured administrative costs for such programs as Medicare are lower than those for private programs, and indeed that is the finding reported below. Unmeasured costs yield a different conclusion.

Organization of This Study

Section II discusses the emergence of the debate over relative administrative costs, and then presents data for the years 2000-2005 on the administrative and other costs of delivering private and public health-insurance benefits, with Medicare used as the representative public program. This analysis offers insights into the question of the administrative cost “savings” yielded by single-payer systems relative to the cost of covering the uninsured, and into the broader question of the economic cost of delivering private and public health-insurance benefits under a full accounting. Section III discusses the conceptual differences between such private and public costs as part of a consideration of a larger question: Why do private plans seeking cost efficiencies in pursuit of greater profits accept nontrivial administrative costs? In simpler terms: Are such costs “wasteful”? And do they serve the interests of consumers? As we will see, the deeper question underlying the issue of relative administrative costs is far more fundamental: Should a health-insurance market be viewed as an institution with which risk-averse individuals and groups can pool risks efficiently? Or should it be viewed as a vehicle with which to redistribute wealth? Section IV discusses some simple economics of health insurance in the context of the current debate over health-care reform in the U.S. Section V offers brief analytic conclusions and policy implications.

II. THE ECONOMIC COST OF DELIVERING HEALTH-INSURANCE BENEFITS

The Emergence of the Administrative Cost Issue

The U.S. Census Bureau estimates that 47 million Americans were without health insurance in 2006. As the uninsured are a central focus of the national debate over reform of the U.S. health-care system, the problem of additional costs—and how to finance them—arises immediately in any discussion of vehicles with which to “cover” the uninsured in the U.S. Perhaps in response, some advocates of a single-payer system for the U.S. more recently have argued that single-payer systems enjoy an important cost advantage over private insurance plans, in the form of administrative costs substantially lower than those observed for the latter.[2] Indeed, as noted above in Section I, the administrative cost savings attendant upon a shift to a single-payer system for the U.S. are purported by some to be sufficiently large that they exceed “the cost of providing full medical care to all of America’s uninsured…”[3]

We turn first to a recent detailed study of health-care costs published by the McKinsey Global Institute, because some advocates of a single-payer system for the U.S. cite figures from that work in support of the proposition that the attendant savings in administrative costs would be sufficient to cover the uninsured.[4] We then turn to our own analysis as described in detail below.

The McKinsey study finds in the context of administrative costs an estimated $98 billion “administrative and insurance gap.” That estimated “gap” in the McKinsey study is not the difference between administrative costs for single-payer and private insurance in the U.S. It is instead the difference between all administrative costs in the U.S. system, both public and private, and a McKinsey-defined hypothetical parameter called “Estimated Spending According to Wealth” (ESAW), an international statistical projection of administrative (and other) health-care spending categories made on the basis of differences in per capita GDP across economies. It is, in other words, the difference between U.S. administrative costs (as defined and estimated in the study) and those of a hypothetical economy overseas with per capita GDP equal to that of the U.S.

Of that $98 billion difference, $14 billion is public, and on the private side includes profits and taxes.[5] Professor Paul Krugman of Princeton University asserts the $98 billion McKinsey estimate to be the “excess administrative costs” of private health insurance relative to a single-payer system, but that redefinition clearly is erroneous.[6] Moreover, Krugman claims that McKinsey estimates that “more than half of the [$98 billion is] accounted for by marketing and underwriting”; but in fact McKinsey estimates $32 billion in “underwriting and marketing expenses” for private insurance.[7]

One problem inherent in the McKinsey study arises with the ESAW analysis of comparative spending across economies: Because single-payer systems tend to be characterized by a broader application of various types of price controls, reported health-care spending will tend to understate the true resource (or opportunity) cost of those programs. The same is true for the rationing that must characterize all single-payer systems: Classes of services denied patients, or classes of patients denied given services, may yield spending savings, but the value of the services not delivered or consumed is a real economic cost not included in reported budgets. To the extent that the services are worth more to patients than their true economic cost, the forgone net value is a real economic cost of single-payer health insurance not reflected in official cost (spending) figures.[8] Krugman, in any event, cites the McKinsey estimate of $77 billion as the cost of providing medical care to the uninsured, concluding that the asserted $98 billion administrative cost savings attendant upon a switch to a single-payer system would be more than sufficient to cover all of the uninsured.[9]

The measurement of health insurance “administrative costs,” of course, can be defined in various ways. The McKinsey analysis, for example, includes as “administrative costs” such obvious categories as underwriting and marketing expenses. It includes as well profits earned and taxes paid by private insurance plans in the U.S., a methodology that is correct in that “profits”—more rigorously, net revenues (or returns) to investors—are the economic cost of attracting capital inputs, just as wages and salaries are the economic cost of attracting labor inputs. Similarly, taxes can be interpreted, perhaps more loosely, as the cost, determined under the public choice processes of direct and indirect democracy, of attracting (or paying for) government “inputs,” that is, services, even if those services bear little relation to the delivery of private insurance benefits.[10]

Government insurance programs, on the other hand, do not earn “profits,” but government still must acquire (or rent) such capital inputs as plant and equipment, the costs of which may or may not be included explicitly in formal budgeting. Similarly, government agencies do not pay taxes, but usually benefit from some array of government services, payment for which, again, may or may not be explicit in budgets. More important, as discussed in detail below, reported budgets are biased downward substantially as a measure of the economic cost of delivering government services, a factor highly relevant for purposes of measuring the true economic cost of government health-insurance programs.

And so it is not only a narrow concept of “administrative costs” that is relevant for purposes of comparing the economic costs of delivering private and government health-insurance benefits. Instead, all such costs are relevant, whether or not they are classified (arbitrarily) as “administrative.” The central purpose of this paper is a comparison of all such costs for private health-insurance programs with those for Medicare.

Direct, Indirect, and Unmeasured Costs of Medicare

As noted above, the economic costs of any health- insurance plan, whether private or public, can be divided into three categories: benefit payments to (or on behalf of) patients, that is, payments for medical services; various costs incurred for administration and other functions complementary with the provision of insurance benefits; and other costs attendant upon the financing and delivery of insurance benefits. For such public programs as Medicare, these latter non-benefit costs may or may not be listed in the budgets of the agencies directly responsible for the given programs; they may be listed in the budgets of other public entities, or they may not be listed (or measured) anywhere at all.

That last case turns out to be quite important: There are substantial economic costs borne by the U.S. economy as a result of the federal tax system that are relevant for purposes of estimating the cost of all federal spending, including that for Medicare, even though such costs do not appear in government budgets. Such costs are unreported because they are difficult to measure; and incentives to measure them may be weak as well. They are analogous to the adverse effects of pollutants emitted during the production of, say, electricity: From the social standpoint, such costs are as relevant as the costs of fuels, as markets and governments make decisions about the appropriate prices and production levels of electricity.[11] Similarly, the adverse GDP effects of taxes are wholly relevant for the evaluation of government spending programs.

Adjustments for Inflation

Because the data used in the analysis described below are for several years, it is necessary to convert (or “deflate”) the annual dollar figures to constant-year figures in order to remove the effects of inflation. The Centers for Medicare and Medicaid Services (CMS), U.S. Department of Health and Human Services, publishes data annually on direct costs for both private health-insurance programs and for government health programs.[12] These data are reported in nominal dollars (not adjusted for inflation); because the production of health insurance by the private sector requires a combination of both health-care resources and other inputs, it is necessary to create a “weighted” deflator capturing the differing inflation rates for health and nonhealth inputs. This is reported in Table 1, as computed for this study from the deflators for medical care and nonmedical services. Also shown is the deflator for federal defense and nondefense expenditures; all of these are used for the purpose of transforming (“deflating”) the direct and indirect data into constant year 2006 dollars.

Private Health Insurance

For private health insurance, CMS publishes data on total premiums—that is, revenues received by insurers from (or on behalf of) policyholders—and on benefit payments, both in total and broken down by subclasses of medical goods and services.[13] Table 2 presents a breakdown of private health-insurance benefits spending for 2000-2005.

One crude measure of the administrative cost of private health insurance is the difference between total premiums and total benefit payments, that is, the net cost of private health insurance. While this difference is likely to include some costs that arguably are not “administrative” in nature[14], use of the total difference between premiums received and benefits paid avoids any issue of downward bias in the measurement of administrative costs for private insurance. Table 3 presents those summary data.

The data reported in Table 3 show that the net cost of private health-insurance programs as a proportion of total premiums increased from about 11 percent in the 2000-2001 period to about 14 percent in the 2004-2005 period. And so, accordingly, that range of 11-14 percent represents our rough estimate of the proportional administrative cost for private health- insurance programs.

Medicare: Direct Reported Administrative Costs

Table 4 presents a breakdown of direct Medicare expenditures as reported in the National Health Expenditures Database, and then deflated.

From Table 4, the direct costs of administrative functions for Medicare, as reported in federal budget data for 2000-2005, are about 3 percent of total Medicare outlays, substantially lower than the 11-14 percent range noted above for private insurance programs.

To some degree, this lower proportional administrative cost characteristic of Medicare results from a measurement bias: Because Medicare beneficiaries are older and thus less healthy as a class than participants in private health-insurance programs, average claims for Medicare are likely to exceed those for private health insurance. For Medicare in 2005, outlays other than for administration were about $345 billion for 42 million enrollees, or about $8200 per enrollee. For private insurance in 2005, with about 165 million enrollees under age 65, benefits were about $621 billion, or about $3765 per enrollee.[15]

If some or most administrative functions display nontrivial scale economies, the rough comparison of administrative cost percentages is biased against the private programs, in that the average cost of administering Medicare claims, because they are likely to be larger than private claims, would be lower as a proportion of claims than is the case for private insurance; this outcome has nothing to do with “efficiency.”[16]

The benefits figures per enrollee noted above suggest that this bias is not likely to be trivial; at the same time, it is hard to believe that it is sufficient in magnitude to make the direct administrative costs of Medicare and private plans comparable as a proportion of outlays or premiums. And so it remains clear that the direct administrative costs of Medicare are lower by some substantial amount (or proportion of total program spending) than those of private insurance programs, although the difference between 3 percent and 11-14 percent for direct costs overstates the true difference, as shown in the next subsection. Whether these additional administrative costs for private programs are “waste” is an issue addressed in Section III.

Medicare: Indirect Reported Administrative Costs

It is clear as well that not all costs relevant for the administration of Medicare—in particular, general government functions and the administration of justice—appear in the Medicare budget; but they are relevant for purposes of computing the cost of delivering Medicare benefits.[17] Table 5 presents these data.

In order to allocate a reasonable proportion of these costs for general government functions to Medicare, we use the share of Medicare spending in total federal outlays, and then apply that proportion to the costs of general government as shown in Table 5. For the administration of justice, we use the share of Medicare spending in total nondefense federal outlays because the defense justice system operates largely outside the federal justice system for civilians. Tables 6 and 7 summarize these calculations.

Medicare: Total Reported Administrative Costs

Table 8 presents the resulting computation of reported direct and indirect non-benefit (“administrative”) costs for Medicare as a proportion of total Medicare outlays.

As noted above, the direct administrative costs of Medicare as reported in the CMS data are about 3 percent of Medicare outlays, roughly one-quarter of the 11-14 percent net cost of private health-insurance premiums.[18] Inclusion of proportional shares of the costs of general federal government functions and the administration of justice increases the non-benefit cost share of Medicare outlays to about 6 percent, or about twice the figure often asserted, and roughly half the figure for private health insurance.[19]

This computation for Medicare is comparable to that estimated by Litow using a different methodology; his estimate for Medicare is a bit over 5 percent for 2003 before adjustment for the average size of claims. With such an adjustment, the Litow estimate for Medicare administrative costs is 6-8 percent.[20] In the short run, of course, general government functions and the administration of justice would continue roughly at current levels even if the Medicare program were to disappear; but in the long run it must be the case that some general functions and some administration of justice by the federal government are driven by Medicare operations. Since there is no a priori reason to believe that Medicare over the long term creates such federal government activity either proportionately larger or smaller than average, the assumption of proportionality incorporated in Tables 6-8 is reasonable.

In short: The reported (direct and indirect) administrative costs for Medicare, at 6 percent of outlays, are about twice the proportion commonly asserted, but still only about half the net cost of private health insurance.

Would the Administrative Cost Savings Cover the Uninsured?

The findings reported above show clearly that the direct and indirect reported administrative costs of Medicare are lower than those for private health insurance programs. (As discussed below, this does not mean that the private administrative costs are “waste.”) Suppose that a single-payer system similar to Medicare were implemented for all Americans: Would the savings in administrative costs be sufficient to provide insurance coverage for those now uninsured?

From Table 3, the net cost of private health insurance in 2005 was $101.4 billion (in year 2006 dollars). Those costs have been rising over time—premiums rose by 7.7 percent in 2006—so let us assume an increase of 5 percent, yielding a figure of $106 billion for 2006, a year for which the Census Bureau reports a population of uninsured Americans of 47 million.[21] At the same time, from Tables 4 and 8 we see that the direct and total (direct and indirect) reported administrative costs for Medicare, respectively, are about 3 percent and 6 percent of total outlays. Some such administrative costs either are fixed or are somewhat invariant with respect to total outlays, while others would rise in proportion to the new public spending, or, in principle, perhaps even more than proportionately. Moreover, because the uninsured disproportionately are a healthy group,[22] the new administrative costs would tend to rise as a proportion of claims.

Accordingly, it is reasonable or even conservative to assume total reported (direct and indirect) administrative costs of 6 percent for the single-payer insurance benefits that would be provided to those now uninsured.

This means that the net saving in administrative costs from adoption of a single-payer system for the U.S. would be not $106 billion (as of 2006), but instead would be about $99.6 billion, that is, 94 percent of $106 billion. For 47 million newly insured, that net saving in administrative costs would yield average insurance benefits of about $2100, in addition to the health-care services that the uninsured consume now. As noted above, Medicare outlays in 2005 other than for administration were about $345 billion for 42 million enrollees, or about $8200 per enrollee. For private insurance in 2005, with about 165 million enrollees under age 65, benefits were about $621 billion, or about $3765 per enrollee; since premiums increased by 7.7 percent in 2006, it is reasonable to assume average benefits of $4055 for 2006.[23]

The uninsured in the U.S. clearly are not deprived of all health care; recent estimates suggest that the uninsured in 2006 consumed $2262 in health-care services on average.[24] Accordingly, the question to be addressed is the magnitude of the increase in health-care consumption by the uninsured were they to be moved into a single-payer system; that such an increase would be observed is incontrovertible. Hadley and Holahan report findings that the uninsured, given insurance coverage, increase their consumption of health-care services by about 75 percent (depending on choices among alternative assumptions).[25] Finkelstein estimates that the expansion of health insurance between 1950 and 1990 explains “about half of the six-fold rise in real per capita health spending” over that time period.[26]

Those empirical findings yield reasonable lower- and upper-bound estimates for the increase in health-care consumption by the uninsured were they to be shifted into a single-payer plan; for a conservative upper-bound assumption, we use half of the Finkelstein finding, or 150 percent. For the lower bound, average consumption would increase by 75 percent from $2262 to about $3959 as of 2006; for the upper bound, the increase would be from $2262 to $5655. Accordingly, the increase in average health-care consumption by those currently uninsured would be in the range of about $1700 to about $3400.

As discussed above, the net savings in reported administrative costs attendant upon a shift to a single-payer system would be about $99.6 billion, or about $2100 for each of the 47 million individuals currently estimated by the Census Bureau as uninsured. For our lower-bound assumption on the increase in health-care consumption by those currently uninsured, this would yield savings ($400 for each of 47 million new enrollees) in reported costs of about $19 billion per year. For our upper-bound assumption, the funding shortfall for reported costs would be about $61 billion ($1300 for each of 47 million new enrollees). Accordingly, a midpoint estimate of the funding shortfall for reported costs would be about $21 billion per year. This estimate is likely to be biased downward because not all of the current health-care consumption by the uninsured is funded by the public sector; moreover, these estimates assume away any increases in the prices of medical goods and services attendant upon an increase in demand engendered by a doubling of the population eligible for Medicare or a similar single-payer program.[27]

In short, the argument that a shift to a single-payer system of health insurance for the U.S. would yield savings in reported administrative costs sufficient to cover all of the uninsured is highly problematic. The next subsection discusses the magnitude and implications of important economic costs that are unmeasured in the context of federally-financed insurance programs, that is, costs in addition to the direct and indirect administrative costs discussed above.

Medicare: Unmeasured Costs

The comparison discussed thus far represents the following conceptual question: If the U.S. were to shift substantial numbers of individuals from the private system of health insurance to a single-payer system similar to Medicare, what change would be observed in total direct and indirect administrative costs as reported by federal agencies and as shown in various parts of the federal budget? In the context of the unmeasured costs of federally-financed health insurance, the conceptual question remains the same: Is there something about federal funding—whether for a single-payer system or for subsidized private insurance—that influences the cost comparison under a full accounting?

The federal government must acquire the resources needed for all of its spending programs through the tax system. Just as administration and oversight functions for both Medicare and private insurance represent real costs of delivering health-insurance benefits, the tax system—again, an institution necessary for the delivery of Medicare benefits—imposes two classes of costs that, similarly, are a real cost of Medicare and, indeed, of all federal programs. The first is the cost of operating the tax revenue system itself; these costs are captured in the general government function discussed above.[28] The second type of cost, however, appears nowhere in government budgets, but is both substantial and necessary for federal revenue operations: It is the real economic cost of the distortions created by the tax system, or the “excess burden” of that system, which takes the form of GDP smaller than otherwise would be the case.

Federal tax instruments are applied to income of various classifications, to transactions, to capital assets, and the like. Those who bear the economic burdens of such taxes attempt, ceteris paribus, to avoid them in whole or in part; and so, particularly in the long run, the taxes affect economic behavior, that is, work effort, saving and investment, transactions, and the like.[29] An obvious example is the (multiple) taxation of capital: By lowering the net returns (or “profitability”) to capital investment, capital taxation reduces that investment. Such distortions have the effect of lowering aggregate output below levels that would prevail in the absence of the taxes; that reduction in aggregate output, however hidden, is termed the “excess burden” of taxation.[30] Gruber notes that:

If there is some action that market participants can undertake to minimize the burden of taxation, they will do so. … [These] attempts to minimize tax burdens have efficiency costs for society. (Emphasis in the original.)[31]

These actions have nothing to do with tax evasion. Instead, because of the tax system, some transactions that would yield net benefits for the economy—work, investment, etc—are avoided, so that the private sector bears a cost greater than a dollar to send a dollar to the federal government. Another example is the excise tax on telephony communications services: fewer resources are devoted to such communications than would be the case in the absence of the excise tax.

This adverse economic effect of various federal tax instruments has been recognized broadly for many years, although there is a range of estimates on the magnitude of the effects. With respect to the income tax, Feldstein has noted that:

The traditional method of analyzing the distorting effects of the income tax greatly underestimates its total deadweight loss as well as the incremental deadweight loss of an increase in income tax rates. … The true deadweight losses are substantially greater than [prior] conventional estimates because the traditional framework ignores the effect of higher income tax rates on tax avoidance through changes in the form of compensation… and through changes in the patterns of consumption…[32]

Note that despite the official use of a payroll tax for purposes of financing Medicare, it is the excess burden imposed by the entire federal tax system that is relevant, because Medicare expenditures are independent of tax revenues yielded by the payroll tax. In effect, the revenues yielded by the payroll tax are deposited, figuratively, in one large federal revenue “pot” used to finance all programs in the unified federal budget, and general revenues yielded by tax instruments other than the payroll tax are deposited in that same “pot” used, in part, to fund Medicare. These taxes necessarily create distortions in economic activity, reducing the aggregate value of economic output as measured by GDP from the level that would prevail in the absence of the taxes.

That excess burden is a real economic cost of all federal spending, including that for Medicare; that is, it is an unavoidable cost of delivering Medicare benefits, and therefore it is appropriate analytically that it be included as a non-benefit (“administrative”) cost of the program.[33] There exists a substantial literature on the magnitude of the economic distortions engendered by the federal tax system; a useful update is provided by Feldstein in two recent papers.[34] Feldstein’s finding, in brief, is that higher marginal tax rates used to finance additional federal spending would impose upon the economy an excess burden of $0.76 per dollar of revenue; that is, it costs the private sector $1.76 (the dollar of tax payments plus $0.76 of economic losses) to send an additional dollar to the federal government, other things held constant.

Because that is a measure of the incremental cost of federal spending, it is reasonable to assume that the average excess burden of existing spending is less than $0.76, because the incremental distortion is likely to rise as spending and tax rates increase. In other words, the taxes needed to fund existing spending impose an excess burden smaller than the taxes needed to fund increased spending. And so we can assume an average excess burden impact of current federal spending smaller than the Feldstein estimate, in order to estimate such existing unmeasured costs, properly, as part of the true cost of Medicare. Moreover, Medicare beneficiaries pay premiums for their participation in the program; such premiums analytically are not taxes but instead are fees paid by program participants. As such, they do not distort economic incentives in the private sector, and so are not relevant for purposes of estimating the excess burden of the taxes needed to finance Medicare benefits. Only the tax-financed portion of Medicare, therefore, is relevant for purposes of estimating the unmeasured non-benefit costs of Medicare.

Table 9 shows total Medicare outlays, total premiums, and the net tax-financed portion of Medicare outlays.

Table 10 presents alternative estimates of the unmeasured excess-burden costs of Medicare under various assumptions about the average excess burden of the tax system.

Medicare: Total Non-Benefit Costs

Table 11 presents a summary of the reported and unmeasured costs of delivering Medicare benefits.

Table 12 presents the non-benefit cost figures from Table 11 as a proportion of reported Medicare outlays.

Recall from the discussion above (Table 3) that the net (administrative) cost of private health insurance is about 11-14 percent of total premiums, and that allocation of a proportional share of general government and justice administration costs to Medicare yields a non-benefit (“administrative”) share of Medicare outlays of about 6 percent (Table 8), reported directly and indirectly. But the lowest plausible assumption about the excess burden engendered by the federal tax system—20 percent—raises the true cost of delivering Medicare benefits to about 24-25 percent of Medicare outlays, or about double the net cost of private health insurance. A more realistic assumption—say, 50 percent—raises the true cost of delivering Medicare benefits to about 52 percent of Medicare outlays, or between four and five times the net cost of private health insurance.

This suggests strongly that economic “savings” in administrative costs purportedly attendant upon adoption of a single-payer health-insurance system for the U.S. would be outweighed greatly by the adverse excess burden effects of the taxes needed to finance such a policy shift. Thus would adoption of such a system increase the true economic costs of U.S. health care in ways not measured by the available cost/spending data. To the extent that adoption of a single-payer system would reduce discipline in the consumption of health-care resources—as discussed below, private “administrative” costs have as a central purpose the imposition of precisely that discipline—costs will rise not only as measured by the official data but also fundamentally as a result of the additional excess burden attendant upon an increase in overall federal spending financed through the tax system, unless the increased spending is offset fully with reduced spending in other parts of the federal budget.

III. ARE THE ADMINISTRATIVE COSTS OF PRIVATE INSURANCE PROGRAMS “WASTEFUL”?

One curious aspect of the debate over the administrative costs of private and public health-insurance programs is the absence of some obvious questions: If the administrative costs systematically borne by private insurers fail to yield services for which consumers are willing to pay, why do profit-seeking insurance companies continue to accept them? And why are the purchasers of health-insurance services, among them the largest companies in the world, willing to pay for them? If there exist important scale economies in administration, why do we not observe a long-term decline in the number of insurers accompanied by an increase in their average size?

It is a cornerstone tenet of market institutions—capitalism—that the pursuit of self-interested goals by individuals and businesses yields as well, as if “by an invisible hand,” the advancement of aggregate economic wellbeing.[35] The proper definition of aggregate economic wellbeing lies beyond the scope of this study; but it is incontrovertible that a competitive market for private health insurance, driven by self-interested consumers and profit-seeking insurers, tends to eliminate cross-subsidies among classes of consumers, claims (or coverage) not included in agreed insurance contracts, and costs not justified by the services, savings, and efficiencies that they yield. The incentives of consumers will drive premiums to reflect the risks that particular consumers impose upon the risk-pooling group.

Consider first cross-subsidies among classes of consumers, that is, premiums for one group lower than the administrative and health-care costs that they impose upon the system. That gap between premiums paid and costs imposed must be financed in some way or the insurer will not be able to stay in business. One obvious way is to charge another group(s) premiums that are greater than the costs that they impose upon the system, by an amount sufficient to finance the gap created by the first group. This would represent a cross-subsidy to the first group from others. But such cross-subsidies would prove untenable in a competitive market for insurance services, because competing insurers would offer lower premiums to the group(s) paying more than the costs that they generate. As those consumers opted for the less-expensive policies, the premiums needed to cover the gap for the first group would rise, leading to ever-greater abandonment of the first group by other consumers.

And so market forces provide powerful incentives for insurers to invest resources in underwriting, that is, the evaluation of individuals and groups in terms of the costs that those consumers can be expected to impose upon the system in the form of health-care consumption and administrative costs, and the alignment of premiums with those expected costs. That is why, as a simple example, smokers pay more than nonsmokers for life insurance. Unless regulations or other legal constraints prevent or impede such underwriting efforts, insurers will be forced by market pressures to align premiums with costs, because insurance—by its very nature as a market for the pooling of risks—is not charity, and indeed cannot be that in a world in which consumers pursuing their own interests can be predicted to opt for premiums lower rather than higher, holding the quality of the insurance product constant.

Note that efforts by individuals and groups to avoid the costs of subsidizing others are not a phenomenon limited to the private sector. Such competition for lower costs is a prominent feature of public finance as well, as various groups prefer—strongly—to enjoy increases in their preferred programs at the expense of others’ programs, and to shift the tax burdens necessary for public spending programs onto others. Neither private nor public health insurance in the context of allocating costs is a charitable endeavor.

Moreover, some would-be consumers of health insurance are not insurable for given conditions (or at all), because they have medical conditions that would yield costs higher than any premium that they would be willing to pay.[36] Their attempts to shift their future costs, both high and relatively predictable, onto other insurance consumers is a standard process of adverse selection, that is, a shift of known costs onto others rather than participation in a system of risk-pooling. The evaluation of applicants’ medical conditions, therefore, also is an administrative cost that avoids the creation of cross-subsidies among consumers, cross-subsidies that are not tenable in a competitive market because they do not serve the interests of consumers.

In short, the administrative costs borne by insurers seeking to align premiums with costs are efficient, in the sense that they provide a somewhat subtle service valued by consumers: avoidance of premiums that subsidize others. They are efficient as well because the presence of cross-subsidies would mean that some consumers would face premiums higher than otherwise would be the case, and so could be predicted to purchase less insurance than would be observed in the absence of the cross-subsidies. Because the pooling of risk in insurance markets is efficient, an artificial reduction in the size of that market means that too little insurance is being purchased, that is, that the allocation of risk is inefficient. In other words, to some degree the presence of cross-subsidies means that too little risk-pooling and too much self-insurance would be observed. The administrative costs borne in efforts to avoid such outcomes thus yield greater resource productivity in the insurance market.

Because competitive pressures tend to drive such cross-subsidies toward zero, regulations and policies that impede such competition automatically tend to preserve the cross-subsidies. Constraints on underwriting are one example; another is the current proscription on the interstate sale of health insurance policies, which prevents consumers from avoiding the costs of mandated coverage and underwriting restrictions imposed by state officials.

Insurers also bear substantial administrative expenses in efforts to scrutinize claims, to ensure that the services for which insurance coverage is claimed actually are included in the insurance contract, and that the prices demanded by providers are within the limits specified in contracts with the providers. Claims for the reimbursement of services and prices not included in the contract would create costs not covered by the agreed premiums, and thus would have to be spread among the other policyholders. But just as the case with the underwriting problem discussed above, an insurer failing to scrutinize claims carefully will lose customers to other insurers that devote greater effort to such claims examination. Such competing insurers will be able to charge lower premiums, satisfying a central preference of consumers.

And so, just as the case with underwriting, resources devoted to the examination of claims serve the interests of consumers, and thus cannot be “wasteful” by definition. Yes, consumers prefer that their claims be paid; but they prefer premiums lower rather than higher as well. Competitive pressures lead insurers to balance such conflicting goals in the context of perceived consumer preferences.[37]

More generally, it is difficult to see why insurers systematically would accept “wasteful” administrative costs, that is, costs yielding services the benefits of which do not justify the costs. Advertising and marketing, for example, often are criticized as components of the administrative costs of private health insurers[38]; while it certainly is true that Medicare engages in far less advertising—Medicare, after all, does not face competitors—that does not mean that private advertising is “wasteful.” Apart from the provision of information about the availability and characteristics of a product, advertising, more subtly, provides market incentives for the delivery of promised product quality, functions that clearly serve the interests of consumers.[39] It is far more reasonable to posit instead that profit-seeking insurers would incur costs, whether “administrative” or not, as long as the incremental benefits yielded by those costs are sufficiently high from the viewpoint of consumers, so that consumers are willing to pay those costs.

Government, on the other hand, is an institution that exists explicitly for the purpose of engendering cross-subsidies among groups, whether through the tax/expenditure system or the regulatory mechanism. Single-payer health insurance by its very nature—it accepts all those eligible, and does not base taxes and fees on health status—must create such subsidies, and the tax system prevents competition on the basis of price.[40] Accordingly, the deeper issue inherent in the debate over relative administrative costs is fundamental: Is the health-insurance system to be viewed as a mechanism with which to pool risks efficiently? Or is it more appropriate to view it as an institution through which wealth is to be redistributed under political and regulatory processes?

To say this a bit differently, it is implausible that the tax system allocates tax burdens in proportion to the benefits of government-spending programs, and so implicitly must engender cross-subsidies among interest groups. Moreover, the argument that health care is a “right,” and that, therefore, cross-subsidies are a beneficial outcome of public health-insurance programs is highly problematic because resources are limited always and everywhere, and so even (or, perhaps, especially) government must make choices among the competing demands of patients for “coverage” for the costs of given medical goods and services.[41] Even in a single-payer system financed by government, health care cannot be a “right” because resources are not infinite.

In short: The implicit premise or assumption on the part of many that administrative spending for private health insurance is wasteful is seriously to be questioned.

Efforts in the private sector to reduce the size and cost of cross-subsidies are efficient economically; but that says only that elimination of such efforts would not yield a free lunch in the form of lower administrative costs without adverse effects. The opposite premise seems to have been endorsed implicitly by some advocates of a single-payer system for the U.S. This is not to say that subsidies for the consumption of health-care services necessarily are inefficient; as a matter of social policy, such subsidies for, say, the poor are supported widely and thus may be efficient in terms of the preferences of consumers/voters even if delivered through the tax system rather than through private charity. The point here is that the reported administrative costs of single-payer systems are not as low relative to private plans as often asserted; and the unmeasured costs of tax finance are very substantial. The issue to be addressed—which lies beyond the scope of this study—is the relative virtues of alternative vehicles with which to deliver such subsidies if it is deemed appropriate to do so.

IV. SOME SIMPLE ECONOMICS OF HEALTH INSURANCE

Advocates of health-sector reform in the U.S. fall essentially into two camps: those in support of some variant of a (government) single-payer system, such as a Medicare-type system expanded to all in the U.S., and those favoring mechanisms designed to increase the importance of competitive incentives and market processes among both patients and providers. The former goal can be summarized reasonably as a centralization of health-care finance and thus administration, while the latter can be summarized as a decentralization (or deregulation) of the health-care sector.

Supporters of reforms intended to yield greater competition and decentralization criticize current public policies that undermine discipline in the consumption of health-care resources, and that introduce various rigidities into the insurance market. The most prominent example is the favored tax treatment of employer-provided health insurance, a tax environment that in practice has yielded private insurance policies with small deductibles and co-payments, that is, benefits very likely to be worth less (on the margin) to consumers than their total costs, but that are chosen because of the tax savings enjoyed privately.[42] It is likely as well that this favored tax treatment has yielded insurance covering more medical services than otherwise would have been the case. In any event, the tax subsidies have yielded a system of private health-care finance in which insurance has evolved largely into a de facto system of pre-payment for most health-care consumption rather than risk-pooling for large and expensive adverse health contingencies.[43] The obvious outcomes of such third-party payment are overconsumption, that is, consumption of some services not worth their total cost by consumers paying substantially less than that total cost, and thus sharply rising social costs for health-care services in the aggregate. Moreover, the bias in favor of employer-provided insurance over, say, individual policies not given the same favored tax treatment under the current system has the effect of reducing labor mobility and creating other distortions.

Advocates of a single-payer system—Medicare for all, so to speak—criticize instead the presence (or purported plight) under the current system of many without health insurance, with an implicit (or explicit) assertion that those individuals are priced out of the market, and therefore consume too few medical services. A single-payer system (ostensibly) would “cover” everyone, yielding greater fairness in terms of both those without health insurance and those suffering disproportionately from adverse health conditions, who otherwise may face higher (“unaffordable”) insurance premiums or who might find themselves uninsurable.[44]

Note that the argument that a single-payer system would “cover” everyone is not the same as saying that all services demanded would be “covered” (i.e., approved for public payment), because resources by definition are limited always and everywhere. Just as market processes in general allocate—that is, ration—resources largely through the price mechanism, a single-payer system cannot avoid similar allocation decisions, except that nonprice criteria are likely to prove relatively more important in practice. Such nonprice criteria include waiting lists (“queuing”), exclusion from coverage of a broader range of services or products by government agencies, exclusion of certain classes of patients, such as the elderly, from such given procedures as organ transplants, and the like. Individuals demanding procedures that are not “covered” in effect are rationed out of the market, except to the extent that they choose to pay privately for the excluded goods and services.[45] In short, both market-based insurance systems and single-payer systems must engage in some form of rationing because resources are limited by definition. “Coverage” for medical expenses is not the same as the delivery of health care, and single-payer “coverage” cannot yield enhanced access to actual health-care services unless costs—that is, budget outlays—are allowed to rise substantially. Those rising costs inexorably would lead to rationing in some form; and government agencies have powerful incentives to reduce budget costs by lowering the prices paid to providers for given services, thus yielding a decline in the quantity of such services supplied over time.[46] For both of those reasons, therefore, greater “coverage” in a real sense is the opposite of health care.

No one disputes the basic law of demand: As the perceived price of a good declines, more of it is demanded. Whether third-party payment for health- care services is organized through a subsidized private insurance market or through a government single-payer system, an increase in the consumption of health-care resources—and thus an increase in aggregate costs—is inevitable. This basic finance problem—sharply rising costs engendered by a population growing and/or growing older—for government insurance programs is well recognized, not only for such U.S. programs as Medicare and Medicaid, but emphatically for single-payer systems in other advanced Western economies as well.

V. CONCLUSIONS

Analysis of the relative virtues and vices of single-payer health insurance lie beyond the scope of this study. Instead, the central focus here is on measurement of the true non-benefit costs of private and public health-insurance systems under a full accounting of each; on the issue of whether adoption of a single payer system of health insurance would yield savings in administrative costs sufficient to provide insurance for all the uninsured; and on an examination of the efficiency of the non-benefit costs of private health insurance.

For the private market, non-benefit costs are about 11-14 percent of total premiums, while the direct administrative costs reported for Medicare are about 3 percent of Medicare outlays. A reasonable allocation of a share of outlays for general government functions and for the administration of justice increase direct and indirect administrative—that is, non-benefit—outlays as reported in the federal budget to about 6 percent of Medicare outlays.

A shift to a single-payer system would yield net savings of about $99.6 billion (as of 2006) in reported administrative costs. This would yield about $2100 in resources for additional health-care consumption for each of the 47 million uninsured. But the increase in that consumption can be estimated conservatively in a range of $1700 to $3400; the lower-bound estimate would yield net aggregate savings of about $19 billion per year, while the upper-bound estimate would yield a net aggregate funding shortfall of about $61 billion annually. Accordingly, a midpoint estimate of the funding shortfall for reported costs would be about $21 billion per year.

Because the federal tax system is an institution necessary for financing all federal programs, including Medicare, the unmeasured costs borne by the economy as a result of the tax system are a real cost of federal spending, and the lowest plausible assumption about the magnitude of that excess burden has the effect of raising the non-benefit costs of Medicare to about 24-25 percent of Medicare outlays, or about double the net cost of private health insurance. A more realistic assumption raises the true cost of delivering Medicare benefits to about 52 percent of Medicare outlays, or about four to five times the net cost of private health insurance.

And so the purported savings in true economic costs that would be yielded by a shift to a single-payer system are highly problematic under a full accounting. Moreover, the administrative and other net costs of private health-insurance programs are very likely to be efficient in terms of satisfying the preferences of consumers. Such benefits of market institutions should not be discarded lightly.

At the same time, it is clear that a social consensus—perhaps even near-unanimity—exists with respect to the proposition that to some substantial degree health-care services ought to be made available to those who cannot afford to pay market prices. And this reality returns us to the basic debate noted above, to wit, whether it is a centralization or a deregulation of the health-care system that would further that goal most effectively. Single-payer systems must ration care and impose various types of price controls on providers, as the budget pressures attendant upon “free” (or low-cost) health care inexorably grow. A deregulated system not tied to employment, on the other hand, would resemble the markets for life insurance or long-term-care insurance: Individuals would have powerful incentives to purchase such policies when young, paying actuarially fair premiums, with efficient risk-pooling in the insurance market yielding coverage for whatever level of health-care expenditures for which individual consumers are willing to pay. The problem of the poor can be addressed in such a system in several straightforward ways, among them the provision of vouchers for the purchase of private insurance plans. A reorientation of the current public debate toward that kind of reform would be salutary.

See Jack A. Meyer, Fighting Medicare Fraud, monograph, Taxpayers Against Fraud Education Fund, July 2006. Meyer finds that the marginal return to a dollar of anti-fraud spending is $15; putting aside quibbles about the specific estimate, this suggests strongly that the federal government spends too little on such efforts.
See, e.g., Representative Pete Stark, "Privatizing Drives Up Cost of Administering Medicare," Wall Street Journal (Letters to the Editor), October 11, 2006.
See Paul Krugman, "The Health Care Racket," New York Times, February 16, 2007.
The most prominent example is Krugman, supra., fn. 3. See Carlos Angrisano, et. al., Accounting for the Cost of Health Care in the United States, McKinsey Global Institute, January 2007 (hereinafter "McKinsey"), pp. 70-74.
This inclusion of profits and taxes perhaps is counterintuitive, but is correct analytically; this is discussed briefly below.
See Krugman, supra., fn. 3; and McKinsey, supra., fn. 4, at 70-72.
Note that McKinsey’s dollar estimates are for 2003, while the percentage figures are for 2004. Contrary to Krugman, McKinsey actually estimates that underwriting and marketing expenses ($32 billion for 2003) were 64 percent (in 2004) of private insurers' selling, general operations, and administrative (SG&A) costs of $50 billion, not of the $98 billion "gap." The latter two dollar computations are for 2003; the implicit McKinsey assumption is that the true percentage figures for 2003 are likely to be close to the estimated percentage figures for 2004. See McKinsey, supra., fn. 4, at 72-73, exhibits 61 and 62.
These issues lie beyond the scope of this study; but see fn. 41 below. For a brief further discussion of this point, see Henry J. Aaron, "The Costs of Health Care Administration in the United States and Canada—Questionable Answers to a Questionable Question," New England J. of Medicine, Vol. 349, No. 8, August 21, 2003, pp. 801-803. If we assume the latest Census Bureau estimate of 47 million uninsured Americans, $77 billion would yield average "coverage" spending of about $1640 as opposed to the actual figure of $2262 in health-care consumption by the uninsured in 2006. See fn. 3, and see the discussion below of prospective health-care consumption by those now uninsured. See also U.S. Census Bureau, Income, Poverty, and Health Insurance Coverage in the United States: 2006 at http://www.census.gov/prod/2007pubs/p60-233.pdf.
See Krugman, supra., fn. 3; and McKinsey, supra., fn. 4, at 77-79. Again, as discussed in Section IV, "coverage" is not the same as health care. The McKinsey analysis of the cost of insuring the uninsured is problematic, in part because average U.S. government spending for the newly insured would be unlikely to be the same as that for, say, a similar European population adjusted only for differences in per capita GDP. See Aaron, supra., fn. 8, at 802. Moreover, even apart from the issue of assumptions about the range of covered medical services and financing methods, an increase in the population of "covered" individuals inexorably will yield increased demands and higher prices. This issue is noted in more detail below.
Or, more rigorously, if the demanders of the insurance services have small demands for the public services financed with the given taxes. This paper addresses the tax incidence issue only in passing, and so does not delve into the difference between the nominal statutory burden of a tax (who sends what size check to the government) and the real economic incidence of a tax (who bears what economic burden after all economic adjustments to the tax are made). But the issue of the aggregate economic burden of the tax system is highly relevant, as discussed below.
Note that to a substantial degree the costs of pollution abatement can be measured. But even if no efforts to control such pollution were made, or if the negative value of the pollution were hard to measure, it still would be appropriate to consider the adverse effects of pollution in decisions about electricity output. The analogous argument for the adverse effects of taxes in the context of government spending is straightforward.
See http://www.cms.hhs.gov/NationalHealthExpendData/01_Overview.asp#TopOfPage, "Historical," various links.
See www.cms.hhs.gov/NationalHealthExpendData/downloads/tables.pdf, Table 12. Data for 2006 will not be released until early 2008.
A possible example is the premium taxes imposed by the states, which essentially are excise taxes on insurance premiums; these taxes vary by state, but for the most part lie in a range of 0-4 percent. Source: Council for Affordable Health Insurance, private communication. For the most part, these premium taxes are not earmarked for government services complementary with the production of private health-insurance services; however, see fn. 10 and the associated discussion in the text, supra.
See Mark E. Litow, "Medicare versus Private Health Insurance: The Cost of Administration,” Milliman, monograph, January 6, 2006. See also Merrill Matthews, "Medicare's Hidden Administrative Costs," Council for Affordable Health Insurance, monograph, January 10, 2006.
16 Consider, for example, ordinary administration of budgets or claims examination. The size of the budgets or of average claims is unlikely to increase the attendant administrative costs proportionately; in the limiting, but sometimes reasonable, case, for example, the cost of scrutinizing a claim arguably is independent of the size of the claim. And so claims will impose administrative costs that fall as a proportion of the size of claims as the latter increase.
See Budget of the United States Government, Fiscal Year 2008, Historical Tables, Table 3.2. General government functions appear as category 800, while the administration of justice appears as category 750. The latter category is relevant to this analysis because (purported) Medicare fraud is a substantial problem to which the Departments of Health and Human Services and Justice devote considerable attention.
See Tables 3 and 4.
With respect to the depreciation of federal office space occupied by Medicare administrators, most such space is provided by the General Services Administration, which charges rents based upon commercial rental rates for similar properties. Under the reasonable assumption that such rents are not substantially below the cost of land acquisition and building construction (due, say, to a weak office rental market), these rental rates implicitly would include depreciation charges for the physical office investment. See CMS Financial Report for Fiscal Year 2006, p. 39, at www.cms.hhs.gov/CFOReport. Other property, plant, equipment, and software are depreciated on a straight-line basis, the costs of which are included in Medicare budgets. See, e.g., CMS Financial Report for Fiscal Year 2006, Note 10, p. 52.
See Litow, supra., fn. 15; and Matthews, supra., fn. 15.
See U.S. Census Bureau, supra., fn. 8.
See fn. 44 below.
See Kaiser Family Foundation, Kaiser Daily Health Policy Report, September 12. 2007.
See Merrill Matthews, "Conventional Wisdom in Health Care Reform, and Why Most of It Is Wrong," monograph, Council for Affordable Health Insurance, May 2007. See also Jack Hadley and John Holahan, "How Much Medical Care Do the Uninsured Use, and Who Pays For It?," Health Affairs, Vol. 22 (2003), pp. W3-66-W3-81 (web edition).
Jack Hadley and John Holahan, "Covering the Uninsured: How Much Would It Cost?", Health Affairs, Vol. 22 (June 4, 2003), pp. W3-250-W3-265 (web edition).
Amy Finkelstein, "The Aggregate Effects of Health Insurance: Evidence From the Introduction of Medicare," Quarterly Journal of Economics, Vol. 122, No. 1 (February 2007), pp. 1-37.
Compare the 47 million uninsured with the Medicare enrollee population of 42 million in 2005. At $2262 on average, the currently uninsured consume about $106 billion in health-care services. At $3959 on average, the currently uninsured would consume about $186 billion in health-care services; at $5655 on average, consumption would be about $266 billion. The midpoint is $226 billion. Medicare expenditures other than for administration in 2005 were about $345 billion.
See Table 5.
See fn. 10.
Strictly speaking, the excess burden (or "deadweight loss") is the difference between aggregate output under the existing tax system, and aggregate output under a different system of "lump-sum" taxes that would yield the same revenues without distorting economic activity. Because government output is not worthless, a zero-tax, zero-outlay, zero-excess burden environment in principle might yield aggregate output lower than that observed under the existing tax system even though, again, the excess burden of taxation would be zero.
Jonathan Gruber, Public Finance and Public Policy, New York: Worth Publishers, 2005, p. 547.
See Martin Feldstein, "Tax Avoidance and the Deadweight Loss of the Income Tax," Review of Economics and Statistics, Vol. 81, No. 4, November 1999, pp. 674-680. See also Jon Gruber and Emmanuel Saez, "The Elasticity of Taxable Income: Evidence and Implications," Journal of Public Economics, Vol. 84, No. 1, April 2002, pp. 1-32
See fn. 10 and the accompanying discussion.
Martin Feldstein, "The Effect of Taxes on Efficiency and Growth," National Bureau of Economic Research Working Paper No. 12201, May 2006; and Martin A. Feldstein, "The Effect of Taxes on Efficiency and Growth," Tax Notes, May 8, 2006, pp. 679-684. See also William A. Niskanen, "The Economic Burden of Taxation," in Mark Wynne, Harvey Rosenblum, and Robert Formaini, eds., The Legacy of Milton and Rose Friedman's Free To Choose: Economic Liberalism at the Turn of the 21st Century, Dallas: Federal Reserve Bank of Dallas, 2004; and the Report of the President's Advisory Panel on Federal Tax Reform, November 1, 2005, p. 36, at www.taxreformpanel.gov/final-report/. See also Benjamin Zycher, “A Preliminary Benefit/Cost Framework for Counterterrorism Public Expenditures,” Rand Corporation MR-1693-RC, May 2003.
"… by directing that industry in such a manner as its produce may be of the greatest value, he intends only his own gain, and he is in this, as in many other cases, led by an invisible hand to promote an end which was no part of his intention.” Adam Smith, An Inquiry Into the Nature and Causes of the Wealth of Nations, London: Methuen and Co., Ltd, Edwin Cannan, ed., 5th ed., 1904, Book IV, Ch. 2, paragraph 9.
Consider a patient suffering from cancer, the prognosis for which is uncertain, but who faces a lengthy and expensive course of treatment. Insurance for this type of consumer, if priced at less than the actuarially expected cost, cannot appropriately be classified as participation in the pooling of risk; it is instead an "adverse selection" process in which the consumer simply shifts known future costs onto others. An extreme example would be an attempt to purchase life insurance by an individual knowing that he has only days to live. It may be the case that such individuals are worthy of compassion and even subsidies, but that is not the same as saying that they are entitled to subsidies from other consumers of insurance services. If subsidies are an appropriate vehicle with which to express such compassion as a worthy social goal—a premise less obviously correct than some argue—then it is appropriate to finance them publicly.
Hyman notes in his chapter on "Sloth," bemusedly, that the failure of Medicare administrators to devote more than trivial amounts of resources to claims examination yields “savings” in the administration of the program viewed as a virtue by the advocates of a single-payer system. See David A. Hyman, Medicare Meets Mephistopheles, Washington D.C.: Cato Institute, 2006, pp. 53-56.
See, e.g., Anna Bernasek, “Health Care Problem? Check the American Psyche," New York Times, December 31, 2006. See also Krugman, supra., fn. 3.
See, e.g., Benjamin Klein and Keith B. Leffler, "The Role of Market Forces in Assuring Contractual Performance," Journal of Political Economy, Vol. 89, No. 4 (1981), pp. 615-641.
For an approving discussion of this inescapable outcome, see Paul Krugman, "Edwards Gets It Right," New York Times, February 9, 2007.
See, e.g., Sally C. Pipes, Miracle Cure, San Francisco: Pacific Research Institute, 2004, for examples from the Canadian system. In "Medicine: Who Decides?", New York Times, December 26, 2005, Krugman, arguing in favor of a single-payer system, concedes this explicitly:"… the public sector… sooner or later [would] have to make key decisions about medical treatment." And: "…health care—including the decision about what treatment is provided—[would become] a public responsibility." See also Cynthia Ramsay, "Michael Moore's Sheer Fantasy: Canada's 'World-Class'Health-Care System," Medical Progress Today, June 22, 2007.
Or, at a minimum, deductibles and co-payments smaller than otherwise would be the case. See, e.g., John F. Cogan, R. Glenn Hubbard, and Daniel P. Kessler, Healthy, Wealthy, & Wise, Washington D.C.: AEI Press, 2005; David Gratzer, The Cure, New York: Encounter Books, 2006; Regina Herzlinger, Who Killed Health Care?, New York: McGraw Hill, 2007; Arnold Kling, Crisis of Abundance, Washington D.C.: Cato Institute, 2006; and Pipes, supra., fn. 41.
"Insurance" traditionally is a vehicle with which individuals purchase participation in a large pool for which losses are relatively predictable because of the ordinary laws of large numbers. By accepting a small, predictable loss—the insurance premium, the deductible, and co-payments—in the current time period, a given policyholder avoids large future losses that can be predicted to occur with some nontrivial probability greater than zero. Risk aversion on the part of policyholders enables insurers to charge premiums higher than the losses expected actuarially, so that the costs of administering insurance programs can be recovered. An equivalent way to view traditional health insurance is as a payment by a policyholder during the current time period that yields increased income during future time periods characterized by adverse health events.
See, e.g., Paul Krugman,"Health Economics 101," New York Times, November 14, 2005. Note that the uninsured in the U.S. are hardly without access to health-care services. About 45 percent are uninsured for six months or less, as for example during the interval between successive employers, and many opt to go without health insurance because of its cost and/or because of a low perceived likelihood that an adverse health condition will occur over, say, the ensuing year. About 57 percent are under age 35, clearly a group disproportionately healthy. Recent analysis finds that the uninsured in 2006 consumed $2262 in health-care services on average. See J.P. Wieske and Merrill Matthews, "Understanding the Uninsured," monograph, Council for Affordable Health Insurance, 2007. See also Matthews, and Hadley and Holahan, supra, fn. 24.
See Krugman, supra., fn. 41.
There is also likely to occur a reduction in actual supply conditions—a downward shift of the supply schedule—but this is a topic outside the focus of this study.