A paper by Selden and Sing (2008) reminds us of what was at stake 45 years ago, when Emmett Hall recommended universal public medical insurance over private-public alternatives. While focusing exclusively on the United States, it also helps to explain why universal pharmacare is being diverted into that same private-public dead end through public "catastrophic" coverage. Governments finance, through many different programs, most US health expenditure. Spending programs - Medicaid, Medicare and others - primarily benefit the unhealthy and unwealthy. However, benefits of the largest program, the tax exemption for private insurance, are heavily tilted towards the highest incomes and are essentially unrelated to health. This pattern (also found in Canada) may help explain political support for private insurance, despite its excessive administrative cost and inability to cover those in greatest need. One pictures paleontologists wandering the wilds of the Gobi Desert, or grubbing in the walls of the Olduvai Gorge or the Red Deer Valley. Yet searching in the musty basements of museums, among specimens collected long before, can also yield important discoveries, or re-discoveries. The analogy was brought to mind by a paper last summer in Health Affairs (Selden and Sing 2008) and by re-reading Barer's (2005) Hall Memorial Lecture.

Selden and Sing, while focusing exclusively on the United States, nevertheless shed light both on the "old bones" of the Hall Commission report, and on the decades-long resilience of debates over private health insurance. They also contribute to explaining why pharmacare in Canada has been driven into the blind alley of "catastrophic" (sic) coverage.

The federal and state governments in the United States spend a great deal of money, directly and indirectly, on healthcare. Despite the general impression that the American healthcare system is "private," the public sector covers more than half the total bill. But who are the beneficiaries of this massive public spending? The funds flow through a number of different channels, some openly reported in public accounts and others more hidden from view. The benefits from the different channels are distributed very differently across the population.

Selden and Sing estimate the distribution of the benefits of each form of public expenditure on healthcare across the (civilian, non-institutionalized) US population. They augment data from the Medical Expenditure Panel Survey (MEPS) with the National Health Expenditure Accounts and the TAXSIM simulation model from the National Bureau of Economic Research. Linking sources yielded a set of 70,099 individual observations.

Table 1 shows the average level of support received from the major American public programs by persons in each of four income classes. For the whole group studied, the average benefit received from public sources was $2,612 per person in 2002. This amounted to an estimated 56.1% of the group's healthcare expenditures.

Table 1. Estimates of per capita public spending on healthcare, by family income and insurance coverage, for the US civilian, non-institutionalized population, 2002
Population subgroup Medicaid/SCHIP Medicare Other public Total Tax expenditure
All $561 $651 $655 $1,867 $745
Family income (relative to federal poverty level):
Below poverty $2,064 $794 $1,121 $3,979 $102
100%-199% $961 $1,052 $818 $2,831 $348
200%-399% $311 $596 $591 $1,498 $716
400%+ $74 $455 $474 $1,022 $1,177
Source: Selden and Sing 2008, Exhibit 4.

Medicaid is a state-based program for persons with low incomes; the federal government contributes financial support but the individual states set criteria for eligibility and levels of support within federal guidelines. (SCHIP, the State Children's Health Insurance Program, is intended to supplement Medicaid with additional benefits for low-income children.) Medicare is a federally funded and administered program for those 65 and over, plus certain special categories in the non-elderly population. "Other" covers a wide range of public programs, individually small but large in total.

Of particular importance, however, are the "tax expenditures" or tax preferences, the favourable tax treatment of particular classes of "private" expenditure on healthcare. These represent public revenue forgone rather than direct program expenditure. They are just as much a cost to government as direct program expenditures, even though they do not show up in the public accounts and must be estimated independently. Selden and Sing estimate these tax expenditures at $214.8 billion in 2002, or 28.5% of the total of $752.9 billion in public contributions. The forgone revenue through tax expenditures is thus greater than expenditures on either Medicare or Medicaid.

The largest single component of these tax expenditures, $147.9 billion, arises from the fact that the premiums for private health insurance purchased by an employer on behalf of employees are a deductible expense for the employer, but are not taxed in the hands of the employee. This creates a powerful incentive for both employers and employees to negotiate and maintain private insurance plans.

As one might expect, Medicaid expenditures are primarily on behalf of the poor. People whose family income was below the federally established poverty line received an estimated average of $2,064 in public benefits; the amount drops sharply as incomes rise until those at or above 400% of the poverty line received, on average, only $74.

Medicare benefits, on the other hand, have quite a different pattern. The criterion for eligibility is age, not income, and the benefits actually peak for those between 100% and 199% of the poverty line. Overall, though, the public expenditure programs have a very pronounced tilt in favour of those at the lowest incomes. Average benefits shrink from $3,979 at the bottom to $1,002 at the top.

The pattern for the tax expenditures is exactly the reverse. Benefits rise sharply with income, from $102 per person below the poverty line to $1,177 for those in the highest income class. The private insurance system thus provides a highly regressive form of public benefit, serving significantly to reduce the overall progressivity of the public financing programs. It reduces the transfer of income "from people who have earned it ... to people who haven't," in Conrad Black's memorable, if gratuitously pejorative, description.

Moreover, the tax expenditures have the further "advantage" that these sums are not open to direct scrutiny in the public accounts. Estimating the value of these benefits requires considerable research effort, let alone allocating them by the income class or other characteristics of the recipients, and the results are always contestable. This pattern thus confirms the insight of a senior Canadian bureaucrat, who noted that programs primarily benefiting the poor are typically overt, while those primarily benefiting the rich are covert. (Well, they would be, wouldn't they?)

Nor is the mitigation of egalitarianism confined to the distribution of benefits by income class. Table 2 shows the distribution of estimated benefits according to the self-reported general health status of those studied. All the public programs are very heavily tilted in favour of the less healthy - as one would expect. Sick people need and use a lot more healthcare, and the various public programs are put in place to help them pay for it.

The tax expenditures would seem to have some other purpose. Their traditional justification was that tax expenditures (by encouraging private insurance) help people get care they need but might not otherwise be able to afford. But tax expenditures assist the sick by subsidizing the healthy - feeding the horses in order to feed the birds.

The value of the public subsidy actually rises slowly as self-reported health status improves, although it drops off for those reporting excellent health. With these subsidies included, the public sector supports an estimated 44.6% of spending for the healthiest Americans; if they were excluded, public sources would cover only 21.4%. The effect on the distribution by income is very similar. Those with the highest incomes have 45.8% of their healthcare costs covered from public sources; remove the tax expenditure subsidies and the proportion falls to 21.1%. By contrast, the contribution of the tax expenditure subsidies to the coverage of the poorest and sickest is negligible. The private health insurance system thus provides a channel for flowing a very significant amount of public money to the healthy and wealthy.

Table 2. Estimates of per capita public spending on healthcare, by health status, for the US civilian, non-institutionalized population, 2002
Population subgroup Medicaid/SCHIP Medicare Other public Total Tax expenditure
All $561 $651 $655 $1,867 $745
Self-reported general health
Excellent $161 $127 $326 $615 $664
Very good $249 $284 $507 $1,040 $794
Good $550 $720 $701 $1,971 $785
Fair $1,876 $2,155 $1,386 $5,417 $778
Poor $4,617 $5,170 $3,257 $13,044 $726
Source: Selden and Sing 2008, Exhibit 3.

These findings are not entirely new. Students of American healthcare have long understood that it is primarily funded by the public sector. Fox and Fronstin (2000) and Woolhandler and Himmelstein (2002) estimated the contribution of direct and indirect public sources in the United States as nearly 60% of the total. Sheils and Haught (2004), in the course of estimating the size of the tax expenditure subsidy for 2004 ($188.5 billion by their method), also estimated its distribution by income class in that year.

Sheils and Haught used a finer breakdown than Selden and Sing, with eight income classes. Tax expenditure benefits continued to increase with family income into ranges well above four times the poverty line. Families with incomes under $10,000 received an average of $102; those with over $100,000 averaged $2,789. These high-income families accounted for about 14% of the population, but received 26.7% of the benefit from tax expenditures. Selden and Sing, however, set the tax preferences in the broader context of public support for healthcare, and permit a much more detailed breakdown of the (estimated) benefits received according to the characteristics of the beneficiaries.

All of which is very interesting, but what does it have to do with Canadians, or anyone else outside the United States? The United States is the world's "odd man out" in its extraordinary reliance on private health insurance. According to the World Health Organization (2008), private prepaid health insurance funded 17.6% of healthcare expenditures worldwide in 2005, compared with 55.9% from governments and 22.5% paid out of pocket. But if one excludes the United States, these percentages change to 6.5%, 62.3% and 28.0% across the remaining 192 countries. The United States accounts for 76.7% of all the private health insurance expenditure in the world. Outside a handful of countries, private insurance makes little or no significant contribution to financing healthcare.

But Canada is one of that handful. We may perceive ourselves as a country characterized by universal public health insurance. Few realize that the WHO places us at number 14 out of 193 countries in the proportion of health expenditures covered by private insurance (12.2% in 2005). A significant proportion of expenditures on both prescription drugs (35.0%) and dental services (52.4%) is financed through private insurance, and that private coverage enjoys exactly the same public tax expenditure subsidies as it does in the United States. And that subsidy is of much greater value to people at higher incomes for exactly the same reasons - they are more likely to have coverage, and they are in higher tax brackets.

Because private coverage does not extend (yet) to hospital care or physicians' services, the subsidies involved are not nearly as impressive as in the United States, and they have attracted very little research. But they are not trivial. Smythe (2001) estimated with 1994 data that the total value of the subsidy for private health insurance in Canada was $2.28 billion; expanding this in proportion to the subsequent growth of private insurance coverage yields $8.1 billion by 2008. Furthermore, Smythe deliberately chose conservative assumptions. Alternative assumptions yielded an estimate of $2.87 billion, or $10.2 billion today.

The "official" estimates are that the public/private split of health expenditures was 70/30 in 2008 (CIHI 2008). Accounting for the tax expenditure subsidy, however, would on Smythe's estimates shift this ratio to 75/25, or on his less conservative assumptions, 76/24. The public sector actually supports a much larger share of Canadian health spending than is indicated in the official figures. But, as in the United States, the covert public spending - revenue forgone - is extremely regressive. In Smythe's 1994 estimates, families with incomes over $80,000 received an average benefit of $225. For those with incomes under $5,000 - fifty cents.

So what does this have to do with Justice Emmett Hall? Well, Hall (like the prime minister who appointed him) was one of a species almost extinct today, a Red Tory. As Barer (2005: 46) notes, Hall "[began] from the very conservative principle that '... community action by the people through their government should be undertaken only when voluntary action leads to lesser objectives or fails to reach essential objectives for sufficient numbers" (Canada 1964: 742) [my emphasis]. But he really did mean both parts of the principle, the Tory and the Red. If voluntary action fails, government should act.

There were, of course, strong voices on the other side. As Barer reminds us, both the Health Insurance Association and the Canadian Medical Association brought forward proposals whose "central feature [was] that the great majority of Canadians could and would become insured through their own means and that the government would need to assist only a relatively small number" (Canada 1964). According to Hall, the commission approached these proposals "sympathetically" and "hopefully." But Hall became convinced, on the basis of the evidence available to him, that "voluntary action [would lead] to lesser objectives or fail to reach essential objectives," and he was therefore driven to his most consequential recommendation - universal public health insurance, administered by governments and financed from taxation.

We now know, of course, that this "central feature" of the alternative proposals was fundamentally wrong. Conveniently, the United States chose, or more accurately drifted into, an insurance system very similar to that which the Canadian Medical Association (CMA) and Canadian insurers had advocated. The results have been available for all to see for nearly 30 years, and continue to grow ever more conclusive.

Private insurance can cover a significant majority of the population. But it covers only about a third of health expenditures, because those with greatest need are excluded. The American elderly are covered relatively well by their federal government; (some of) the poor are covered by more or less mingy state Medicaid programs, and over 15% have no coverage at all. Hall was bang on in worrying about the high cost to government of covering those left out of the private insurance market.

He was equally prescient in emphasizing the high administrative costs of the private insurance system. In this he was decades ahead of most students of healthcare (and, in particular, of most economists). It was left to two Harvard physicians, Steffie Woolhandler and David Himmelstein, to calculate and draw attention to the hundreds of billions of dollars of pure administrative waste generated by private insurance mechanisms. In the process, they have also shown that private health insurers in Canada have even higher administrative overheads than those in the United States. They burden our system less because their scope is more restricted.

It is less clear whether Hall appreciated that a significant proportion of the population would be left entirely uncovered by the CMA and Health Insurance Association proposals, although he did emphasize the administrative difficulties of providing coverage for the residual population left behind by private insurers.

So Hall "got it right." But this raises a couple of related questions. First, why did the private insurers and the CMA get it so badly wrong? And second, why is private insurance back on the table today? In particular, why does pharmacare on the Canadian medicare model - universal, comprehensive, first-dollar- and tax-financed - keep getting pushed off the table? Instead, we have drifted to "Little America," financing pharmaceuticals in the American way and with the same results.

The answer to the first question is, I think, pretty simple. The private insurers didn't get it wrong - follow the money. Advocates of private coverage, supplemented by government subsidies, were pursuing different, and in Hall's terms, "lesser," objectives. Private insurers, in particular, are responsible to and only to their shareholders. Hall's recommendations would, and did, push them out of a lucrative market. They may or may not have understood that their proposals would fail "to reach essential objectives for sufficient numbers"; but that was simply irrelevant, then and now. They certainly understood that their proposals would be much more costly for Canadians. But that was exactly the point; those costs would be their revenues.

The position of the CMA is a little more nuanced, requiring a balance of the economic interests of its members against the well-being of their patients. Many Canadian physicians were genuinely concerned for their patients' access to care, as well as for the economic hardship that payment could impose. But the CMA also calculated that a universal public system would confront physicians with a public payer willing and (to some extent) able to contain their then-escalating share of national income. Private insurers have neither incentive nor capacity to do this; nor do governments that are responsible for paying for only a small "rump" of relatively poor and vulnerable people. As readers of this journal all know, the escalating share of national income devoted to healthcare slowed markedly after medicare was enacted; in the United States, it exploded. The CMA also "got it right," but had other objectives.

Exactly the same pattern of interests has played out in the debates over pharmacare, but this time the private insurers and the pharmaceutical industry appear to have won. Public "catastrophic" coverage, with a high deductible, could remove the embarrassment of the wholly uninsured while leaving plenty of room and market for private insurers under the deductible. It also preserves a fragmented payment system in which the market power of pharmaceutical companies can be fully exploited without meeting any countervailing power from a single public purchaser. High deductible coverage will thus preserve the past trend of higher prices and expenditures for Canadian patients, taxpayers and employers, corresponding to continuing escalation of pharmaceutical industry revenues. It didn't have to be this way; there are other, much better models. But as Brennus said: "Vae victis!" ("Woe to the vanquished!") To the (political) conquerors belong the spoils.

But why has the Canadian public been so thoroughly defeated on this one? Let's return to Selden and Sing. The private insurance system provides a two-pronged mechanism by which high-income people can protect themselves against the potential redistributive impact of a public insurance system. Not only are private premiums unrelated to income, but they attract a public subsidy that actually increases with income. Tax finance requires high-income people to pay more, regardless of their needs and use, but private insurance with public subsidy permits them to pay less, after tax, for the same coverage. And the numbers are big.

The steady growth in income inequality in Canada over the past quarter-century may thus have strengthened a silent "fifth column" in the upper half of the income distribution, a fifth column willing to open the city gates to the private insurance and pharmaceutical industries, and beginning to erode medicare as well. Total costs are higher, a lot higher, in a privately insured environment, owing to massive administrative waste, excessive advertising, misdirected research and fat pharmaceutical profits, but the share borne at the upper end of the income distribution will be much lower, thanks in no small part to the tax expenditure subsidy.

The class war? We lost. Catastrophically.


Un article de Selden et Sing (2008) nous rappelle l'enjeu qui prévalait, il y a 45 ans, alors qu'Emmett Hall recommandait un régime d'assurance maladie public et universel au détriment des options public-privé. Les auteurs du présent article indiquent pourquoi le régime d'assurance médicaments universel est voué à la même impasse public-privé, en raison d'une couverture publique "catastrophique". Au moyen de divers programmes, les gouvernements financent la plupart des dépenses en santé aux États-Unis. Les programmes de dépenses - Medicaid, Medicare et autres - profitent principalement aux moins nantis et à ceux qui sont en moins bonne santé. Toutefois, le plus important programme, soit l'exonération d'impôt pour l'assurance privée, présente des avantages principalement pour les plus nantis, et n'est pas essentiellement lié à la santé. Ce schéma (qu'on retrouve également au Canada) peut expliquer l'appui politique en faveur de l'assurance privée en dépit de ses coûts d'administration excessifs et de son inaptitude à offrir une couverture pour les ceux qui en ont le plus besoin besoins.


Barer, M.L. 2005. "Evidence, Interests and Knowledge Translation: Reflections of an Unrepentant Zombie Chaser." Healthcare Quarterly 8(1): 46-53.

Canada. 1964. Royal Commission on Health Services (Hall Report). Volume 1. Ottawa: Queen's Printer.

Canadian Institute for Health Information (CIHI). 2008. National Health Expenditure Trends in Canada 1975-2008. Ottawa: Author.

Fox, D.M. and P. Fronstin. 2000. "Public Spending for Health Care Approaches 60 Percent." Health Affairs 19(2): 271-73.

Selden, T.M. and M. Sing. 2008. "The Distribution of Public Spending for Health Care in the United States, 2002." Health Affairs DOI 10.1377/hlthaff.27.5, w349-359.

Sheils, J. and R. Haught. 2004 (February 25). "The Cost of Tax-Exempt Health Benefits in 2004." Health Affairs Web Exclusives. Retrieved December 13, 2008. < http://content.healthaffairs.org/cgi/content/ full/hlthaff.w4.106v1/DC1 > .

Smythe, J.G. 2001. "Tax Subsidization of Employer-Provided Health Care Insurance in Canada: Incidence Analysis." Unpublished working paper, Department of Economics, University of Alberta, Edmonton.

Woolhandler, S. and D.U. Himmelstein. 2002. "Paying for National Health Insurance - and Not Getting It." Health Affairs 21(4): 88-98.

World Health Organization (WHO). 2008. World Health Statistics 2008. Geneva: Author. Retrieved December 13, 2008. < www.who.int/whosis/whostat/2008/en/index.html > .