Still here, still flawed, still wrong: the case against the case for taxing the sick

Steven Lewis, MA

CMAJ 1998;159:497-9

Articles in the Controversy section appear in the form of a debate. Mr. Lewis was invited to respond to Dr. Gordon and colleagues' article (page 493). The rebuttal of Dr. Gordon and Stephen Lewis follow on page 500.

Reprint requests to: Steven Lewis, Chief Executive Officer, Health Services Utilization and Research Commission, Box 46, 103 Hospital Dr., Saskatoon SK S7N 0W8; fax 306 655-1462; lewiss@sdh.sk.ca

© 1998 Canadian Medical Association

• Letter: Give it and they will spend
• Funding medicare: Letter 1, Letter 2

Undefended assertions and inconsistent logic

To get our attention, Gordon and colleagues strike an ominous pose: they declare that the principles underlying the Canadian health care system are at risk because the barbarians are at the gate. "It may be impossible for Canada to resist the US market-driven health care industry because of our close economic ties," they state. Follow us or we're doomed. Unless I've missed it, we have had close economic ties with the US for most of our history and have found it eminently possible to resist adopting its health care principles and organization. Somehow, legislation, political will and public vigilance have delivered us (well, most of us) from temptation. If the US system were widely praised, either in the US or elsewhere, perhaps we would be at risk. It isn't, and we aren't.

Second, Gordon and colleagues (a) state that quality has been compromised, citing a newspaper article, not data, as proof, thus establishing that (b) more money is needed in the system, but (c) we should not obtain it through increases in general taxation, because (d) "Canada's tax burden is quite substantial, especially its high marginal income tax rates." All 4 propositions are highly debatable. To address only the fourth, let us concede that comparatively high overall and marginal tax rates often offend both those who pay them and purveyors of business school ethics. For society, they are a problem only if one decides they are a problem, or if evidence suggests that they interfere with economic goals or principles of distributive justice. If a nation decides to support public financing for reasons of equity, efficiency, health or political culture, it will tax its citizens higher than one reflexively wedded to a private-sector approach. Is this a "problem"? Gordon and colleagues assert that it is, just as others assert that minimum wage laws interfere with the moral uplifting of the poor. Neither assertion is proof, or even a case in itself.

But even on their own grounds, Gordon and colleagues contradict themselves. The logic of their case is that taxes are too high now and that we can't raise them further, yet their proposed solution to raise additional revenues is to impose a tax, but only on those already most taxed and only when they're sick enough to require health care.

Third, against the objection that a taxable benefit scheme taxes the sick, the authors respond that any co-insurance system with a deductible for claims is the same thing. Indeed, deductibles and co-payments are also taxes on the sick. If the goal is to avoid taxing sick people when and because they are sick (which the public system does admirably well), it is hard to achieve it by taxing, say, half of them when they're sick. Exempting low-income people and putting a ceiling on annual taxable benefits simply puts a slightly progressive gloss on a grotesque discrimination.

Fourth, early in their article the authors concede that "user fees disproportionately inhibit access to needed care in lower socioeconomic groups." Yet in their Discussion section, they maintain that proposing to tax the sick would have only the positive effect of making patients better decision-makers: keen-eyed adjudicators of competing drugs with a properly diminished sense of entitlement, wisdom reinforced by the prospect of a consumption tax. You can't have it both ways. If a few dollars demonstrably deter utilization among poor people,^1,2 why won't a few hundred dollars of utilization-induced tax deter the not-so-poor? No research has proved that patient decisions improve when simultaneously estimating the impact of a service on health and pocketbook.

Fifth, against the charge that the user-based tax might be seen as shifting some tax liability from the general public to those who use the system, Gordon and colleagues urge us to consider "that general revenues support the infrastructure of the system that is required by all, and the taxable benefit system is directly related to individual utilization." How much of the system do the authors think consists of infrastructure used by all as opposed to services used by individuals? According to the Canadian Institute for Health Information³ 13% of health expenditures in Canada went to general infrastructure (prepayment administration, public health, health research and other) in 1994. There is no way to insulate the sick from the increased liability built into the proposal.

Flawed revenue estimates

Suppose, for a moment, that we overlooked the logical flaws and morally dubious elements of the argument. Have the authors estimated potential revenues accurately? Do their assumptions stand up?

At the heart of the matter is a fundamentally wrong assumption. Gordon and colleagues calculate potential revenue based on the average use of services by age and income group, but they are insensitive to the distribution of service use in any given year. Poor people use more services than others^4,5 — a good thing if we support needs-based utilization. More fundamentally, average per capita service costs tell us nothing about typical use — the mean and the mode are far apart. Evans⁶ has cited US data that estimates 2% of the population consume 41% of the services, and 10% consume 72% in any given year. This distribution is surely plausible in Canada: when you're young, either you need and use very few services or you use a lot, and this is generally true (at higher levels on average) for elderly people as well.^7,8 (A blip in the data occurs among women having babies, meaning that non-exempt families would be hit with a tax bill to accompany their bundles of joy.)

I have recalculated the approximate revenues that could be generated taking into account the realities to which the scheme of Gordon and colleagues is oblivious (Table 1). The total is under $3 billion, or little more than half the $5.55 billion envisioned. Suppose my distribution assumptions are wildly wrong (unlikely) and that the bottom 90% of users consume not 28%, but half again as much — 42% — of services. That would raise the take by another billion at most. And who will pay? The sickest 10% of the taxable group will ante up nearly 30% of the total, averaging $670 per year, while the healthier 90% will consist of many who pay nothing and many who pay far less than their unfortunately ill fellow citizens. And who are the sick? The old, on balance.

But we're not finished yet. Gordon and colleagues assume that it is straightforward to calculate the value of services that individuals consume. No one now cares much whether estimates of the cost of a particular hospital day are based on precise and transparent accounting or are grouped with someone else's day in the same diagnosis-related group or case-mix group or long-term care facility. Enter tax consequences and people will start to care a great deal, creating a demand for the kind of Byzantine accounting apparatus that exists, at staggering expense, in the US. We ought not underestimate the cost of building and managing such systems. If it costs $50 to refine and itemize one's taxable benefit each year, the tab would be $635 million. If it costs $200 — and remember, we're itemizing the tests, the consumables, the days, the visits, with lots of 1-800 numbers and financial officers to explain to (often) elderly people what their 12-page printout means — we're up to a cool $2.54 billion, reducing the net haul dangerously close to zero.

Peculiar incentives

The system proposed by Gordon and colleagues creates some peculiar incentives. Deferring tax is preferable to paying it now; one might delay that surgery from November to January to push back the tax another year. Conversely, if you've surpassed your $2000 limit by April, you might be inclined to "gorge" on elective procedures during the same year since there will be no further tax bite. Finally, if one spouse is very ill and the other is well, those with the option to do so will load income onto the healthy person to ensure that the high service user is exempt from the tax bite. A system that rewards games begets games.

Can't we talk about something else?

In these difficult and volatile times we can ill afford to spend intellectual capital on fatally flawed schemes that replace simple and progressive funding mechanisms with complex and regressive alternatives based on assumptions both empirically false and logically incoherent. If Canada's health care system requires more money — a debatable proposition — progressive taxation gets it fairly and simply. If one is worried (properly) about increased privatization, then fund more services through the tax system because this holds the best prospects for equity and efficiency. The whole point of medicare is to tax universally in order to confer benefits specifically. The earth is not flat, and taxing the sick, nuances aside, is indefensible in a society that takes fairness seriously.

Is there another agenda behind Gordon and colleagues' viewpoint? They have cleverly pitched their piece as an alternative to privatization. In reality it is an alternative to public funding and its non-discriminatory progressive taxation. Remember that a sophism is an argument intending to deceive.

Send a letter to the editor Envoyez une lettre à la rédaction

1. Beck RG, Horne JM. Utilization of publicly insured health services in Saskatchewan before, during and after copayment. Med Care 1980;18(8):787-806.
2. Kupor SA, Liu YC, Lee J, Yoshikawa A. The effect of copayments and income on the utilization of medical care by subscribers to Japan's National Health Insurance System. Int J Health Serv 1995;25(2):295-312.
3. Canadian Institute for Health Information. National health expenditure trends, 1975­1997. Ottawa: The Institute; 1997.
4. Mustard CA, Frohlich N. Socioeconomic status and the health of the population. Med Care 1995;33(12 Suppl):DS43-54.
5. Roos NP, Mustard CA. Variation in health and health care use by socioeconomic status in Winnipeg, Canada: Does the system work well? Yes and no. Milbank Q 1997;75(1):89-111.
6. Evans RG. Going for the gold: the redistributive agenda behind market-based health care reform. J Health Polit Policy Law 1997;22(2):427-65.
7. Roos NP, Shapiro E, Tate R. Does a small minority of elderly account for a majority of health care expenditures? A sixteen-year perspective. Milbank Q 1989;67(3-4):347-69.
8. Cairney J, Arnold R. Social class, health and aging: socioeconomic determinants of self-reported morbidity among the noninstitutionalized elderly in Canada. Can J Public Health 1996;87(3):199-203.