May 2000 Fraser Forum: Curing Medicare

[Previous][Contents] [Next]

Martin Zelder

Medicare, while intended to embody generosity and compassion, especially for the poor, does not live up to these noble intentions. It does not live up to these intentions because it is defectively designed.

The compassionate ideal, of course, was that medical care should be free at the point of use for all Canadians. This seemed like a good idea at the time it was formulated via the Hospital Insurance and Diagnostic Services Act (1958) and the Medical Care Act (1968), but we've learned a lot since then. What economists and public policy analysts have learned is that "free" doesn't necessarily imply "compassionate."

What does "free" imply? It implies long waits for many vital forms of care - much longer waits than in other, comparable countries. The most recent finding:

• Canadians typically wait 5 months for a cranial MRI; Americans only wait 3 days;
• For knee replacement, Canadians wait 5 months; Americans only wait 3 weeks (Bell, Crystal, Detsky, and Redelmeier, 1998).

Other studies find that Canadians also wait longer than Swedes and Germans for urgent and elective cardiovascular surgery (Carroll, Horn, Soderfeldt, James, and Malmberg (1995), and Collins-Nakai, Huysmans, and Scully (1992)).

And Canadian waiting is getting worse over time; consider table 1. Averaged across all specialties, waiting time has grown 37 percent (3.6 weeks) over the period 1994-1998. Waiting time for chemotherapy grew 44 percent over that same period.

Why do Canadians wait so long for treatment? Their long waits arise from two fundamental defects in the government's intervention in the medical system. These two defective elements are a badly designed insurance plan, and distorted incentives for spending health dollars in government-controlled hospitals.

The underlying problem with the insurance scheme is "moral hazard," or overuse. Economists first acquired a more precise understanding of moral hazard in the 1960s. The idea is that the existence of an insurance contract affects the insured individual's behaviour. More specifically, it means that an insured individual will use insured services to the point where their benefit to her is far below their cost to society.

Because of the slow development of the theory of moral hazard, the founders of medicare didn't realize how serious a problem it might be. As leading medicare historian Malcolm Taylor describes the perspective at the time of Mackenzie King's thwarted 1945 plan: "No one really knew what a comprehensive public health insurance program would cost.... [T]here was no way of knowing what utilization would be if the total population were insured" (Taylor, 1978, p. 14). After the Saskatchewan Hospital Services Plan of 1946 commenced, "The costs of the plan... were considerably in excess of estimates," according to Taylor (Taylor, 1978, p. 104).

What the economists subsequently discovered was that coinsurance payments, or user fees, were a desirable and powerful mechanism to restrain moral hazard. But even after this insight, and the warning signs noted by Taylor, medicare has persisted as an institution without user fees, for the most part.

Why hasn't Canada taken advantage of this growth in social-science knowledge? Perhaps it's because politicians have perennially attacked user fees. And they've been emboldened by medicare's defenders in academia who claim that user fees do not necessarily decrease "overall use" of medical services (Evans, 1994, p. 334).

These claims have been thoroughly refuted by one of the leading social-science research projects of the century, the RAND Health Insurance Experiment (Newhouse, 1993). The RAND researchers randomly assigned about 2,000 randomly-selected nonelderly families to a variety of different insurance plans. The plans varied in terms of their user fee - 0 to 95 percent of the bill - up to a maximum spending limit - 5, 10, or 15 percent of family income (or $1,000 US, whichever was smaller). The families were followed for three to five years in order to understand how assignment to different insurance groups affected health spending and health outcomes.

The RAND researchers found that those families who paid 25 percent out-of-pocket (never exceeding $1,000 per year) incurred annual health care costs, on average, of $826 (US). By comparison, those in the "Canadian" group (0 percent coinsurance rate) incurred annual costs of $1,019 (US). This means that a 25 percent coinsurance rate led to a reduction in annual spending of $193 (US), a 19 percent reduction.

The truly remarkable finding contained in the RAND analysis, however, relates to the change in health status among the families studied. Before-and-after comprehensive measures of health status permitted the RAND researchers to determine whether members of the "Canadian" plan, who received more health care, had better success in improving and maintaining their health than those who paid 25 percent out-of-pocket or more (and received less health care). Extraordinarily, access to "free" health care did not benefit the "Canadians," with very minor, albeit important, exceptions.

The exceptions - those whose health was benefitted by the "Canadian" plan - were the "sick poor," those with low incomes who had high blood pressure, vision problems, dental problems, or anaemia at the start of the experiment. But "free" care had no beneficial health impacts outside of those "sick poor," approximately 6 percent of the population.

And, in fact, when other changes in government spending are taken into account, the Canadian poor may actually have been harmed by free care. Economists Lindsay and Zycher (1984) found that prior to Established Program Financing in 1977, each additional dollar of government health spending crowded out $.31 of social welfare spending. It's seriously debatable whether the small fraction of that dollar of health spending which went to the poor made up for the $.31 cents of welfare spending they lost.

But in the RAND experiment, for the other 94 percent of the population, free care provided no health benefits. This means that much of the additional health services currently consumed due to the zero-coinsurance "accessibility" of medicare are wasted. So, we should provide free care to the poor (without making them worse off by cutting other social programs), but redefine "accessibility" to mean a 25 percent coinsurance rate (up to some modest expenditure limit) for the rest of us.

The fact that there is extensive moral hazard within the current system is reflected in evidence regarding the connection between government health spending and waiting time. Specifically, I analyzed whether provinces in which more public money was spent per person on health care had shorter waiting times (over the period 1992-1998). Controlling for other underlying factors which differed among the provinces, I found that there was no statistical relationship between public health spending per capita and waiting time. In other words, provinces which spent more government money per person had no shorter (and no longer) waiting times than provinces which spent less. This reveals moral hazard in that it reflects money being spent on valueless care, preventing money from being spent where it is valued, and thus increasing waiting times for higher-valued care.

Further analysis I did showed that the money that is being spent isn't being spent in the right places. Specifically, waiting times could be reduced if existing spending were reallocated into spending on drugs and "other professionals."

These findings illustrate the second substantial defect of medicare: the inefficiency of government enterprise. Although ostensibly private non-profit firms, Canadian hospitals are government-controlled, both in terms of overall funding and in terms of resource allocation within hospitals. Consequently, it's not surprising that more spending doesn't reduce waiting, and that existing spending is misallocated. As University of Rochester economist Eric Hanushek commented about the fact that the vast majority of studies find no connection between higher spending and improved student performance in American public schools: "If few incentives exist to reward improved performance, it should not be surprising to find that resources are not systematically used in a fashion that improves performance" (Hanushek, 1996, p. 23). In other words, resources will not tend to be allocated optimally in a system which does not reward optimal allocation. With regard to hospitals, of the 15 studies which compare the efficiency of for-profit and government hospitals, 8 find that for-profits are lower-cost for a given level of quality, while only 3 find the reverse (4 find no difference) (Zelder, 2000).

Ultimately, medicare has proven to be an object lesson in the economics of government failure. Despite abundant evidence that medical insurance should involve a modest coinsurance payment for the non-poor, medicare does not. Despite abundant evidence that government enterprise is inefficient compared to private enterprise, medicare relies on government enterprise. In the end, Duplessis was right in his concern about Mackenzie King's proposal: "Health insurance is dangerous" (Duplessis, 1951). Medicare, designed and perpetuated without adequate social-scientific understanding, endangers Canadian compassion and perhaps even Canadian lives. Economists were slow to understand insurance. Now that they do, it is time to cure medicare.

[This is an edited version of Martin Zelder's testimony before the Senate Standing Committee on Social Affairs, Science and Technology, on April 6, 2000, in Ottawa.]

References

Bell, Chaim M., Matthew Crystal, Allan S. Detsky, and Donald A. Redelmeier (1998). "Shopping Around for Hospital Services." Journal of the American Medical Association 279:1015-1017.

Carroll, Richard J., Susan D. Horn, Bjorn Soderfeldt, Brent C. James, and Lars Malmberg (1995). "International Comparison of Waiting Times for Selected Cardiovascular Procedures." Journal of the American College of Cardiology 25:557-563.

Collins-Nakai, R.L., Huysmans, H.A., and Scully, H.E. (1992). "Task Force 5: Access to Cardiovascular Care: An International Comparison," Journal of the American College of Cardiology 19:1477-1485.

Duplessis, Maurice (1951). Montreal Star. October 5, cited in Taylor (1978).

Evans, Robert G. (1984). Strained Mercy: The Economics of Canadian Health Care. Toronto: Butterworths.

Hanusek, Eric A. (1996). "Measuring Invest- ment in Education." Journal of Economic Perspectives 10:9-30.

Lindsay, Cotton M., and Benjamin Zycher (1984). "Substitution in Public Spending: Who Pays for Canadian National Health Insurance?" Economic Inquiry 22:337-359.

Newhouse, Joseph P. and the Insurance Experiment Group (1993). Free for All? Lessons from the RAND Health Insurance Experiment. Cambridge: Harvard University Press.

Taylor, Malcolm G. (1978). Health Insurance and Canadian Public Policy: The Seven Decisions that Created the Canadian Health Insurance System. Montreal: McGill-Queen's University Press.

Zelder, Martin (2000). How Private Hospital Competition Can Improve Canadian Health Care. Vancouver: The Fraser Institute. u

[Previous][Contents] [Next]

info@fraserinstitute.ca
4th Floor, 1770 Burrard Street, Vancouver, BC, Canada, V6J 3G7
Tel: (604) 688-0221 Fax: (604) 688-8539 Book Orders: 1-800-665-3558 ext. 580

You can contact us at the above email address for any comments or information requests. Please report any dead links or technical problems.