TCS Daily

The Myth of Massive Health Care Waste

By Arnold Kling - March 21, 2005 12:00 AM

"If a president takes office thinking that good policymaking is a simple matter of declaring his desire to...eliminate 'waste, fraud and abuse,' he is ill-prepared for the complexities of the job."
-- Jeffrey Frankel, Harvard University

Economists do not believe that a politician should be taken seriously when he suggests that government spending can be reduced simply by eliminating waste, fraud, and abuse. Instead, we believe that the only way to bring down spending is to make large, politically difficult reductions in major programs.


Similarly, I do not believe that health-care reformers should be taken seriously when they suggest that spending on health care in the United States could be brought down to the levels in other countries simply by eliminating waste. Instead, it is far more likely that bringing down health care spending would require a significant reduction in the quality of health care that Americans receive.


According to those who subscribe to the myth of massive waste in health care spending, the large discrepancy in the share of GDP devoted to health care (15 percent in the United States, compared with less than 10 percent in many other developed countries) reflects the inferiority of our system. They take our higher spending level as irrefutable proof of the inefficiency of our system of private and public financing relative to a more socialized approach.


Instead, I am prepared to make the following bet: ten years from now, it will be objectively clear that the United States provided significantly better health care to its citizens between 1990 and 2005 than did other developed countries. From the vantage point of 2015, the policy blunder of the past fifteen years will not be that the United States spent too much on health care, but that other countries spent too little. The socialized systems, forced to ration health care because tax revenues are not sufficient to pay for state-of-the-art care, are constraining their citizens from being diagnosed and treated as well as Americans.


I am not denying that waste exists. However, I contend that the difference between health care spending in the United States and that in other countries cannot be accounted for by the wasteful items that critics have identified.


The Usual Suspects


To examine the hypothesis that waste accounts for the difference between U.S. health care spending and spending in other countries, let us round up the usual suspects:


  • spending in the last year of life
  • drug company profits and advertising
  • administrative overhead


An urban legend has it that close to half of all health care spending comes in the last year of life. The facts are somewhat different. The most thorough study, by Donald Hoover, et al, finds that 27 percent of Medicare spending takes place during the last year of life.


Overall, 22 percent of health care spending on people over 65 takes place in the last year of life. However, only 1/3 of U.S. health care spending is for people 65 and older. Thus, as a percentage of overall U.S. health care spending, spending on the last year of life amounts to about 7 percent. That is high, but not staggering.


The high rate of spending in the last year of life may not be easy to avoid. When people throw around the claim that much money is wasted on the terminally ill, the implication is that doctors are hovering over dying patients, larding them with futile diagnostic procedures and surgeries. Although some of this takes place, it is not the primary cause of heavy spending in the last year of life.


The most expensive people to care for are invalids. And the simple fact is that people are more likely to be invalids in the last year of life than earlier in life. So, in any population, you will observe health care spending rise as you get closer to the death of the person. This Rand Study by Joanne Lynn and David M. Adamson provides an insightful perspective.


"Improved public health and medical treatments have translated into far fewer deaths from acute causes such as childbirth or infections. Currently, the most common causes of death are heart disease, cancer, stroke, chronic respiratory disease, injury, and diabetes. Dementia and multi-factorial frailty shape the last years of life for a large part of the population...Americans will usually spend two or more of their final years disabled enough to need someone else to help with routine activities of daily living because of chronic illness."


Thus, the 7 percent that we spend on the last year of life is more a reflection of the changing nature of old age, which in turn reflects improved care of acute illnesses. However, even if we emptied the nursing homes and spent nothing on patients in the last year of life, this would still reduce our health care spending as a proportion of GDP by only 1 percentage point, which would keep it several percentage points above the level in other countries. Contrary to myth, the magnitude of what the United States spends on patients in the last year of life is not a factor in our excessive spending relative to other developed countries.


Another usual suspect is the evil pharmaceutical industry. However, total profits of pharmaceutical companies are about one-half of one percent of GDP. In the short run, stringent price regulation could reduce health care spending by perhaps one or two tenths of one percent of GDP. The long run effects of reducing the incentive to develop pharmaceuticals could be adverse, because pharmaceuticals often substitute for more expensive therapies.


Another suspected cause of the high cost of health care is malpractice premiums and defensive medicine. I am in favor of malpractice reform. However, the fact is that such reform will not make much of a dent in our nation's health care budget. Malpractice reform is not a magic bullet, either.


Overhead and Efficiency


A more plausible suspect is administrative overhead in the U.S. health care system. For example, a study by Steffie Woolhandler, et al concluded that, "administration accounted for 31.0 percent of health care expenditures in the United States and 16.7 percent of health care expenditures in Canada." Their view is that switching to a Canadian style national health insurance system would reduce administrative costs.


I wish that economists had been involved in doing the study. My concern is that the measures of administrative cost may be misleading. For example, the authors make much of the fact that the U.S. health care system hired administrative workers at a higher rate than the Canadian system.


However, hiring administrative workers is not by itself an indication of inefficiency. In a fee-for-service system as in the United States, physicians have an incentive to spend their time doing procedures. They will want to off-load as much paperwork as possible to clerical staff. In a different health care system, where physicians are paid something more like a flat salary, two factors are at play. One is that there is less paperwork, which is good. The other factor, however, is that physicians have less incentive to offload paperwork, because spending time doing administrative work themselves will not lower their incomes. More administrative workers could be a symptom of more paperwork, or a more efficient system for handling paperwork, or both. A careful study needs to sort this out.


Switching from a fee-for-service system to a straight-salary system would reduce measured administrative costs. However, it produces a different kind of economic inefficiency. It breaks the connection between work effort and pay for health care providers. Doctors will be paid for showing up, not necessarily for putting in a full day's work.


Reducing administrative overhead is a desirable goal in U.S. health care. However, many factors are involved. The main potential benefit from nationalizing the health care system would come from replacing the fee-for-service model. How much this would save is debatable, but using the most extravagant estimates it is unlikely to reduce our health care spending by more than 2 percent of GDP.


The Real Culprits


While the usual suspects receive attention that is disproportionate to their true impact on U.S. health care spending, two important factors receive relatively little attention: physician compensation; and the utilization of high-tech procedures. Both of these are much higher in the United States than elsewhere.


Physicians are paid more than twice as much in the United States as in other developed countries. Because physician services are about one fourth of all health care spending, we could eliminate one eighth of our health care spending by reducing doctor salaries to the levels of other countries.


The other big factor is utilization of high-tech procedures, such as MRI's, CT scans, and open-heart surgery. If Americans would cut back on the utilization of these procedures, that would reduce health care spending by hundreds of billions of dollars.


The question is whether our medical care would deteriorate if we were to pay our doctors much less while at the same time reducing our utilization of expensive capital resources. It seems reasonable to conjecture that the quality of diagnosis and treatment ultimately would suffer.


Infant Mortality and Longevity


One of the reasons that pundits are fixated on health care costs is that health care benefits are difficult to measure. Over the next decade, I believe that we must and we will develop more useful statistics on health care outcomes and health care quality.


In the meantime, two commonly-used statistics used to compare health care quality across countries are not particularly informative. These are infant mortality and longevity.


Most health care spending is not focused on reducing infant mortality. Infant mortality is very low in all industrial countries, so any differences across countries are of little significance. It is questionable whether differences in measured rates of infant mortality are due to differences in definition. The United States attempts to save most premature infants and counts the failures as infant mortality. In some other countries, deaths of premature infants may be treated for statistical purposes as incomplete pregnancies, which takes them out of their infant mortality statistics.


Longevity calculations are not a sensitive measure of improvement in medical care. In my essay on lifespan, I showed how the longevity number is calculated as a peculiar weighted average of the survival rates for different population cohorts. I produced a simplified example in which the longevity number came out to be 68.9 years. In that example, suppose that 10 percent of the people who otherwise would die at age 60 instead receive treatment that allows them to live at least to age 80, when they die at the rate of other 80 year-olds. In that case, the overall longevity number would increase by less than 1.5 years, to 70.3 years. In international comparisons, such an increase easily could be swamped by other demographic and genetic factors.


Moreover, even if we controlled for other factors, the increase in longevity due to medical treatment will take many years to work its way into the actual longevity calculations. For example, my wife was treated for breast cancer a few years ago. If she had not been treated, she might still be alive today, but she would almost certainly have died by age 55. As of today, therefore, her successful treatment counts for no increase in longevity as it is conventionally calculated. In twenty or thirty years, however, the difference will be quite noticeable (certainly to me).


In another ten or fifteen years, it may be possible to document a significant increase in life extension for people over the age of 55 in the United States compared to what is now occurring in other countries. However, as the example of my wife illustrates, superior medical care will not necessarily show up in the backward-looking statistics that are calculated currently.


Two Hypotheses


There are two hypotheses about the difference in health care spending between the United States and other countries. One hypothesis is that it is due to a higher proportion of waste in the U.S. health care system. The other hypothesis is that Americans are getting better health care.


There is not a shred of doubt in my mind that the U.S. health care system is rife with inefficiency. I will have suggestions for alleviating the problems in future essays.


What I seriously doubt, however, is that the inefficiency of our system so dramatically exceeds that in other health care systems that we have no better quality health care to show for our expenditures. The more carefully one examines the facts pertaining to the arguments used to justify that hypothesis, the less convincing it becomes.


At this point, my personal bet would be that most of what we are getting for our larger expenditure is better health care. All that I know for certain is that more research and better measurement are needed before sober people can draw conclusions.


TCS Daily Archives