Health Care's Muddled Incentives
OCTOBER 23, 2009 by ARNOLD KLING
On the topic of health care, what empirical observations are reliable? Unfortunately, many “facts” come freighted with a great deal of ideological baggage. Those skeptical of markets, who favor a large role for government in health care, tend to emphasize statistics that disparage the American healthcare system. For supporters of markets, it is tempting to try to fight back by finding data that reflects well on American health care.
I think that the best strategy is to arrive at the most accurate understanding, even if it means that America’s healthcare system does not come out looking superior.
This article will summarize empirical observations that I consider important in looking at the U.S. healthcare system. First, I look at international comparisons. Next, I look at the implications of studies that compare health care in different regions within the United States.
Differences Across Countries
People often speak as if the healthcare system is relatively free-market in the United States and much more socialized in other developed countries. However, whether this is the case depends on how one classifies private health insurance.
About 12 percent of personal medical expenditures in the United States are paid out of pocket, which is slightly below that of nearly all other developed countries, including Canada. Where the United States differs significantly from other countries is in how the remaining 85–90 percent of health-care expenditure is financed. About half that remainder is financed by private health insurance, with the other half paid for by government programs, such as Medicare and Medicaid. In other developed countries, healthcare spending is mostly financed out of a government budget.
Private health insurance in the United States, like government health insurance, insulates the consumer from the cost of medical care. Strong regulatory and tax incentives induce businesses to include a large health insurance component in employee compensation. Looking at how private health insurance operates in this country, and why it takes the form it does, I detect considerable influence from government policy. It is hard to tell what kind of insurance, if any, would emerge in a market unaffected by government subsidies, taxes, and regulations.
Overall, I would argue that on the demand side, the United States healthcare system is essentially socialized. The fact is that close to 90 percent of health-care spending is paid for by third parties, which means that individuals in this country generally experience a socialized process for obtaining medical services.
The United States deviates most from socialized systems on the supply side. What I believe is most distinctive about the U.S. healthcare system is that it combines fee-for-service compensation for healthcare providers with fairly unlimited access to medical services. This means that supply is limited neither by rationing nor by absence of compensation nor by any fixed government budget. Notwithstanding considerable regulation of medical practice, the supply of health care is relatively unsocialized.
The combination of socialized demand and relatively unsocialized supply explains one of the key facts about health care, which is that the United States spends far more on health care than other industrialized countries. We spend close to 17 percent of GDP on it, while other countries generally spend just over 10 percent. Because our GDP per capita is also higher than other countries’, the differences in healthcare spending are even more dramatic in terms of absolute dollars per capita. We spend almost twice as much per person on health care as many other industrialized countries.
This additional spending is not inherently problematic. Indeed, one would think that it is a good thing that we can afford to spend so much on medical care. Unfortunately, another fact about international healthcare comparisons is that outcomes differ very little between the United States and other countries. One can find higher survival rates in the United States for people with some forms of cancer, but overall, life expectancy does not appear to be better here.
The combination of higher spending with equivalent health outcomes suggests that the U.S. healthcare system is relatively less cost-effective than that of other countries. One can think of many possible reasons for this.
Perhaps healthcare providers earn higher rents in the United States.
Perhaps private health insurance absorbs much higher overhead than government health insurance.
Perhaps in the United States people make extravagant use of medical procedures that have high costs and low benefits.
In my view, the evidence to support (1) or (2) is very weak. Suppose we were to take away the profits of private health insurance companies and drug companies. Also eliminate the excess of what doctors are paid relative to the median salaries of other workers compared to that ratio in other countries, after adjusting for differences in specialization and hours worked. Overall, the effect on total healthcare spending would be small. I like the way David Goldhill puts it: “For fun, let’s imagine confiscating all the profits of all the famously greedy health-insurance companies. That would pay for four days of health care for all Americans. Let’s add in the profits of the 10 biggest rapacious U.S. drug companies. Another 7 days. Indeed, confiscating all the profits of all American companies, in every industry, wouldn’t cover even five months of our health-care expenses.” (“How American Health Care Killed My Father,” The Atlantic, September 2009.)
On the other hand, one can document dramatic increases over the past 30 years in what I call “premium medicine,” meaning medical procedures that use physical and human capital intensively. Among doctors, the number of specialists per capita has increased sharply. As an example of the use of physical capital, we now do more than 50 million CT scans and 25 million MRIs per year.
We will see below that there is strong evidence that Americans make extravagant use of medical procedures with high costs and low benefits. Thus the rise of premium medicine also coincides with a rise in wasteful spending.
Differences Within the United States
One of the most important findings in healthcare research is that within Medicare there are large regional variations in spending with no difference in outcomes for similar patients. The studies directly measure the amount of medical services received, such as the number of doctor visits per week. Thus there can be no doubt that the quantity and type of medical services consumed, not prices or administrative expenses, account for the large differences in spending across regions. Finally, the studies carefully control for patient characteristics, so that the absence of difference in outcomes is a reliable finding.
The studies of regional differences offer clear and compelling evidence that large variations in the utilization of medical services can be associated with no differences in health outcomes. If this is true across regions within the United States, then, in the absence of some striking piece of evidence to the contrary, I think it is reasonable to take it as true across countries as well. Again, it points to Americans’ use of medical procedures with high costs and low benefits as the best explanation of why the United States spends more on health care than other countries without seeing correspondingly better outcomes.
The differences in Medicare spending across regions are striking. In contrast, the differences between the way American medicine works within Medicare and under private health insurance are small. Relative to other countries, we seem to have the same large discrepancy in spending per person among our elderly population as we do in our younger population. This is consistent with the thesis that the institutional factors that make America different are on the supply side (easier availability of treatments) rather than on the demand side (the financing mechanism). America is an outlier in terms of spending both among people insured privately and among people insured by government.
There is considerable controversy over what the regional differences in health spending—without corresponding differences in outcomes—imply about the value of medical care. At one extreme, George Mason University economist Robin Hanson argues that we could dramatically cut medical care without adverse implications for health (“Cut Medicine in Half,” Cato Unbound, September 10, 2007).
On the other hand, some economists point to the tremendous value of improvements in health and longevity in recent decades. If the value of health improvements over the past 30 years exceeds $1 million per person (as Kevin Murphy and Robert Topel argue), then how can our medical services not be cost effective? (Kevin Murphy and Robert Topel, “The Value of Health and Longevity,” NBER Working Paper No. 11405, June 2005.)
The discrepancy between the way health care looks if we examine differences across regions or over time can be explained in a number of ways. One hypothesis is that the big improvements in health and longevity that we have observed over the years are due to factors largely unrelated to premium medicine: better nutrition, less physical labor, better prenatal and postnatal care, vaccinations, and antibiotics.
Another hypothesis is known as the “flat-of-the-curve” theory. The idea is that medical care offers diminishing returns. The marginal benefit of the most useful treatments is high. However, as one adds more and more spending, the marginal benefit declines. People nonetheless keep undergoing treatments as long as they can hope for any marginal benefit, no matter how small. This means that they consume along the “flat of the curve,” where the marginal benefit is close to zero.
Hanson thinks medical treatment can be harmful as well as helpful. He explains the observation that higher spending is not associated with better outcomes by suggesting that more treatment improves outcomes for some patients but worsens outcomes for others, with a net effect near zero. (“Showing That You Care: The Evolution of Health Altruism,” Medical Hypotheses, Volume 70, Issue 4, 724–742.)
What is puzzling is we know there is considerable research and innovation in medical treatment. We also are confident that at least some of these innovations are successful. For Hanson’s view to be correct, there would have to be as many harmful innovations as helpful ones. One wonders why we do not then observe the market gradually filtering out the harmful innovations and retaining only the successful treatments.
I should point out that in a market system, it would be of no concern to me whether other Americans spend their money on health care wisely or not. Whether you waste your money on ineffective medical treatments or on other consumer goods is of none of my business. Wasteful healthcare spending is a collective problem only because we have collectivized so much of what we spend.
In spite of the large presence of private health-insurance companies, one should view the American healthcare system as largely socialized on the demand side. Healthcare “reform” could entrench this socialized system even more by inducing or requiring more people who are currently uninsured to obtain policies that insulate them from the costs of their medical care. (For more on the difference between insulation and true health insurance, see chapter 5 of my book, Crisis of Abundance: Rethinking How We Pay for Health Care, Cato Institute.)
One wonders what might emerge were the government to refrain from subsidizing or regulating health insurance. If other consumers were like me, then the insurance market would be captured by companies offering policies that pay off only in the case of extreme catastrophic illness. However, if other Americans were like me, we would not have elected leaders who champion so much government involvement in health care.