Cato Policy Analysis No. 211 June 23, 1994

Policy Analysis

Why Health Care Costs Too Much

by Stan Liebowitz

Stan Liebowitz is a professor of managerial economics in the Management School of the University of Texas at Dallas.


Executive Summary

Health care costs have increased dramatically over the last few decades and are now thought to be excessively high. That has caused the current political reevaluation of our health care system, including its funding and performance.

This study is an analysis of the causes of the increase in health care costs. The major culprit in the seemingly endless rise in health care costs is found to be the removal of the patient as a major participant in the financial and medical choices that are currently being made by others in the name of the patient.

The increasing share of medical bills paid by third-party payers (insurance companies and governments) and the disastrous consequences are documented. Patients overuse medical resources since those resources appear to be free or almost free. Producers of medical equipment create new and more expensive devices, even if they are of only marginal benefit, since third-party payers create a guaranteed market. Attempts to rein in those costs have led to a blizzard of paperwork but proven ineffective in controlling costs.

The cure for the present problems is straightforward: the patient must once again be made the central actor in the medical marketplace. Patients need to be given the same motivations to economize on medical care that they have to economize in other markets. Tax laws need to be rewritten. The use of medical savings accounts needs to be promoted. High-deductible health insurance should be encouraged.

Returning the patient, and normal market principles, to center stage is all that is necessary to bring the costs of health care under control.

Introduction

One would practically have to be a modern Rip van Winkle not to be aware of the fact that the percentage of the gross national product devoted to health care has been rising for several decades. That fact figures prominently in the claim that health care is devouring too many of America's resources and that, therefore, the health system needs to be overhauled. The infamous growth of medical care relative to GNP is shown in Figure 1.

Fig 1
Source: Health Care Financing Administration

The focus on the share of GNP devoted to health care is somewhat unusual. For example, there does not seem to be any concern over the share of our wealth that is devoted to shoes, or automobiles, or housing. Moreover, there are many products, such as recreational activities, whose share of GNP rises as our wealth increases, yet there is no concomitant clamor to reduce our expenditure on them, as there is on health care.(1) The increasing share of GNP devoted to health care, by itself, is not evidence that the health care market is in need of repair.

More telling are attributes of the health care delivery system that make it inefficient, foremost among which is the

reliance on third parties (insurance companies and the government) to pay most medical costs. In 1990 third parties paid 77 cents of each dollar of medical expense. Because patients pay an average of only 23 cents on each dollar of medical expense, there is only a weak linkage between any consumer's use of medical resources and the payments made by that consumer. When the direct linkage between use of medical facilities and payment is broken, medical consumers lose their incentive to economize on their use of medical resources.

Another factor that usually portends inefficiency in any market is a high degree of government intervention in it, as the extensive literature examining government organizations has demonstrated.

Analysis indicates that our high medical costs are the result of various government policies that have removed patients as purchasers in the medical marketplace. While that state of affairs may be no more than the unlucky result of misguided policies, it is detrimental to the health of medical markets and, if improperly diagnosed, may eventually prove deadly to the literal health of many Americans.

Unfortunately, the proper diagnosis of our medical problems has been obscured by the demonizing of certain components of the medical industry. For example, the Clinton administration has at various times blamed the pharmaceutical industry, medical specialists, and health insurance companies for causing high prices and excessive medical expenditure. Such charges miss the underlying reasons for the current poor health of the medical delivery system, and diminish our ability to repair it. The failure to understand the causes of increased medical costs is apparent also in the Clinton proposal to revamp our health care system, which unabashedly increases our reliance on government and third-party payments.

Several competing proposals, however, have been suggested. Among them are some that adopt, at least in part, the medical savings accounts and tax-law changes proposed by John Goodman and Gerald Musgrave in Patient Power.(2) Central to the Patient Power approach is the weakening of third-party payment mechanisms and the reestablishment of the patient as both the consumer and the purchaser of medical services. By putting consumers back in control of their money, we can restore the vitality of the medical sector.

The Varieties of Excessive Costs

The excessive costs of our current medical system can be classified into three major categories:

The first, and by far the largest excess cost, is due to the current overuse of medical resources by patients. Overuse is the rational response of consumers who do not have to pay the entire cost of the medical services they use. The causes of those excess costs are Medicaid, Medicare, and tax laws that provide incentives for individuals to have their employers purchase their medical care in the form of private health insurance.

The second category of excess cost consists of administrative and paperwork costs that are unnecessary for the provision of health care, but that have come into existence because of the current patchwork of third-party payers and their attempts to control their increasing costs by closely monitoring the behavior of doctors and patients. Even worse is the fact that those cost-containment activities do not seem to have contained costs very well.

The third excess cost is associated with the fear of malpractice suits. Administering medically unnecessary tests and procedures helps to insulate doctors and hospitals from the potential wrath of patients or their families when inevitable accidents occur in medical treatment or when treatments just do not work.

In some sense each of those costs has been brought about by the retreat from a market-based system of medical delivery. The first two of them could have been avoided if patients had been given incentives to make their own choices about medical care. The third cost could have been controlled if the courts had allowed patients and medical providers to use market contracts to detail liability in case of unforeseen accidents.

The Cost from Overusing Medical Resources

Largely ignored in much of the current debate over health care is the excessive use of medical resources by ordinary Americans. No politicians are giving speeches blaming the average citizens of the country for overusing medical care. There are no fireside chats with the president asking citizens to stop seeing doctors so often, asking parents to have their children "tough it out" and not see the doctor for every little scratch, asking the elderly to give up that extra year or two of life. Politicians are not so foolish.

But turning a blind eye to the consumption of medical resources by patients is a mistake. If the country is overusing medical resources, patients must bear responsibility for much of that overuse. We cannot cut our medical expenditures without reducing our consumption of medical resources. Fortunately, we know why patients overuse medical resources, and we know how to solve the problem. Unfortunately, the political will to enact correctives to the problem is not as easily come by, and the current administration in Washington seems to prefer to make empty promises to reduce costs while at the same time increasing medical services.

The concept of "excessive" medical use has a very precise meaning in economic analysis. When the marginal value of the resources used in a medical treatment is greater than the marginal value provided to the patient by the medical treatment, then the medical treatment is classified as "excessive." Note that the economic concept does not require that the medical treatment be without value altogether.

That definition needs to be contrasted with that of the medical community, which typically defines "excessive" treatment as a treatment that is not medically beneficial, as in the claim that cesarean sections are performed in many cases where they serve no positive medical purpose. The medical definition of "excessive" is similar to that of "fraudulent." Patients purportedly accept unneeded treat ment because they are misled by doctors. Yet the economic concept of "excessive" does not require any deceit or fraud at all. It merely requires that patients receive treatment that the patients themselves value at less than the cost of the treatment.

The economic concept of excess use of medical resources is illustrated in Figure 2, which is a version of a simple diagram that can be found in virtually any introductory economics textbook. In Figure 2 medical care is simplified into a single unidimensional concept for the purposes of illustration, but the ideas contained in the diagram are perfectly general and can apply to any particular medical procedure. The downward sloping line represents the value to patients of increasing amounts of medical care. Not all patients have the same value for a given procedure: some patients are not likely to live a useful or productive life, even with treatment; others expect to be able to live many productive years afterwards; still others prefer to preserve resources for their children and forgo treatment. Various persons, therefore, will have different values for identical medical procedures, since the impact of the procedure on their lives will be different. Their differing values for the medical service are arrayed in order, measured by dollars, from highest to lowest, in Figure 2. In the jargon of economics, it is a demand curve.

Fig 2

The upward sloping line represents the value of resources that are used when providing medical services. Doctors, nurses, hospitals, and the other resources currently used in providing medical care could be productively put to use in other activities. Thus, the provision of medical services is a cost to society, in the sense that resources that are used to provide medical care cannot then be used for something else. The measure of the value of lost resources is known as the opportunity cost of producing medical services.

In Figure 2, the cost of providing additional units of medical service is shown by the upward sloping line, which is usually called a supply curve. It is shown to slope upward because it is often (but not necessarily) thought that the resources used first in this market are best suited for medical uses relative to other uses, and those used last are poorly suited to medical uses.

It is a simple matter to determine the optimal quantity of medical services in a diagram such as Figure 2, and students in introductory economics classes have been doing so for decades. The quantity of medical services Q* is the optimal amount of medical service.

That can be understood by examining the implications of other quantities of medical service. For quantities of medical service greater than Q*, a unit of additional medical service is of lower value to patients than is the cost of providing it. In other words, patients would prefer cash equal to the value of the resources used to provide the medical services to receiving the medical services. Thus, it impoverishes patients and society to produce medical services when the recipient of the service would prefer those resources to be used for a different purpose. Similarly, for quantities less than Q*, patients value an additional unit of medical service more than they value the resources used to provide that unit of medical service. Producing the extra unit of medical service would enhance the well-being of patients and society. Thus, if the extra unit is not produced, society is deprived of a potential gain. Therefore, the quantity Q* is the efficient output. At Q*, the net value (value to consumers minus resources used up) of medical services is maximized.

Unfortunately, the current medical system does not induce patients to choose the efficient quantity Q*. Because patients largely have their medical bills paid by third parties, it is rational for them to consume medical services even when the value of those medical services is less than the value of the resources used to provide them.

Third-party payments are of two forms. First, most patients have private health insurance, usually provided by their employers.(3) A typical feature of such insurance is that when insured patients go to doctors, or hospitals, they pay only a small part of the actual cost of the visit, known as a copayment. Second, most patients without private health insurance are covered by government health insurance, either Medicare or Medicaid. Those patients also pay only a portion of the actual costs of the medical resources they use. As a result, there are very few persons who actually pay their entire health care bills out-of-pocket.

Figure 2 can be used to illustrate the situation in which patients pay zero out-of-pocket expense for medical procedures. Although zero out-of-pocket expense is something of an exaggeration (such expenditure is actually 23 percent), that assumption makes the issue easier to understand. In that case, patients have no reason to refuse any medical procedure, no matter how little the value of the procedure might be to the patient.(4) The quantity of medical services that patients will request will be Q1. The extent of the unnecessary medical services is given by the difference between Q1 and Q*. Those excess medical procedures have some value (given by area D), but their value is too low to justify the expense of the procedure.

The unshaded rectangle in Figure 2 represents the expenditures that society would make for medical care if it were provided in a fully functioning marketplace. It is merely the product of the price P* and the quantity Q*. The shaded region represents the excessively high expenditures that occur when third parties pay for all medical care. It is equal to the product of the excess quantity, Q1, and the higher price of medical care, P1, minus the product of P* and Q*.

Some of the excess expenditure goes to sellers of medical services, indicated by areas B and C. The extra revenue going to providers may explain why they have been willing participants in the movement away from consumer payment for medical care.(5) Some of the excess expenditure produces value to consumers, given by area D. But some of the excess expenditure is pure waste, known to economists as deadweight loss, and given by the triangular portion of the shaded area indicated as A.

The excess consumption at a point such as Q1 will likely take the form of excess quality since, in some sense, quality and quantity are interchangeable. Too many hospitals might contain expensive state-of-the-art equipment; too many patients might occupy singleor double-occupancy rooms rather than wards. Overall, the quality of care will be too high, even though there clearly is some value in the additional care. We have chosen a Cadillac of health care systems when a Chevrolet is more in line with our willingness to pay. It is understandable that some commentators are reluctant to characterize the problem of excess quality as a "crisis." Of course, it is not really the quality of health care that is in crisis; it is the financing. Making monthly payments on a Cadillac can seem like a crisis to someone making Chevrolet wages. Too much of an economic good can be as harmful as too little.

The Impact of Third-Party Payment on Medical Spending

Measuring excessive use of a product is a difficult and usually imprecise task. The best that can be hoped for is a crude estimate, and even that will require some rather broad generalizations, such as lumping many disparate medical resources into a single whole.

The analysis consists first of measuring the relationship between third-party payments and changes in the use of medical resources. Then the current use of medical resources is compared to the resources that would have been used if patients had paid for their own health care (Q* in Figure 2). The difference measures the excessive use of medical resources.

Third-party payment mechanisms are now very common, although before World War II individuals generally purchased medical services just like any other economic commodity and paid for them just like any other economic commodity--out of their own pockets. But during the war many companies began to offer medical benefits as a way to avoid price controls and to take advantage of the tax code. As shown in Figure 3, there was an explosion of private health care coverage shortly after World War II.(6)

Fig 3
Source: Based on data from Health Insurance Association of America, 1991 Source Book of Health Insurance Data.

In addition, 1965 was the year in which the government introduced Medicare and Medicaid, which pay for much of the medical care of the elderly and the poor. The combination of the increased use of private health insurance and in creased government payments in the last few decades has reduced the out-of-pocket expenses of consumers dramatically. Figure 4 shows the fall in out-of-pocket expenses since 1960.(7)

After 1960 the fall in out-of-pocket expense was mainly due to increases in government expenditure. Although not shown, there was a significant decrease in out-of-pocket expenses due to increases in private health insurance in the 1940s and 1950s. The overriding conclusion to be drawn from Figure 4 is clear, however. The role of third-party payment has increased significantly in the last few decades.

Some of the most compelling evidence that third-party payments alter the use of medical resources comes from a study performed under the auspices of the RAND Corporation in the late 1970s.(8) That study assigned families to four health insurance plans with differing coinsurance provisions and deductibles. Coinsurance is the percentage of medical bills paid out-of-pocket by the patient. The deductible measures the maximum total dollar amount that a family will pay out-of-pocket before the plan will drop the coinsurance requirement and pick up the entire medical bill. Some families had zero coinsurance, meaning that the plan paid all of their medical bills, while other families had to pay up to 95 percent of the cost of their medical bills, until their bills reached a total deductible level of $1,000 in 1973 dollars, which is the equivalent of approximately $2,850 in today's dollars.(9)

Fig 4
Source: Based on data form Health Insurance Association of America, 1991 Source Book of Health Insurance Data.

The RAND researchers observed how the different coinsurance rates influenced the use of medical resources by 2,500 families for three to five years. They found very pronounced changes in the use of medical resources, depending on the extent of third-party payments. In particular, families with no coinsurance (complete third-party payments) used 53 percent more hospital services (measured in dollars) and 63 percent more visits to doctors, drugs, and the like than did the group that paid 95 percent coinsurance. Overall, the total use of medical resources was 58 percent greater for the group with no coinsurance. Thus, there is clear indication that the use of medical resources by patients varies dramatically with the existence of third-party payment mechanisms.

Figure 5 shows the relative medical expenditures for each of four groups in the experiment. As the share that patients pay drops below 50 percent, the use of medical resources increases dramatically. It is interesting that this experiment did not find increased use of medical resources as the out-of-pocket share dropped from 95 percent to 50 percent. That may mean that consumers do not begin to overuse medical resources seriously until they pay less than half the cost, or it may just be a statistical anomaly, as the authors of the RAND study point out.

Fig 5
Source: Data from RAND health insurance experiment, cited in Joseph Newhouse et al., "Some Interim Results from a Controlled Trial of Cost Sharing in Health Insurance," New England Journal of Medicine, December 17, 1981.

The decrease in use of medical resources by families with high copayments might have been thought to decrease their health. One of the criticisms that has been made of high-deductible, high-copayment medical plans is that they discourage inexpensive preventative medicine, causing higher medical payments down the road. But the RAND study found no significant difference in health outcomes. In addition, a study by Robert Brook and others, reported in the New England Journal of Medicine, concluded that free medical care did not appear to improve the health of the participants.(10) Thus, it appears that the excessive costs associated with excessive use of medical resources do not materially improve the health of those receiving that care, a result that should not be surprising. Most persons are likely to be willing to pay for inexpensive medical services that provide enhanced future health. Moreover, the prevention of disease is most often associated with activities that individuals engage in for their own reasons, and that are not strongly related to visits to doctors (e.g., they stop smoking or they exercise).

It should also be noted that the RAND study was conducted in such a way as to underestimate the impact of third-party payments on total medical expenditures, because the impact of third-party payments in the experiment could not appreciably influence the price of medical resources, since the number of participants in the study was too small a percentage of the market to have influenced market prices. However, if the measured increase in use of medical resources found by the RAND researchers were duplicated throughout the country by millions of patients, as more and more of them switched to third-party payments, the price of medical resources could be expected to rise, and the increase in expenses could be expected to be larger than that found in the RAND experiment.

The RAND experiment is not the only estimate of the response of consumers to medical payments. A large number of other studies conclude that medical consumers do respond to price changes, and the degree of response found is often similar to that reported in the RAND study. There is virtual unanimity in the belief that higher levels of third-party payment will increase the use of medical facilities by patients.(11)

In the RAND study patients responded within a few years to changes in third-party payments. Yet it is likely that, for society as a whole, the complete reaction to changes in third-party payments might take a longer time to work through the system. Once third parties pay for a large share of total costs, technologies that might not have been cost effective when the patient was paying the full cost will be demanded by patients.

A simple analogy can be used to illustrate the impact of third-party payment on the growth rate of medical expenses. If the government told citizens that it would pay 80 percent of the cost of each automobile purchased, most citizens would march right out to their local dealerships and order very expensive cars. Automobile manufacturers, sensing profits in the air, would begin to offer far more standard equipment and would begin to offer more new types of equipment than they had previously. What was formerly a luxury car would become commonplace, and new, more luxurious automobiles would be produced. The newest technologies would be used (rather like those used in jet fighters), since the cost to the consumer would be only a fraction of the actual cost. Thus, the growth in automobile expenditures caused by the third-party payments could go on for many years.

A similar story can be told about the health care industry. Although the RAND experiment indicated that consumers responded quickly to third-party payments, the longer run consequences might continue for decades. It is possible to examine that hypothesis by comparing the growth in expenditures over several decades for various medical products that have considerable variation in the degree of third-party payment. As most persons who have experienced the choices available with different health insurance policies can testify, medical services related to dental and vision care (eyeglasses) and drugs or medical appliances tend to have much higher out-of-pocket expenses than hospital stays or visits to doctors. Figure 6 indicates that major differences exist in the share of out-of-pocket expenses borne by the patient for various categories of care.(12) In 1990 third parties paid virtually all hospital bills (95 percent), making hospitalization essentially a free good for most Americans, and only 20 percent of physicians' bills were paid by patients. On the other hand, 53 percent of dental bills, 74 percent of drug expenses, and 68 percent of eyeglasses bills were paid by patients.(13)

If third-party payments influence the growth of medical expenditures, then the increased use of medical resources in the past few decades should differ for the various types of medical services. That prediction is generally borne out, as shown in Figure 7, which shows the growth in each of the medical sectors, relative to their 1965 amounts, after controlling for the effects of inflation. Thus, the total costs of hospitalization increased more than 350 percent from 1965 to 1990, even after controlling for general inflation. During the same period, physician payments went up almost 250 percent, yet costs for dentists, drugs and appliances, and vision care went up only 150 to 200 percent. At the same time, real GNP went up by 94 percent.(14) It should be no surprise, then, that medical costs are gobbling up larger and larger shares of GNP.(15)

Fig 6
Source: Based on data from Health Insurance Association of America, 1991 Source Book of Health Insurance Data.

Fig 7
Source: Based on data from Health Insurance Association of America, 1991 Source Book of Health Insurance Data.

It is particularly ironic that drug manufacturers have been singled out by the Clinton administration as being responsible for the spiraling costs of health care in light of the fact that the growth in drug expenditures is far less than the growth in overall medical costs, particularly hospitalization. The relationship between growth in expenditure and out-of-pocket payment is more clearly seen in Figure 8. For five categories of medical care, the share of costs paid by patients is related to the growth in real expenditure over a 25-year period. The relationship is as expected: medical categories with low levels of third-party payments (high out-of-pocket expense) had the smallest increase in total expenditures. Although there are only five data points--and the small number of observations requires that we be cautious in trying to generalize the results--it is still noteworthy that the results indicate a powerful relationship between the level of coinsurance and the growth of medical expenditures.

Fig 8

The three medical categories having a relatively small third-party payments--dental services, drug products, and vision products--also show growth rates that are not a great deal higher than the overall growth of GNP (194 percent). The two categories with the greatest level of third-party payments experienced the greatest growth. The rank correlation between the growth of medical costs and the share of the medical bill paid out-of-pocket is perfect.

The line drawn through the points is a linear regression line.(16) It is almost a perfect fit through the points, and its interpretation is straightforward: the larger the share of medical expense paid by the patient, the smaller the growth in expenditure on that medical product.(17) Extrapolating the line to a point where patients pay completely for medical services would lead to the conclusion that medical services would grow at a rate only slightly greater than the overall growth in GNP.(18) Table 1 gives the expected share of GNP devoted to medical care in 1990 for various levels of third-party payments, based on the results shown in Figure 8.

Table 1
Medical Care's 1990 Share of GNP for Different
Third-Party Payments (1965 base)
If Patients Had Paid (%) Percentage of GNP in 1990 Would Have Been
100 6.8
75 8.7
50 10.6
25 12.6
0 14.5

Thus, the evidence indicates that if the effect of third-party payment had been eliminated, the growth in medical expense would have been much smaller than it actually has been. The reality is that medical expenditures have risen from 4.4 percent of GNP in 1950 to 12.2 percent in 1990 to over 14 percent today. That increase has occurred under a regime of increasing third-party payments. Yet without third-party payments, the growth rate of medical care would have been much smaller, and the "crisis" in health care would not have been a crisis at all.

But even the figures in Table 1 estimating the importance of medical care under regimes of low third-party payments will, to some extent, overestimate the importance of medical care as a percentage of GNP. That is because the base year, 1965, was already severely tainted by the influence of third-party payments, and thus the level of medical spending was already significantly higher than it would have been had third-party payments not been as high as they were.

Finally, it is disconcerting to note that two of the three categories that have experienced the smallest increase in total expenditure--dental and vision products--are going to be brought under the umbrella of third-party payments in the proposed Clinton health plan, a policy that will ensure that our current problems will get worse. Instead of trying to duplicate the relatively good performance of dentistry, eye care, and drugs in the relatively profligate categories of hospitalization and physician payments, the Clinton administration appears determined to impose the egregious performance of hospitalization and physician expenses on the few areas not currently suffering from an explosion in costs.

Evaluating Excessive Output

The historical evidence just examined indicates that with no third-party payments, the medical bill for the nation would be less than 7 percent of GNP instead of the current level of 14 percent. Stated another way, current spending is approximately double the level it would have been if third-party payments had not existed. However, since insurance for calamitous medical bills is valuable, so is some level of third-party payment. Assuming that the alternative to the current system will still leave thirdparty payments in the vicinity of 25 percent implies, based on Table 1, that the share of GNP devoted to medical care would be in the range of 8 to 9 percent. In dollar terms, that translates into a conclusion that for 1992, under a system with third-party payments in the vicinity of 25 percent, medical spending would have been approximately $300 billion less than the actual payments.(19) That is not to say that the excessive $300 billion provides no value, but that it provides less value than cost and would not have been spent if patients had been making the financial decisions.(20)

Although that estimate of excessive expenditure may seem like a fairly enormous sum, it is actually quite conservative. Other analyses in the literature provide a much larger estimate of the increased use of medical resources. Martin Feldstein estimates that for hospital care, the largest single component of health care, the increase in expense that would be caused by a change from complete out of-pocket expenses to complete third-party payments might be as high as 250 percent.(21)

Even the RAND study, which provided an underestimate of the impact of third-party payment, concluded that virtually complete third-party payments would increase medical costs by at least 60 percent relative to what they would have been with much lower third-party payments, a result not far from that found in the historical data.

Excessive Costs of Monitoring

Much of the public debate over health care centers on the amount of paperwork that is required. Hospitals and doctors fill out a plethora of forms for health insurance companies and for the government. But in fact the paperwork (administrative) costs of the current system are not the largest unnecessary costs in our medical system.

It is possible to gauge total administrative costs by focusing on the administrative costs of health insurance, the component of administrative costs that appears to be most precisely measured. Comparing those costs to the other costs of the health care system (Figure 9) makes it clear that the measured administrative costs of running health insurance companies are not a large proportion of the total. Indeed, in 1990 they came to 5.81 percent of the total cost of our health care system.(22)

However, health insurance administrative costs are only a part of the true administrative costs of the current system. After all, hospitals and physicians have enormous amounts of paperwork, much of which they send to the health insurance companies. Yet only the costs to the insurance companies are included in Figure 9. Still, the administrative costs of running health insurance companies should mirror the costs that hospitals and physicians incur, since the forms go back and forth between those parties. If so, then the growth of one category of administrative costs will reflect the growth in other categories of administrative costs.

As a first approximation, administrative costs could be expected to grow at about the same rate as other medical costs, since some administration is necessary. If administrative costs are excessively high, and if the excess has not been in the health system from the beginning, then we should find that administrative costs have increased by more than other medical costs. Figure 10 compares the growth of total medical costs with the growth in the cost of administering private health insurance since 1965.

Fig 9
Source: Based on data from Health Insurance Association of America, 1991 Source Book of Health Insurance Data.

Fig 10
Source: Based on data from Health Insurance Association of America, 1991 Source Book of Health Insurance Data.

Figure 10 indicates that administrative costs increased slightly less than overall medical costs from 1965 to 1975, but that since 1975 they have grown very rapidly.(23) Thus, there is some evidence to indicate that administrative costs might be too high. It is important to understand just why those costs might have started to grow so rapidly after 1975.

The most likely explanation seems to be the emergence of Professional Standards Review Organizations (PSROs) in 1972 and Professional Review Organizations (PROs) 10 years later. Those organizations are privately contracted agents of the government that review the decisions made by doctors and other health professionals, purportedly to save taxpayers money on Medicare and Medicaid cases by eliminating unnecessary or wasteful expenditures. Since private health insurance companies act as fiscal intermediaries for the government's Medicare program, the reviews are bound to affect their costs as well. The flip side is Utilization Review (UR), a system very similar to PROs, in that private health insurance companies hire third parties to review the behavior of doctors. All three systems give doctors incen tives to document all aspects of care, since otherwise they might not be compensated by the third-party payer.

If the entire difference in the growth of administrative costs and other medical costs since 1965 were taken to be excess administrative costs, then 50 percent of current administrative costs would be excessive. And it is inter esting that there is no evidence that the extra administrative costs have lowered overall medical costs, so that the supposition that the new administrative costs do not provide any value seems plausible. Of course, the true test of whether the additional administrative costs are worthwhile requires comparing actual medical costs with what the costs would have been without the additional administrative costs, a test that I do not attempt here, nor am I aware of any such calculation by others. Since administrative costs for health insurance were somewhat less than $50 billion in 1990, and doctors and hospitals must duplicate those costs, we can conservatively assume that total administrative costs are at least $100 billion. Then, if the costs were 50 percent too high, the excessive administrative costs would be $33 billion.

There are many estimates of the excess costs of administering health care, as might be expected given the difficulty in measuring them.(24) It has been claimed that current administrative costs are twice as high as they should be, and that as much as 10 percent of medical expenditure is excess administrative costs. Still, even those estimates indicate that excessive administrative costs are small ($83 billion in 1992) compared to the excessive use of medical resources due to third-party payments.

The current Clinton heath plan claims that there will be large savings in administrative costs and that those savings will help to cover the cost of health insurance for some 37 million Americans who are thought not to have health insurance at any moment.(25) But those cost savings are predicated on there being fewer forms to be filled out, since there will be fewer insurers. But if the additional administrative costs are due to wasteful utilization reviews by third-party payers, there would be no reason to expect administrative costs to fall, given that the Clinton plan expands the role of third-party payers. The Clinton plan also adds entire new layers of government bureaucracy, which, if history is any guide, seems most unlikely to reduce overall administrative waste.

How Patient Power Lowers Health Care Costs

The Patient Power plan avoids excessive costs--both those associated with excessive use of medical care and those associated with excessive administrative burdens. It reduces medical expenditures by giving patients an incentive to use medical care efficiently, rather than overusing it. To that end, tax laws would be altered. The tax break extended for the purchase of health coverage would be allowed only for basic, no-frills catastrophic insurance policies. No longer would patients be faced with a choice of having their employer pay for small medical bills with before-tax dollars or paying out-of-pocket costs with aftertax dollars. Thus, there would be no reason for them to prefer to have insurance pay for most medical bills, and insurance policies would no longer carry small copayments. As we have seen, that is the most crucial element in stopping the soaring increase in health care costs without clumsy, government-imposed price controls.

The Patient Power plan would allow patients to selfinsure (meaning that patients themselves pay for the treatment of their illnesses) for many potential medical bills through medical savings accounts that would go hand in hand with the tax changes. It typically costs an employer more than $4,800 to provide health insurance for a worker, her spouse, and two children. Under the Patient Power plan, employers would purchase only catastrophic policies for workers, and workers would deposit the savings in premiums in medical savings accounts. The medical savings accounts could be used to pay for small, routine medical bills not covered by catastrophic health insurance. If the account was not used to pay for medical bills, the owner could roll it over into an IRA to be used for other purposes after retirement. Patients would have an incentive not to use their medical savings accounts except for medical care that they deemed worth the money, since they would benefit directly from economizing on medical care. In addition, selfinsurance eliminates the paperwork involved with having third parties pay medical bills. It also eliminates the costs of having third parties monitor the transactions between patient and doctor, thereby greatly reducing administration costs generated by PSROs, PROs, and URs.

Second, Patient Power would reduce state regulations that currently mandate many benefits that must be provided by each health insurance policy sold in a state irrespective of patients' wants or needs. Such regulations drive up the price of health insurance and make the purchase of a policy less attractive for persons who are not interested in the extra benefits mandated by the state. If consumers are allowed to purchase insurance that is tailored to their specific needs without having to comply with state mandates, they will be happier and will save money.

Finally, Patient Power would reform tort law to allow patients and doctors to contract in advance to rationally insure against accidents or errors.

Conclusion

The moral of this story is crystal clear: third-party payment mechanisms have raised the total consumption of medical resources to unprecedented levels. The excessive use of medical resources due to third-party payments was estimated to be over $300 billion and the excessive administrative costs to be in the vicinity of $33 billion.

To lower the currently very large medical expenditures in the United States, the third-party payment system must be reined in. Putting the patient back in control of the medical purchasing decision is the most effective way to control third-party mechanisms, while still providing a safety net for Americans.

The worst policy that we could follow would be to increase third-party payments and reduce copayments. Yet that is exactly what is proposed by the Clinton administra tion. The evidence makes it abundantly clear that the current increase in medical bills will only be exacerbated by the Clinton plan and that rising costs will quickly run into the spending caps contained in the Clinton plan. That plan would be greatly improved if it were to impose high copayments on patients instead of low copayments, and if it were to keep predictable and relatively inexpensive medical costs, such as dentistry and eye care, out of the thirdparty payment system. But even if those changes were made, the Clinton plan would still create a large government bureaucracy controlling and limiting consumer choices, and it still would contain the dreadful idea of spending caps as a means of reducing medical costs.

The Patient Power plan is much more likely to reduce health care costs.

Notes

(1) Recreational expenditures, relative to disposable in come, increased from 5.0 percent in 1958 to 7.1 percent in 1988, according to statistics reported in Harold Vogel, Entertainment Industry Economics (Cambridge: Cambridge University Press, 1990), p. 348.

(2) John C. Goodman and Gerald L. Musgrave, Patient Power: Solving America's Health Care Crisis (Washington: Cato Institute, 1992).

(3) Ibid. That appears to be an outgrowth of two factors. First, during World War II price controls were in place at a time when employers were looking to increase the pay of workers. Providing additional fringe benefits allowed employers to circumvent price controls, and fringe benefits thus became a common part of an employee's compensation. Second, tax laws allow employers to deduct medical insurance premiums, whereas individuals have no such right (unless their medical bills are large enough for them to declare them as itemized deductions, which is certainly not the usual case). Obviously, those factors provide a strong incentive for most employees to purchase their medical insurance through their employers.

(4) Note that the inconvenience of the medical procedure, lost wages, pain, and so on, are taken into account in the patient's valuation of the medical procedure. Thus, a patient will request any procedure for which all "psychic" costs are less than the benefits, ignoring the monetary costs of the procedure itself.

(5) Although the leadership of the American Medical Associa tion originally opposed Medicare in 1965, they were against it for philosophical reasons and actually predicted that it would increase revenues going to doctors. Their opposition ended when most doctors realized the bonanza that it provid ed. See Edward Annis, Code Blue (Washington: Regnery Gate way, 1993).

(6) Health Insurance Association of America, 1991 Source Book of Health Insurance Data (Washington: HIAA, 1992), Table 2.2.

(7) Ibid., Table 4.4.

(8) Joseph Newhouse et al., "Some Interim Results from a Controlled Trial of Cost Sharing in Health Insurance," New England Journal of Medicine, December 17, 1981.

(9) Using the GDP deflator found in Robert J. Gordon, Macro economics (New York: Harper-Collins, 1993), appendix A.

(10) Robert Brook et al., "Does Free Care Improve Adults' Health? Results from a Randomized Clinical Trial," New England Journal of Medicine, December 8, 1983.

(11) Alan Sorkin reports on 20 estimates of price elasticity for various medical services. Price elasticity measures the responsiveness of consumption to changes in price. The majority are between 0.2 and 1, which is consistent with the RAND experiment. Sorkin, Health Economics (New York: Lex ington Books, 1992), p. 31.

(12) Health Insurance Association of America, Table 4.1.

(13) A large portion of expenditures on drugs is for overthe-counter products, which most medical plans do not cover.

(14) Ibid., Tables 2.2, 4.1, and 4.4; and Gordon, appendix A.

(15) There were three other categories of expenditure that were not included: nursing home expenditures, other health services, and other professional services. Since the con tents of categories with the term "other" is unclear, and might change dramatically over time, I followed the common practice of removing them. Nursing home expenditures in creased dramatically over the period, but that is probably more attributable to the decline in the extended family and the increase in life span than it is to any increase in medical use. Nursing homes, after all, generally do not respond to specific health problems so much as to old age and general inability to look after oneself. (Younger family members used to look after elderly relatives.) In addition, regulation of nursing homes during the period raised their costs significantly, according to Goodman and Musgrave, p. 107. Had those categories been included, the statistical confidence in the relationship between out-ofpocket costs and growth in expenditure would have weakened considerably, although the direction of the relationship would have been the same.

(16) Obviously, the change in expenditures on those catego ries of medical care are likely to depend on many variables other than just the change in copayments. Some of those factors are changes in age cohorts, changes in medical technology (which itself is likely to be affected by the copayment rate), and changes in diet and exercise. Never theless, these results are consistent with those of prior studies, are statistically significant, and are not related in a clear way to potential left-out factors.

(17) Regressing on the share of out-of-pocket expenses (OOPE) gave the equation: Growth in Expenditures (GIE) = -255 _ OOPE + 478. The t-statistic on OOPE is 12.34 and the rsquared (adjusted) is .97. Those coefficients imply that if out-of-pocket expenses had been 100 percent, medical expen diture would have grown in 1990 to only 223 percent (478 255) of its 1965 value. The GNP in 1990 was 194 percent of its 1965 value, so medical care would have remained almost constant as a percentage of GNP.

(18) This estimate implies a higher growth in medical expen diture than some others suggest. For example, Sorkin re ports that most estimates of income elasticity of medical care are in the range of .5 to .7, meaning that medical expenses would be expected to grow only 60 percent as fast as income, holding everything else constant.

(19) In 1992 spending on medical care was approximately $830 billion. Assuming that our best alternative is to have third-party payments in the vicinity of 25 percent, we would expect current spending to be approximately 60 percent too high. Thus, for 1992, under a system with smaller thirdparty payments, medical spending would have been only $520 billion, and, therefore, slightly over $300 billion in health care spending was excessive.

(20) Measuring the actual deadweight losses associated with excessive expenditures is an imprecise task. Martin Feld stein calculated the possible deadweight losses from the overuse of hospitalization. On the basis of the 1969 outof-pocket expense of 33 percent, he estimated that the dead weight loss ran from a low of 23 percent of total hospital revenues to a high of 67 percent of total revenues. Felds tein, Hospital Costs and Health Insurance (Cambridge, Mass.: Harvard University Press, 1981), chap. 6. With current outof-pocket expenses for hospitalization running at only 5 percent, we would expect even larger deadweight losses than he found. Feldstein reports (p. 99) that Mark Pauly mea sured welfare loss at $450 million for 1963, which is 15 percent of 1963 total health expenditures, as reported in Health Insurance Association of America, Table 4.4.

(21) Feldstein, p. 66. He reports estimates that the elas ticity of hospital days with respect to price is between .5 and .7. Assuming that .6 is the appropriate number, and assuming that it is an arc elasticity, decreasing payment from the market price to zero would increase usage by 250 percent.

(22) Health Insurance Association of America, Table 4.1.

(23) It should be noted that using data on personal consump tion expenditure, which exclude Medicaid and some other government spending (public health, research, construction) and which go back to 1950, there is no evidence that admin istrative costs grew more rapidly than other medical expens es from 1950 to 1965.

(24) Some of the higher estimates come from Steffie Wool handler and David Himmelstein, "Administrative Costs of U.S. Health Care," New England Journal of Medicine, May 2, 1991; they claim that about 20 percent of medical spending is administrative in nature. They also claim that health care administration costs in Canada were only about 10 percent of health care spending, and conclude, therefore, that about half of the U.S. administrative expense was wasteful. They also note that there was a very significant increase in wasteful administrative costs between 1983 and 1987, for which they blame the increased use of cost-containment mechanisms. There are good reasons to be suspicious of those results, as reported in a critique of the study by the Health Insurance Association of America in the May 30, 1991, issue of Medical Benefits.

(25) Although approximately 37 million Americans may not have health insurance on any given day, only about 7 million fail to have health insurance for an entire year. Most uninsured individuals are only temporarily uninsured.

Published by the Cato Institute, Policy Analysis is a regular series evaluating government policies and offering proposals for reform. Nothing in Policy Analysis should be construed as necessarily reflecting the views of the Cato Institute or as an attempt to aid or hinder the passage of any bill before Congress. Contact the Cato Institute for reprint permission. Printed copies of Policy Analysis are $6.00 each ($3.00 in bulk). To order, or for a complete listing of available studies, write to: Policy Analysis, Cato Institute, 1000 Massachusetts Avenue NW, Washington, D.C. 20001. (202)842-0200 FAX (202)842-3490 E-mail subscriptions@cato.org

1994 The Cato Institute
Please send comments to webmaster