The Demand for Medical Care. Factors that Determine the Demand for Health Care This topic explores the demand side of the medical care market. The topic highlights: • the theoretical derivation of the demand curve for medical services • economic and noneconomic variables that influence the demand for medical services • the impact of health insurance on the demand for medical services • the concept of elasticity of demand • a review of the empirical literature concerning the factors that determine the demand for medical care • an examination of health spending in the United States • a review of the sources and uses of health care funds in the United States. Many people have the misconception that economic theory has little relevance to the demand for medical care because economic factors are not important when an individual needs urgent medical attention. Recall Joe in Topic 1, who awoke one night with a pain in his chest and realized he was having a heart attack. It is highly unlikely that he and his wife considered the price of medical care as Joe was rushed to the hospital. However, most visits to a physician's office and the majority of visits to a hospital emergency room are not of a life-threatening nature. Thus, for many medical care transactions, there is sufficient time to make conscious choices, and price often plays an important role in the determination of choices. Results of a survey of various types of health care providers and insurers substantiate the critical role price plays in determining the demand for medical care (Winslow, 1994). According to the survey, price was ranked as more important than patient satisfaction or access to doctors, among other factors, in determining the economic success of health care providers.
The Demand for Medical Care and the Law of Demand To derive the demand curve for medical care, we must first establish the relation between the quantity of medical services and utility. Recall from Topic 2 that the stock of health can be treated as a durable good that generates utility and is subject to the law of diminishing marginal utility. As a reminder, note that we continue to ignore the intermediate step between the stock of health, the services it provides, and utility. This means that each incremental improvement in health generates successively smaller additions to total utility. We also know that medical services are an input in the production of health because a person consumes medical care services for the express purpose of maintaining, restoring, or improving health. However, the law of diminishing marginal productivity causes the marginal improvement to health brought about by each additional unit of medical care consumed to decrease. From this discussion, it follows that medical care indirectly provides utility. Specifically, medical care helps to produce health, which in turn generates utility. Consequently, utility can be specified as a function of the quantity of medical care. Figure 5-1 depicts the relation between the level of medical care consumed and utility. Utility is specified on the vertical axis, and the quantity of medical care (q) is measured on the horizontal axis. The shape of the total utility curve indicates that utility increases at a decreasing rate with respect to medical care, or that medical care services are subject to diminishing marginal utility. Marginal utility decreases because (1) each successive unit of medical care generates a smaller improvement in health than the previous unit (due to the law of diminishing marginal productivity) and (2) each increase in health, in turn, generates a smaller increase in utility (due to the law of diminishing marginal utility).
The Utility-Maximizing Rule Given market prices at a point in time, consumers must decide which combination of goods and services, including medical care, to purchase with their fixed incomes. According to microeconomic theory, each consumer chooses the bundle of goods and services that maximizes utility. Without working through the mathematics underlying the process, logic dictates that consumer utility is maximized when the marginal utility gained from the last dollar spent on each product is equal across all goods and services purchased. That is, assuming all prices are known, income is spent over the period in question, and all products are subject to the law of diminishing marginal utility. The shape of the utility curve illustrates that total utility increases at a decreasing rate with respect to the level of medical care consumed. The curve has a bow shape for two reasons. First, each additional unit of medical care consumed results in a smaller increase in health than the previous unit because of the law of diminishing marginal productivity. Second, each additional improvement in health generates a smaller increase in utility because of the law of diminishing marginal utility.
The shape of the utility curve illustrates that total utility increases at a decreasing rate with respect to the level of medical care consumed. The curve has a bow shape for two reasons. First, each additional unit of medical care consumed results in a smaller increase in health than the previous unit because of the law of diminishing marginal productivity. Second, each additional improvement in health generates a smaller increase in utility because of the law of diminishing marginal utility.
This condition is known as the utility-maximizing rule, and it basically states that total utility reaches its peak when the consumer receives the maximum “bang for the buck” in terms of marginal utility per dollar of income from each and every good. In mathematical terms, the rule states that utility is maximized when (5-1) MUq/Pq = MUZ/PZ, where MUq represents the marginal utility received from the last unit of medical care purchased, q, and MUz equals the marginal utility derived from the last unit of all other goods, z. The latter good is often referred to as a composite good in economics. To illustrate why the utility-maximizing rule must hold, suppose that (5-2) MUq/Pq > MUZ/PZ,
In this case, the last dollar spent on medical care generates more additional utility than the last dollar spent on all other goods. The consumer can increase total utility by reallocating expenditures and purchasing more units of medical care and fewer units of all other goods. As the consumer purchases more medical services at the expense of all other goods (remember that the consumer’s income and the composite good’s price are fixed), the marginal utility of medical care falls and the marginal utility of other goods increases. This, in turn, causes the value of MUq/Pq to fall and the value of MUz/Pz to increase. The consumer purchases additional medical services until the equality in Equation 5-1 again holds, or the last dollar spent on each product generates the same amount of additional satisfaction. At this point, total utility is maximized and any further changes in spending patterns will negatively affect total utility.
The Law of Demand The equilibrium condition specified in Equation 5-1 can be used to trace out the demand curve for a particular medical service, such as physician services. For simplicity, assume the prices of all other goods and income remain constant and initially the consumer is purchasing the optimal mix of physician services and all other goods. Now assume the price of physician services increases. In this case, MUq/Pq is less than MUz/Pz (where MUq and Pq represent the marginal utility and price of physician services, respectively). Consequently, the consumer receives more satisfaction per dollar from consuming all other goods. In reaction to the price increase, the consumer purchases fewer units of physician services and more units of all other goods. This reallocation continues until MUq/Pq increases and MUz/Pz decreases and the equilibrium condition of Equation 5-1 is again in force such that the last dollar spent on each good generates an equal amount of utility. Thus, an inverse relation exists between the price and the quantity demanded of physician services. If the price of physician services continually changes, we can determine a number of points representing the relation between the price and the quantity demanded of physician services. Using this information, we can map out a demand curve like the one depicted in Figure 5-2, where the horizontal axis indicates the amount of physician services consumed (as measured by the number of visits, for example) and the vertical axis equals the price of physician services. The curve is downward sloping and reflects the inverse relation between the price and the quantity demanded of physician services, ceteris paribus. For example, if the price of physician services equals P0, the consumer is willing and able to purchase q 0. Notice that if the price falls to P1, the consumer purchases q 1 amount of physician services.
The individual demand curve for physician services is downward sloping, illustrating that quantity demanded increases as the price of physician services drops. Utility analysis, or the income and substitution effects, can be used to derive this inverse relationship, which is called the law of demand.
In this case, price represents the per-unit out-of-pocket expense the consumer incurs when purchasing medical services from a physician. As such, it equals the amount the consumer must pay after the impact of third-party payments has been taken into account. Naturally, if the visit to the physician is not covered by a third party, the actual price of the visit equals the out-of-pocket expense. The substitution and income effects associated with a price change offer another theoretical justification of the inverse relationship between price and quantity demanded. Both of these effects predict that a higher price will lead to a smaller quantity demanded and, conversely, a lower price will result in a greater quantity demanded. According to the substitution effect, a decrease in the price of physician services causes the consumer to substitute away from the relatively higher-priced medical goods, such as hospital outpatient services, and purchase more physician services. That is, lower-priced services are substituted for higher-priced ones. As a result, the quantity demanded of physician services increases as price decreases. According to the income effect, a lower price also increases the real purchasing power of the consumer. Because medical care is assumed to be a normal good (that is, the quantity demanded of medical services increases with income), the quantity demanded of physician services increases with the rise in purchasing power. That also generates an inverse relation between price and quantity demanded because as price falls, real income increases and quantity demanded rises. Taken together, the substitution and income effects indicate that the quantity demanded of physician services decreases as price increases. In summary, Figure 5-2 captures the inverse relationship between the price the consumer pays for medical care (in this instance, physician services) and the quantity demanded. The curve represents the amount of medical care the consumer is willing and able to purchase at every price. Utility analysis, or the income and substitution effects, can be used to generate this relationship. This inverse relationship is sometimes referred to as the law of demand. It is important to note that the demand for medical care is a derived demand, because it depends on the demand for good health. A visit to a dentist illustrates this point. An individual receives no utility directly from having a cavity filled. Rather, utility is generated from an improvement in dental health. Of course, other economic and noneconomic variables also influence the demand for health care. Unlike price, which causes a movement along the demand curve, other factors influence the quantity demanded by altering the position of the demand curve. These other economic and noneconomic determinants of demand are the topic of the next section.
Other Economic Demand-Side Factors Income is another economic variable that affects the demand for medical services. Because medical care is generally assumed to be a normal good, any increase in income, which represents an increase in purchasing power, should cause the demand for medical services to rise. Figure 5-3 illustrates what happens to the demand for physician services when income increases. The increase in income causes the demand curve to shift to the right, from d 0 to d 1, because at each price the consumer is willing and able to purchase more physician services. Similarly, for each quantity of medical services, the consumer is willing to pay a higher price. This is attributable to the fact that at least some portion of the increase in income is spent on physician services. Conversely, a decrease in income causes the demand curve to shift to the left. Some goods are referred to as inferior goods. This is because the demand for these goods decreases as income increases. A classic nonmedical example is hamburger. As real income increases, the consumer may prefer to buy more expensive cuts of meat and purchase less hamburger. In the medical sector, hospital outpatient services may be an example of an inferior good. As income increases, the consumer may prefer to visit a private physician to receive individual care rather than outpatient services. As a result, the demand for outpatient services may decrease as income increases. Some researchers have found that tooth extractions represent an inferior dental service.
Medical care is assumed to be a normal good, which means that as income increases the consumer spends at least a portion of the increase in purchasing power on additional physician services. As a result, the individual demand curve for physician services shifts to the right, from d0 to dv when income increases. At each price, the consumer is now willing and able to purchase more physician services.
The demand for a specific type of medical service is also likely to depend on the prices of other goods, particularly other types of medical services. If two or more goods are jointly used for consumption purposes, economists say that they are complements in consumption: Because the goods are consumed together, an increase in the price of one good inversely influences the demand for the other. For example, the demand for eyewear (that is, glasses or contact lenses) and the services of an optometrist are likely to be highly complementary. Normally, an individual has an eye examination before purchasing eyewear. If these two goods are complements in consumption, the demand for optometric services should increase in response to a drop in the price of eyewear. As a result, the demand curve for optometric services shifts to the right. Another example of a complementary relation exists between obstetric and pediatric services. An increase in the price of pediatric services should inversely influence the demand for obstetric services. If, for example, a woman postpones pregnancy because of the high cost of pediatric services, her demand for obstetric services also falls. The demand curve for obstetric services shifts to the left. It is also possible for two or more goods to satisfy the same wants or provide the same characteristics. If that is the case, economists say that these goods are substitutes in consumption: The demand for one good is directly related to a change in the price of a substitute good. For example, suppose physician services and hospital outpatient services are substitutes in consumption. As the price of outpatient services increases, the consumer is likely to alter consumption patterns and purchase more physician services because the price of a visit to the doctor is cheaper in relative terms. That causes the demand curve for physician services to shift to the right. Generic and brand-name drugs provide another example of two substitute goods. The demand for brand-name drugs should decrease with a decline in the price of generic drugs. If so, the demand curve for brand-name drugs shifts to the left. Finally, eyeglasses and contact lenses are likely to be substitutes in consumption. Time costs also influence the quantity demanded of medical services. Time costs include the monetary cost of travel, such as bus fare or gasoline, plus the opportunity cost of time. The opportunity cost of an individual’s time represents the dollar value of the activities the person forgoes when acquiring medical services. For example, if a plumber who earns $50 an hour takes two hours off from work to visit a dentist, the opportunity cost of the time equals $100. The implication is that the opportunity cost of time is directly related to a person’s wage rate. Given time costs, it is not surprising that children and elderly people often fill doctors’ waiting rooms. Time costs can accrue while traveling to and from a medical provider, waiting to see the provider, and experiencing delays in securing an appointment. In other words, travel costs increase the farther an individual has to travel to see a physician, the longer the wait at the doctor’s office, and the longer the delay in getting an appointment. It stands to reason that the demand for medical care falls as time costs increase (that is, as the demand curve shifts to the left).
The Relationship between Health Insurance and the Demand for Medical Care The growth of health insurance coverage is one of the most significant developments in the health care field over the past several decades. It has had a profound influence on the allocation of resources within the medical care market, primarily through its impact on the out-of-pocket prices of health care services. Out-of-pocket payments for health care dropped from almost half of total expenditures in 1960 to approximately one-seventh in 2003. Even more striking, out-of-pocket payments for hospital care fell from 20.7 percent in 1960 to a mere 3.2 percent in 2003. Given that various features are associated with health insurance policies, it is impossible to discuss the economic implications of each one. Here we will focus on three of the more common features of health insurance policies: coinsurance, copayments, and deductibles. Coinsurance and Copayments. Many health insurance plans, particularly private plans, have a coinsurance component. Under a coinsurance plan, the consumer pays some fixed percentage of the cost of health care and the insurance carrier picks up the other portion. For example, under a plan with a coinsurance rate of 20 percent (a common arrangement), the consumer pays 20 cents out of every dollar spent on health care and the carrier picks up the remaining 80 cents. As you can imagine, an insurance plan like this one has a significant impact on the demand for health care because it effectively lowers the out-ofpocket price of health care by 80 percent. Let’s begin our discussion of coinsurance coverage by looking at the demand curve for medical care from an alternative perspective. We normally think of the demand curve as revealing the amount of a good that a consumer is willing and able to buy at various prices. However, a demand curve also shows the consumer’s willingness to pay (or marginal benefit) for each unit of a good. The negative slope of the curve indicates that the willingness to pay falls as more of the good is consumed due to the law of diminishing marginal utility. For example, the demand curve d WO (WO = without insurance) in Figure 5-4 represents the consumer’s demand or willingness to pay for office visits in the absence of health insurance coverage. This “effective” demand curve reveals that the consumer is willing to pay $50 for the fifth office visit. If $50 is the market price paid by the consumer, she visits the physician five times during the year in the process of maximizing utility because any additional office visits do not yield benefits that compensate for their higher out-of-pocket costs. Notice that the consumer’s willingness to pay for the first four visits, as revealed by the effective demand curve, exceeds the market price of $50. The difference between the willingness to pay and the market price paid is referred to as a customer surplus and, in this example, reflects the net benefits received from visiting the doctor the first four times. As discussed in Topic 8, market price considers both supply and demand conditions. The demand curves in Figure 5-4 represent the effective and nominal demands of an individual. Individual demands must be horizontally summed to arrive at a market demand and then interacted with supply to determine the market price. Now suppose the consumer acquires a health insurance plan that requires her to pay a certain fraction, C0, of the actual price, P. In this case, the insurance coverage drives a wedge between the willingness to pay, or effective demand, and the actual price, or “nominal” demand, for the office visits. Because the utility-maximizing consumer determines the optimal number of times to visit the physician by equating her willingness to pay (or marginal benefit) to the out-of-pocket price (marginal cost), the relationship between the actual and out-of-pocket price can be specified by the following equation: (5-3) Pw = CoP. Here Pw stands for the consumer’s willingness to pay for the last visit, and C0 represents the coinsurance amount. If we solve Equation 5-3 for the actual price, we get (5-4) P = Pw/Co. Because the coinsurance, C0, is less than 1, it follows that the actual price paid, or nominal demand, for office visits is greater than the out-of-pocket price the consumer pays. For example, if she is willing to pay $50 for five visits to a doctor and the coinsurance is 20 percent of the full price, the actual price equals $250 per visit, or $50/0.2.
The graph illustrates how a coinsurance health plan impacts the individual demand curve for physician visits. The demand curve labeled dwo is the individual's effective demand without coinsurance while the demand curve labeled C/WI is with coinsurance. The nominal demand curve dwi traces out the total price for various physician visits and captures that portion paid by consumers as out-of-pocket payments as well as that portion paid by the insurance carrier. If you draw a vertical line from any point on the nominal demand curve to the horizontal axis, you can break down the amount paid by consumers (from the horizontal axis to the dwo curve) and the amount paid by the insurance carrier (the wedge between the dwi and dwo curves). As the coinsurance rate falls, dwi rotates upward and pivots off the point where the two curves cross the horizontal axis.
The nominal demand curve labeled d WI (WI = with insurance) in Figure 5-4 reflects the total price paid for medical services that takes into account the coinsurance paid by the insured. The vertical distance between d WI and the horizontal axis represents the total price for office visits, which can be broken down into the amount the consumer pays and the amount the insurance carrier pays. The portion of the total price the consumer pays as an out-ofpocket payment equals the distance between the horizontal axis and the d WO demand curve. The remaining distance between the two curves represents the amount the insurance carrier pays. It represents the wedge that coinsurance drives between the consumer’s willingness to pay, or effective demand, and the total price paid, or nominal demand. It is easy to see from this analysis that a reduction in the coinsurance rate causes the nominal demand curve d WI to rotate clockwise and pivot off the point where d WO crosses the horizontal axis. At a zero willingness-to-pay price, insurance has no bearing on quantity demanded because medical care is a free good to the individual. In addition, the nominal demand curve d WI becomes steeper as the coinsurance, C0, decreases in value as indicated by Equation 5-4. That makes intuitive sense, because we expect the consumer to become less sensitive to changes in the total price as the coinsurance declines. In the case where the consumer has full coverage (C0 = 0), the nominal demand curve d WI rotates out to its fullest extent and becomes completely vertical. This is shown in Figure 5-5. Because the consumer faces a zero price, she consumes medical care as though it were a free good, when in reality it has a nonzero price. Equation 5-4 can be used to illustrate that point. As C0 approaches zero, the total price is potentially infinity even when Pw equals zero. Coinsurance should not be confused with a copayment. A copayment represents a fixed amount paid by the consumer that is independent of the market price or actual costs of medical care. For example, a person may be required to pay $10 for each office visit regardless of the actual fee negotiated by the health insurer with the physician. Like a lower coinsurance rate, a reduced copayment results in a movement down the effective demand curve and typically leads to greater quantity of care demanded. But unlike a change in the coinsurance rate, a change in the copayment does not cause a rotation of the nominal demand because the consumer’s portion of the bill is independent rather than proportional to nominal demand (that is, the actual price paid).
The graph illustrates the situation in which the individual has complete medical coverage and the coinsurance rate is zero. Notice that the nominal demand curve is vertical because the individual faces a zero out-of-pocket price and visits the physician without regard to the actual price.
Also unlike coinsurance, a copayment does not automatically change with an adjustment in the costs of providing medical care. For example, suppose, in response to higher production costs, a physician negotiates a higher price with the insurer for each office visit so that the market price increases from $100 to $150. An insured individual who is responsible for paying 20 percent of the cost now faces a $10 increase in his coinsurance from $2 0 to $30 per office visit. However, an insured individual who is required to pay a copayment of $10 per office visit is unaffected by the higher negotiated price for an office visit (at least until the insurance policy is renegotiated). Thus, compared to a copayment, coinsurance makes consumers more sensitive to the actual market price of medical care. Deductibles. Many insurance policies have a deductible whereby the consumer must pay out of pocket a fixed amount of health care costs per calendar year before coverage begins. For example, the plan may call for the individual to pay the initial $200 of health care expenses with a limit of $500 per family per year. Once the deductible is met, the insurance carrier pays all or some portion of the remaining medical bills, depending on how the plan is specified. From the insurance carrier’s perspective, the purpose of a deductible is to lower costs. This is accomplished in two ways. First, the deductible is likely to lower administrative costs because fewer small claims will be filed over the course of a year. Second, the deductible is likely to have a negative impact on the demand for health care. The extent to which this is true, however, is difficult to determine and depends on such factors as the cost of the medical episode, the point in time when the medical care is demanded, and the probability of needing additional medical care for the remainder of the period. To illustrate, assume a new deductible is put in place at the beginning of each calendar year. Once the deductible is met, the consumer has full medical coverage. It is easy to see that the extent to which a deductible influences the demand for medical services for any one medical episode is likely to be inversely related to the cost of the medical services involved. For example, if the consumer faces a potentially large medical bill for an operation, the existence of a deductible is likely to have little impact on demand. This is because, in relative terms, the deductible represents very little money. On the other hand, a deductible may play a crucial role in the decision to purchase medical care if the cost of such care is relatively inexpensive. In this case, the out-of-pocket cost is substantial relative to the total cost, and the consumer may elect not to purchase the medical care or postpone the purchase to a later date. It is slightly more difficult to understand how the health of the individual, along with the time of the year, influences the impact of a deductible on demand. The best way to explain this is with an example. Consider a normally healthy individual who contracts the flu late in November and has incurred no medical expenses up to this point. Under these circumstances, he may be less inclined to visit the doctor. This is because he will have little opportunity to take advantage of the fact that health care is a free good after he makes his initial visit to the physician and fulfills the deductible. On the other hand, this same individual is much more likely to visit the physician if he catches the flu early in February and his overall health is such that he can expect to visit the physician three or four more times over the remainder of the year. By visiting the doctor and meeting the deductible, he lowers the cost of any future visits to zero for the rest of the year. Therefore, a deductible is likely to have the greatest negative impact on the demand for medical care when the cost of the medical episode is low, the need for care is late in the calendar year, and the probability of needing future care is slight because the person is in good health.
Moral Hazard Before we leave the subject of the impact of insurance on the demand for medical care, we need to introduce the concept of moral hazard. Moral hazard refers to the situation in which consumers alter their behavior when provided with health insurance. For example, health insurance may induce consumers to take fewer precautions to prevent illnesses or to shop very little for the best medical prices. In addition, insured consumers may purchase more medical care than they otherwise would have without insurance coverage. Let’s illustrate this point by referring to Figure 5-4. According to the graph, a consumer without insurance purchases five units of medical services at a price of $50 per unit. If that consumer acquires full medical coverage such that the insurer’s coinsurance rate, C0, equals zero, the quantity demanded of medical care increases to the point where the demand curve crosses the horizontal axis. At this point, the consumer consumes medical care as though it were a free good because she faces a zero price. Thus, any extension of medical insurance coverage has the potential to increase the consumption of medical care because consumers no longer pay the full price. The availability and extensiveness of health insurance may have a profound effect on medical care expenditures. Topic 6 examines the implications of moral hazard in more detail.
Noneconomic Determinants of the Demand for Medical Care Four general noneconomic factors influence the demand for medical services: tastes and preferences, physical and mental profile, state of health, and quality of care. Taste and preference factors include personal characteristics such as marital status, education, and lifestyle, which might affect how people value their healthy time (that is, their marginal utility of health), or might lead to a greater preference for certain types of medical services. Marital status is likely to impact the demand for health care in the marketplace primarily through its effect on the production of health care in the home. A married individual is likely to demand less medical care, particularly hospital care, because of the availability of a spouse to care for him at home, such as when recuperating from an illness. The impact of education on the demand for medical care is difficult to predict. On the one hand, a consumer with additional education may be more willing to seek medical care to slow down the rate of health depreciation because that consumer may have a better understanding of the potential impact of medical care on health. As an example, an individual with a high level of education may be more inclined to visit a dentist for periodic examinations. Thus, we should observe a direct relation between educational attainment and demand. On the other hand, an individual with a high level of education may make more efficient use of home-produced health care services to slow down the rate of health depreciation and, as a result, demand fewer medical care services. For example, such an individual may be more likely to understand the value of preventive medicine (such as proper diet and exercise). In addition, the individual may be more likely to recognize the early warning signs of illness and be more apt to visit a health care provider when symptoms first occur. As a result, health care problems are addressed early when treatment has a greater probability of success and is less costly. That means that we should observe an inverse relation between the level of education and the demand for medical care, particularly acute care. Finally, lifestyle variables, such as whether the individual smokes cigarettes or drinks alcohol in excessive amounts, affect health status and consequently the amount of health care demanded. For example, a person may try to compensate for the detrimental health impact of smoking by consuming more health care services. That translates into an increased demand for medical care. The profile variable considers the impact of such factors as gender, race/ethnicity, and age on the demand for medical services. For example, females generally demand more health care services than males primarily because of childbearing. In addition, certain diseases, such as cardiovascular disease, osteoporosis, immunologic diseases (such as thyroid disease and rheumatoid arthritis), mental disorders, and Alzheimer’s disease, are more prevalent in women than men (Miller, 1994). Age also plays a vital role in determining the demand for medical care. As we stated in Topic 2, as an individual ages, the overall stock of health depreciates more rapidly. To compensate for this loss in health, the demand for medical care is likely to increase with age, at least beyond the middle years (the demand curve shifts to the right). Thus, we should observe a direct relation between age and the demand for medical care. State of health controls for the fact that sicker people demand more medical services, everything else held constant. As you might expect, health status and the demand for health care are also likely to be directly related to the severity of the illness. For example, a person who is born with a medical problem, such as hemophilia, is likely to have a much higher than average demand for medical care. In economics jargon, an individual who is endowed with less health is likely to demand more medical care in an attempt to augment the overall stock of health. As another example, Fuchs and Frank (2002) find an increased use of medical care, both inpatient and outpatient care, among Medicare recipients living in highly polluted Metropolitan areas of the United States. The relationship holds even after controlling for population, education, income, racial composition, and cigarette use. Finally, although nebulous and impossible to quantify, the quality of care is also likely to impact the demand for medical care. Because quality cannot be measured directly, it is usually assumed to be positively related to the amount and types of inputs used to produce medical care. Feldstein (1967, pp. 158-62) defines the quality of care as “a catch-all term to denote the general level of amenities to patients as well as additional expenditures on professional staff and equipment.” For example, a consumer may feel that larger hospitals provide better-quality care than smaller ones because they have more specialists on staff along with more sophisticated equipment. Or, that same individual may think that physicians who have graduated from prestigious medical schools provide a higher quality of care than those who have not. It matters little whether the difference in the quality of medical care provided is real or illusory. What matters is that the consumer perceives that differences in quality actually exist. With regard to the previous example, it is certainly not the case that larger hospitals provide better care for all types of hospital services. However, if the consumer generally feels that larger hospitals provide better services, the demand for medical services at larger hospitals will be higher than at smaller ones. As Feldstein’s definition indicates, quality can also depend on things that have little to do with the actual production of effective medical care. For example, the consumer may prefer a physician who has a pleasant office with a comfortable waiting room along with courteous nurses. Thus, any increase in the quality of care provided is likely to increase that consumer’s demand for medical care regardless of whether it affects the actual production of health care. Before we move on, we must distinguish between a movement along the demand curve and a shift of the curve. A change in the price of medical services generates a change in the quantity demanded, and this is represented by a movement along the demand curve. If any of the other factors change, such as income or time costs, the demand curve for medical services shifts. This shift is referred to as a change in demand. Thus, a change in the quantity demanded is illustrated by a movement along the demand curve, while a change in demand is illustrated by a shift of the curve. In summary, let’s review the variables we expect to influence an individual’s demand for medical care. Economic theory indicates that the demand equation should look something like the following: (5-5) Quantity demanded = f (out-of-pocket price, income, time costs, prices of substitutes and complements, tastes and preferences, profile, state of health, and quality of care)
Equation 5-5 states that the quantity demanded of medical services is a function of, or depends on, the general factors listed. Note that a change in the first factor results in a movement along a given demand curve, whereas an adjustment in the other factors produce a shift of the demand curve. A rightward shift indicates a greater demand and a leftward shift reveals a lower demand.
The Market Demand for Medical Care Up to now, we have been discussing the individual’s demand for medical care services. The market demand for medical care, such as physician services, equals the total demand by all consumers in a given market. In graphical terms, we can construct the market demand curve for medical care services by horizontally summing the individual demand curves. This curve represents the amount of medical services that the entire market is willing and able to purchase at every given price. For example, if the average price of a visit to a doctor is $50 and at this price consumer A is willing to see a physician three times over the course of a year while consumer B is willing to make four visits, the total, or market, demand for physician services is seven visits per year at $50 per visit. The market demand curve is downward sloping for the same reasons the individual demand curves are downward sloping. In addition, the factors that shift the individual demand curves also shift the overall market demand curve, providing the changes take place on a marketwide basis. The market demand curve also shifts if the overall number of consumers in the market increases or decreases. For example, the demand for medical care in a particular community may increase if it experiences an influx of new residents. This causes the market demand curve to shift to the right. The development of a market demand curve allows us to distinguish between the intensive and extensive margins. The intensive margin refers to how much more or less of a product consumers buy when its price changes. The extensive margin captures how many more or fewer people buy a product when its price changes. Obviously, this is an important distinction to make for a product like medical care. Many medical purchases such as surgeries happen only once for a particular individual. As another example, an individual can have a particular tooth pulled only once. This is also a one-shot purchase that either happens or does not happen. If the price of tooth extraction falls, however, we may still observe a inverse relationship between the price and number of teeth extracted. That is because at the extensive margin, more consumers elect to purchase this onetime form of dental services as price falls. Consequently, quantity demanded may increase with a reduction in price because of changes that occur at the intensive and extensive margins.
The Fuzzy Demand Curve Up to this point, we have assumed the market demand curve for medical care is a well-defined line, implying a precise relation between price and quantity demanded. In reality this is usually not the case, and we need to refer to the derivation of the demand curve for medical care to see why. Recall that the demand for medical care is a derived demand and depends on the demand for health and the extent to which medical care influences the production of health. The relation between medical care and health, however, is far from exact. That is because there is a considerable lack of medical knowledge concerning the efficacy of certain types of medical interventions. As a result, health care providers disagree about the treatment of some types of medical problems, and the demand for medical services becomes fuzzy. For example, there is debate among physicians concerning when surgery is necessary for elderly males with prostate cancer. In addition, in some instances consumers may lack the information or medical knowledge they need to make informed choices. Consequently, consumers tend to rely heavily on the advice of their physicians when making such decisions – as when a particular medical test or surgery is necessary. The implication is that physicians, rather than consumers, choose medical services, which makes the demand curve fuzzier. Further complicating matters is the inability to accurately measure medical care, an issue we touched on earlier. For example, how do we measure the quantity of medical care produced during a one-hour therapy session with a psychiatrist? All these factors combined make it extremely difficult to accurately delineate the relation between the price and the quantity demanded of medical care. In other words, the relation between price and quantity demanded is rather fuzzy (Aaron, 1991). A more accurate depiction of the relation between price and quantity may not be a well-defined line but a gray band similar to the one depicted in Figure 5-6. Two implications are associated with the fuzzy demand curve. First, for a given price, we may observe some variation in the quantity or types of medical services rendered. Indeed, researchers have documented variations in physician practice styles across geographical areas (see, for example, Phelps, 1992); we take up that discussion in Topic 12. Second, for a given quantity or type of medical service, we are likely to witness price differences. For example, Feldstein (1988) reported a substantial variation in physician fees for similar procedures in the same geographical area. We must stress, however, that the existence of the band is unlikely to detract from the inverse relation between the price and the quantity demanded of medical care as suggested by the empirical evidence that follows.
The gray band represents the possible fuzziness of the demand for medical care given uncertainty and the role of the physician.
Elasticities Economic theory gives us insights into the factors that influence the demand for medical care along with the direction of their influence. For example, we know that if the price of physician services increases by 15 percent, the quantity demanded falls. But by how much does it fall? Is there any way to determine whether the decrease is substantial or negligible? The answer is yes, with the help of a measure economists call an elasticity. Elasticity measures that responsiveness of quantity demanded to a change in an independent factor.
Own-Price Elasticity of Demand The most common elasticity is the own-price elasticity of demand. This measure gauges the extent to which consumers alter their consumption of a good or service when its own price changes. The formula looks like this: (5-6) E D = %ΔQd/%ΔP where ED denotes the price elasticity of demand, %ΔQD represents the percentage change in quantity demanded, and %ΔP is the percentage change in price. As you can see from the formula, ED is a simple ratio that equals the percentage change in quantity demanded divided by the percentage change in price. Because elasticity is specified as a ratio of two percentage changes, it is scale free. This makes it much easier to compare elasticities across different goods. For example, we can compare the price elasticity of demand for physician services with that for nursing home care and not concern ourselves with the fact that the demand for physician services is usually measured in terms of the number of visits while the demand for nursing home care is measured in terms of the number of inpatient days. The point elasticity formula can be used to calculate the elasticity of demand if the changes in the variables are small. The formula equals (ΔQd/Qd)/(ΔP/P). For readers with a background in calculus, it equals (dQD/QD)/(dP/P) if the changes are infinitesimally small. The value of ED is negative and reflects the inverse relationship between price and quantity demanded. In economics, the normal practice is to take the absolute value of the price elasticity of demand measure, or | ED|, and eliminate the minus sign. If the price elasticity of demand is greater than 1 in absolute terms (| ED| > 1), the demand for the product is referred to as price elastic. In arithmetic terms, | ED| > 1 if the absolute value of the percentage change in price is smaller than the absolute value of the change in the quantity demanded, or |%ΔP| < |%ΔQd|. For example, if the price elasticity of demand for dental services equals 1.2, this means the quantity consumed falls by 12 percent if the price of dental care increases by 10 percent, ceteris paribus. The price elasticity of demand is referred to as inelastic if | ED| < 1 but greater than zero. In this case, |%ΔP| < |%ΔQd|, or the percentage change in price is greater than the percentage in quantity demanded in absolute value terms. For example, if the elasticity of demand for physician services equals 0.6, a 10 percent decrease in price leads to a 6 percent increase in quantity demanded. If | ED| happens to equal 1 because |%ΔP| equals |%ΔQd|, the price elasticity of demand is unit elastic. This implies that a 10 percent decrease in the price of the product leads to a 10 percent increase in the quantity demanded. A demand curve that is vertical is said to be perfectly inelastic because no change occurs in the quantity demanded when the price changes. In mathematical terms, ED equals zero because %ΔQd equals zero. At the other extreme, if the demand curve is horizontal, it is referred to as being perfectly elastic and | ED | equals infinity ('). Any change in price leads to an infinite change in the quantity demanded. It stands to reason that the more elastic the demand for the product, the greater the response of quantity to a given change in price. Compare the effects of a 10 percent decrease in price on two goods – one with a price elasticity of -0.1 and another with a price elasticity of -26. In the first case, the quantity demanded increases by only 1 percent, while in the second case, it increases by 26 percent. We can also use the elasticity of demand to make inferences regarding the slope of the demand curve. Generally, the more elastic the demand for the product, the flatter the demand curve at any given price. This also means the curve is relatively steep at any given point for an inelastic demand. Consider the two linear demand curves that intersect at point P0, Q0 in Figure 5-7. If the price of the product increases to P1, the quantity demanded decreases to Qa off the flat curve (Da) and to Qb off the steep curve (Db). Therefore, the same percentage increase in price generates a smaller percentage decrease in the quantity demanded for the steeper curve Db than for the flatter curve Da at a similar price of P0. This means demand must be more price elastic for curve Da than for curve Db over the range P0 to P1. Table 51 summarizes our discussion thus far on price elasticity of demand.
The steep demand curve, Db, is relatively inelastic and illustrates that an increase in price from P0 to P1 generates only a modest decrease in quantity demanded from Q 0 to QbThe flatter demand curve, D a , is relatively elastic and, in this case, the same increase in price for P0 to P1 generates a much larger decrease in quantity demanded from Q 0 to Q a .
The own-price elasticity of demand varies greatly across products, and economists point to several factors that determine its value. Among the factors most often mentioned are the portion of the consumer’s budget allocated to the good, the amount of time involved in the purchasing decision, the extent to which the good is a necessity, and the availability of substitutes. Briefly, as the portion of a consumer’s budget allocated to a good increases, the consumer is likely to become much more sensitive to price changes. Demand should therefore become more elastic. An increase in the decision-making time frame is also likely to make demand more elastic. If the consumer has more time to make informed choices, he or she is likely to react more strongly to price changes. Because the consumer typically pays a small portion of the cost of medical services because of insurance, and because medical services are sometimes of an urgent nature, these two considerations suggest that in many cases, the demand for medical services is inelastic with respect to price.
If a good is a necessity, such as a basic foodstuff, the own-price elasticity should be relatively inelastic. The product is purchased with little regard for price because it is needed. Basic phone service might be considered another example of a necessity. Because our society depends so heavily on the phone as a form of communication, it is difficult to imagine a household functioning effectively without one. Naturally, basic health care falls into the same category. If an individual needs a particular medical service, such as an operation or a drug, and if not having it greatly affects the quality of life, we can expect that person’s demand to be inelastic with respect to price. In addition, when a person needs a particular medical service in a life-or-death situation, demand is likely to be perfectly inelastic because the medical service must be purchased regardless of price if the person has sufficient income. Given that many medical services are necessities, we expect the overall demand for medical services to be somewhat inelastic. A word of caution, however: This does not mean the amount of health care demanded does not react to changes in price. Rather, it means a given percentage change in price generates a small percentage change in the quantity demanded of medical services. For some types of medical care, however, demand may be more elastic. Elective medical care, such as cosmetic surgery, may fall into this category, because in most instances it is considered a luxury rather than a necessity. As a result, price may play an important role in the decision to have the surgery. To a lesser degree, dentist services and eyewear might fall into this category. In fact, any medical service that can be postponed is likely to display some degree of price elasticity. The availability of substitutes is another determinant of price elasticity. As we saw earlier, various types of medical services may serve as substitutes for one another. The larger the number of substitutes, the greater the opportunity to do some comparison shopping. As a result, the quantity demanded of any medical service is likely to be more sensitive to price changes when alternative means of acquiring medical care are available. The own-price elasticity of demand for any given product should be directly related to the number of substitutes available. Stated another way, demand should become more price elastic as the number of substitutes expands. One implication is that the demand for an individual medical service or an individual medical care provider is likely to be more elastic than the market demand for medical care. One more point concerning the elasticity of demand needs to be discussed before we leave this subject. The own-price elasticity of demand can be used to predict what happens to total health expenditures if price increases or decreases. Total revenues (or total expenditures, from the consumer’s perspective) equal price times quantity. In mathematical notation, (5-7) TR = PQD, where TR represents total revenue. Demand theory tells us that as the price of a product increases, the quantity demanded decreases, or that P and Qd move in opposite directions. Whether total revenue increases or decreases when the price changes is dictated by the relative rates at which both variables change, or the elasticity of demand. Consider an increase in the price of physician services where demand is inelastic. This means that | %ΔQd| < | %ΔP|, or that the percentage increase in price is larger than the percentage decrease in quantity demanded in absolute value terms. In terms of Equation 5-7, P increases faster than Qd falls. This means total revenue must increase with a higher price. If demand happens to be elastic, the opposite occurs: Quantity demanded falls faster than price increases, and, as a result, total revenue decreases. No change occurs in total revenue when demand is unit elastic because the increase in price is matched by the same percentage decrease in quantity demanded. We leave it to you to work out the implications of a price decrease on total revenue when demand is elastic, inelastic, and unit elastic.
Other Types of Elasticity The concept of elasticity can be used to measure the sensitivity of quantity demanded to other demand-side factors as well. The income elasticity of demand represents the percentage change in quantity demanded divided by the percentage change in income, or EY = %ΔQd/%ΔY, where %ΔY equals the percentage change in income. It quantifies the extent to which the demand for a product changes when real income changes. If Ey is positive, the good is referred to as a normal good because any increase in income leads to an increase in quantity demanded. For example, if Ey equals 0.78, this means a 10 percent increase in income causes the quantity consumed to increase by 7.8 percent. An inferior good is one for which Ey is negative and an increase in income leads to a decrease in the amount consumed. For most types of medical care, the income elasticity of demand should be larger than zero. The cross-price elasticity (Ec) measures the extent to which the demand for a product changes when the price of another good is altered. In mathematical terms, E c = %ΔQx/%ΔPz, where the numerator represents the percentage change in the demand for good × and the denominator equals the percentage change in the price of good Z. If Ec is negative, we can infer that the two goods are complements in consumption. Returning to our earlier example, the cross-price elasticity between the demand for optometric services and the price of eyewear should be negative. If the price of eyewear increases, the demand for optometric services should drop. Two goods are substitutes in consumption when the cross-price elasticity is positive. For example, the cross-price elasticity of the demand for physician services with respect to the price of hospital outpatient services may turn out to be positive. Naturally, if Ec equals zero, the demand for the product is independent of the price of the other product.
Empirical Estimation Numerous studies have attempted to empirically quantify how various factors influence the demand for medical care. Although the studies varied widely in terms of methodology and scope of analysis, certain broad conclusions emerged. Generally, some form of Equation 5-5 is estimated with the use of regression analysis. Unfortunately, the dependent variable representing the amount of medical services consumed is very difficult to measure. Ideally, quantity demanded should capture both the utilization and the intensity of medical services. Data of these kinds are unavailable, so usually only some utilization measure, such as number of physician visits or hospital patient days, is used to measure the quantity demanded of medical services. Proxy variables are then included as independent variables to control for variations in quality. A failure to properly control for quality biases the results. That is because changes in demand may be attributed to changes in other variables when in fact they are the result of differences in the quality of care provided. The measurement of the out-of-pocket price of medical care also presents a problem for economists. This problem has become more severe in recent years given the increasing role of third-party payers. In a perfect world, the out-of-pocket price of medical services should equal the amount the consumer pays after the impact of insurance has been considered. Unfortunately, such data are rarely available, and economists often have to resort to using such variables as the average price of medical services rendered. An additional variable is then included in the equation to control for the presence of health insurance. The price variable should negatively affect the demand for medical care, while the presence of insurance should positively influence quantity demanded. An income variable is included to capture the impact of purchasing power on demand, while time cost variables control for the effects of travel and waiting costs on demand. We expect the income variable to have a positive effect on demand and the time cost variables to have a negative impact. The prices of various substitutes and complements in consumption should also be included in the regression equation. This has become even more important in recent years as medical markets have become more interrelated. For example, if we are trying to assess the quantity demanded of inpatient services at a hospital, we should control for the prices of hospital outpatient services (potentially a substitute service) and physician services (potentially a complementary service). The remaining factors (tastes and preferences, rate of health depreciation, stock of health, and quality of care) are referred to as control variables and capture the impact that various noneconomic factors may have on the demand for health care services.
Own-Price, Income, Cross-Price, and Time-Cost Elasticity Estimates Overall, the empirical literature on the elasticity of demand for primary health care is rich and spans the globe. Table 5-2 provides just a sample of the studies on the topic. Although the range of price elasticity estimates is broad, studies tend to find the demand for primary health care to be relatively inelastic. For example, studies using medical expenditures as the dependent variable find the own-price elasticity of demand to vary from -0.04 to -0.7. Other studies that look at the demand for hospital and physician services find similar results. Taken as a whole, the estimates suggest that the own-price elasticity of demand for primary health care hovers from -0.1 to -0.7, which means that a 10 percent increase in the out-ofpocket price of medical services leads to a 1 to 7 percent decrease in the quantity demanded. The inelastic estimates also imply that total expenditures on hospital and physician services increase with a greater out-of-pocket price, ceteris paribus.
In general, the research indicates that the demand for other types of medical care is slightly more price elastic than the demand for primary care. That is not at all surprising given that the percentage of out-of-pocket payments tend to be the lowest for hospital and physician services. Everything else held constant, consumers should become more price sensitive as the portion of the bill paid out of pocket increases. For example, Manning and Phelps (1979) found the demand for dental services to be slightly more price elastic and to vary by type of service provided and the sex and age of the patient. The price elasticity of demand for dental services by adult females appears to vary between -0.5 and -0.7, and the demand for dental services by adult males and children seems to be slightly more price elastic. The demand for nursing home services also appears to be more price elastic than primary medical services. Chiswick (1976) found the own-price elasticity for nursing home services to equal -2.3, and Lamberton et al. (1986) estimated that it equals -0.76. Finally, Headen (1993) found the own-price elasticity for the probability of entering a nursing home to be -0.7. The empirical estimates for the income elasticity of demand vary widely and merit discussion. Studies using household, or individual, data generally find health care to be a normal good with an income elasticity below 1.0. These results are in direct contrast to studies that utilize country-level data to look at the relation between income and health care expenditures either over time or across countries. The goal of these studies is to ascertain how economic growth impacts national health care expenditures. Generally, these studies find the aggregate income elasticity to be slightly above 1. For example, Newhouse (1977) finds the income elasticity to range between 1.13 and 1.31, while Parkin et al. (1987) estimate the rate to be slightly below +1. Finally, Leu (1986), Gerdtham et al. (1992), and Murray et al. (1994) agree with Newhouse and find the aggregate income elasticity to be above 1. This difference between the micro and macro estimates is interesting and deserves explanation. According to Newhouse, the difference exists because, for example, within the United States at any point in time the average consumer pays only a small portion of the price of medical care (approximately 14 percent in 2003), while over time the country as a whole must pay the full price of health care. As the out-of-pocket price of health care falls for the average consumer, the income elasticity should also fall because the consumer is less conscious of price. For example, if the out-of-pocket price of health care falls to zero, then the average individual is going to consume health care regardless of income. The income elasticity in the extreme equals zero. The country, as a whole, however, must face the entire burden of the cost of health care and, as a result, is going to be much more sensitive to price and income. One of the more interesting questions concerning this research has to do with whether health care is a luxury good. Economists define a luxury good as one that has an income elasticity above 1.0. In this case, an increase in income leads to an even larger increase in the quantity consumed of the good. For example, assume that the income elasticity of a good equals 1.5. In this case, a 10 percent increase in income leads to a 15 percent increase in the consumption of the good. Naturally, this means that the portion of one’s budget allocated to the consumption of the good also increases with income. If the aggregate income elasticity of health care is above 1.0, this may provide a demand-side explanation as to why health care expenditures in the United States as a portion of national income have increased over the past few decades. As the U.S. economy grew over the past few decades and income per capita expanded, the nation allocated a greater portion of its income to health care because it is a luxury good. Consequently, the health care sector received a larger slice of the economic pie. Time costs also appear to have a significant impact on the demand for medical services. In fact, research indicates that the travel time elasticity of demand is approximately equal to the own-price elasticity of demand. According to Acton (1975) and Phelps and Newhouse (1974), the travel time elasticity of demand ranges from -0.14 to -0.51. Using a data set generated in the United Kingdom, Gravelle et al. (2002) found elasticity of admissions with respect to distance to equal -0.35. Taken together, these studies suggest that a 10 percent increase in travel time reduces the quantity demanded for medical services by roughly 3 percent. It also appears that consumers place a value on the time spent waiting for medical services. McCarthy (1985) found the wait time elasticity to range from -0.36 to -1.14, while Martin and Smith (1999) found it to equal -0.20. In addition, Gravelle et al. (2002) estimated the elasticity of admissions with respect to waiting time to equal -0.25. Time costs also influence the decision to acquire medical care; Frank et al. (1995) found the elasticity of travel time costs on the probability of a timely completion of childhood immunization to be roughly -0.08. The extent to which various types of medical services serve as substitutes or complements in consumption is not clear at this time. For example, there appears to be little consensus as to whether inpatient and outpatient hospital services are substitutes or complements. Davis and Russell (1972) found the cross-price elasticity between the price of inpatient services and number of outpatient visits to vary between 0.85 and 1.46, indicating that they are substitutes. These results were later qualitatively confirmed by Gold (1984). Thus, as the price of inpatient services at a hospital increases, consumers rely more on outpatient services to save money. Freiberg and Scutchfield (1976), on the other hand, found that no substitution occurs between these two types of hospital services. At the other extreme, Manning et al. (1987) suggested that they are complements in consumption. A similar debate in the literature concerns whether physician and hospital inpatient or outpatient services are substitutes or complements.
The Impact of Insurance on the Demand for Medical Care The growth of health insurance, both public and private, has had a profound impact on the demand for medical care. Instead of reviewing the results from the many studies that analyzed the impact of insurance on the demand for health care, we will focus on a study conducted by the RAND Corporation (Manning et al., 1987). The RAND Health Insurance Study (HIS) is without doubt the most comprehensive study to date. Families from six sites (Dayton, Ohio; Seattle, Washington; Fitchburg, Massachusetts; Charleston, South Carolina; Georgetown County, South Carolina; and Franklin County, Massachusetts) were enrolled in various types of health insurance plans in a controlled experiment to test the impact of differences in insurance coverage on the demand for medical care. The present discussion focuses on the results published by Manning et al. (1987). However, a number of other articles analyze the data from the RAND HIS study. Among them are Newhouse et al. (1981), Keeler and Rolph (1983), O’Grady et al. (1985), Manning et al. (1985), Leibowitz et al. (1985b), Leibowitz et al. (1985a), and Manning et al. (1986). For a summary of the entire RAND HIS study, consult Newhouse and the Insurance Experiment Group (1993). The category excludes visits for radiology, anesthesiology, or pathology services. The third, fourth, and fifth columns list, respectively, total expenditures per person for outpatient, inpatient, and all medical services, excluding dental care and psychotherapy. The sixth column indicates the probability of using any medical services over the course of the year. The results largely confirm our expectations concerning the impact of coinsurance on the demand for health care. As the level of coinsurance rises, or the out-of-pocket price of medical care increases, consumers demand less medical care. The number of face-to-face visits decreased from 4.55 per year when health care was a free good to 2.73 when the consumer paid 95 percent of the bill. This represents a decrease in visits of 40 percent. The largest drop in visits took place between the free plan and the 25 percent coinsurance plan. This overall decrease in visits was matched by an identical drop in outpatient expenses from $340 to $203 per year. According to Manning et al. (1987), this indicates that as the out-ofpocket price of medical care increases, consumers reduce medical expenditures largely by cutting back on the number of visits to health care providers and not on the amount spent on each visit. It is interesting to note that the authors reported no significant differences in the amount spent on inpatient services across plans. This, they concluded, was the result of the $1,000 cap put on out-of-pocket expenditures. In 70 percent of the cases where people were admitted for inpatient services, the cost exceeded the $1,000 limit. The last two columns in Table 5-3 also largely support our expectations regarding the impact of insurance on the demand for medical services. In every case, as the level of coinsurance increased, the probability of using any medical services, along with total medical expenditures, diminished. The only exception occurred between the 25 and 50 percent coinsurance rates for total medical expenditures. In one phase of the study, families were randomly assigned to fourteen different fee-for-service plans. The plans varied in terms of the consumer coinsurance rate and the upper limit on annual out-of-pocket expenses. Every plan had a maximum limit of $1,000 in out-of-pocket expenses per year. Table 5-3 presents selected results for five of the plans: free (0 coinsurance rate), 25 percent coinsurance rate, 50 percent coinsurance rate, 95 percent coinsurance rate, and individual deductible. The individual deductible plan had a 95 percent coinsurance rate for outpatient services, subject to a limit of $150 per person or $450 per family, and free inpatient care. Essentially, an individual or a family with this plan receives free medical care after meeting the deductible for outpatient expenditures. In Table 5-3, face-to-face visits equal the number of visits per year to a medical provider, such as a physician.
* The chi-square test was used to test the null hypothesis of no difference among the five plan means. In each instance, the chi-square statistic was significant to at least the 5 percent level. The only exception was for inpatient dollars.
Finally, the results from the individual deductible plan illustrate the negative impact of deductibles on the consumption of medical care. In every instance, less medical care was consumed with the deductibles than would have been the case if medical care had been a free good. It seems that individuals with this plan consumed medical services at a rate somewhere between the 25 and 95 percent coinsurance rate. The results also indicate that the own-price elasticity of demand is sensitive to the level of insurance. When the level of coinsurance ranged from 25 to 95 percent, the elasticities of demand for all care and outpatient care were calculated as -0.14 and -0.21. These numbers decreased to -0.10 and -0.13 when the level of coinsurance ranged from 0 to 25 percent. This makes economic sense. As the level of coinsurance drops, consumers become less sensitive to price changes due to lower out-of-pocket payments. In conclusion, the results from the RAND HIS study point to the significant impact of health insurance on the demand for medical care. It is apparent that if either the rate of coinsurance or the deductible falls, the amount of health care consumed increases.
The Impact of Noneconomic Factors on the Demand for Medical Services The empirical research also indicates that a host of other factors, such as tastes and preferences or the stock of health, affect the demand for medical care. Researchers generally agree that age and severity of illness directly influence the demand for medical care, while the overall health of the individual inversely affects the demand for care. There does not, however, appear to be a consensus concerning the impact of education on the demand for health care. This may indicate that the direct impact of education on the demand for medical care (a greater willingness to seek care) is offset by the inverse effect (a greater ability to produce health care at home) or that more research needs to be done in this area. It is interesting to note that a few researchers have focused specifically on the effect of medical knowledge on the demand for medical care. Unlike the results for general education, a positive relationship appears to exist between consumers’ medical knowledge and the demand for medical care. This means that consumers with a more extensive background in medicine tend to consume more medical services. For example, Kenkel (1990) found that consumers’ medical knowledge is positively related to the probability of visiting a physician for medical care, while Hsieh and Lin (1997) uncovered that those elderly who had a greater understanding of health were more likely to acquire preventive medical care. Both studies suggest that consumers with a lack of medical knowledge tend to underestimate the impact of medical care on overall health and, as a result, fail to consume an appropriate amount. It may also be the case that more medical information enhances the ability of an individual to effectively consume medical care, causing the marginal product of medical care to increase (consult Topic 2). As a result, the demand for various types of medical care increases with consumer information. Finally, Hsieh and Lin (1997) found that years of schooling, whether the individual worked in the health care field, medical insurance, and income all positively influenced the level of health information acquired. They also found that age and whether the individual drank or smoked inversely affected the quantity of health information collected. It appears that older people acquire less new knowledge because they have fewer years to live and reap any reward from that knowledge, while individuals who drink or smoke receive less utility from any good health that may result from added medical knowledge. Every day decisions are made in the health care sector concerning the best, or most efficient, amount of medical care to provide. At some juncture in the decision-making process, the all-important question becomes: At what point do the added costs of providing more medical care outweigh the benefits in terms of improved health? In practice, the answer to this question is complex because costs and benefits depend on such factors as the availability of medical resources, patient preferences, and the severity of illnesses. Consider an adult who complains to his or her physician about chest pains during an annual physical exam. The first thing the physician must do is determine the seriousness of the problem. The pain could simply be the result of stress or could be a sign of more serious trouble, such as an impending heart attack (remember Joe at the beginning of Topic 1?). When confronted with a patient's chest pains, a physician faces several options. For example, one clinical professor of medicine says, To assess chest pain . . . we can take a history and a physical examination for $100; do an exercise test for $500; perform a nuclear stress test for $1,500; or do coronary angiography for $5,000. Each escalation in diagnostic approach improves the accuracy of diagnosis from 50 percent to 60 to 80 to 100 percent. (Rubenstein, 1994) Basically, the best medical procedure is chosen by comparing the incremental costs of progressively more expensive medical tests with the benefits of additional medical information provided by greater diagnostic capabilities.
Cost Identification Analysis The first type of analysis we will consider is cost identification. Generally speaking, cost identification studies measure the total cost of a given medical condition or type of health behavior on the overall economy. The total cost imposed on society by a medical condition or a health behavior can be broken down into three major components: 1. Direct medical care costs 2. Direct nonmedical costs 3. Indirect costs Direct medical care costs encompass all costs incurred by medical care providers, such as hospitals, physicians, and nursing homes. They include such costs as the cost of all necessary medical tests and examinations, the cost of administering medical care, and the cost of any follow-up treatments. Direct nonmedical costs represent all monetary costs imposed on any nonmedical care personnel, including patients. For the patient, direct nonmedical costs include the cost of transportation to and from the medical care provider, in addition to any other costs borne directly by the patient. For example, the patient may require home care or have specific dietary restrictions. Others may also be influenced by the treatment. For example, the cost of instituting a substance abuse program in the workplace includes not only the direct medical costs of drug and alcohol rehabilitation but also any nonmedical costs the firm incurs while implementing and overseeing the program. Family members may be financially affected as well. Indirect costs consist primarily of the time costs associated with implementation of the treatment. Indirect costs include the opportunity cost of the patient’s (or anyone else’s) time that the program affects, especially because many health behaviors and medical conditions result in lost productivity due to injury, disability, or loss of life. Consider the substance abuse program previously discussed. Costs should reflect the opportunity costs of the time needed to educate workers about the potential dangers of substance abuse. The time cost is borne by the employer and equals the value of forgone production. By and large, cost identification studies consider the direct medical care and indirect costs associated with medical actions or adverse health behaviors. For example, Druss et al. (2001) estimated the total economic cost of five chronic medical conditions – mood disorders, diabetes, heart disease, asthma, and hypertension – in 1996. In their estimates, the authors considered medical costs as well as work loss. Out of the five conditions, hypertension was by far the most costly medical condition with a total cost of $121.8 billion annually, of which slightly more than 90 percent was accounted for in health care costs. The next largest was mood disorders, $66.4 billion, followed by diabetes, $57.6 billion, heart disease, $42.4 billion, and asthma, $31.2 billion. In another study, Meltzer et al. (1999) estimated that an influenza pandemic in the United States would result in 89,000 to 207,000 deaths, 314,000 to 734,000 hospitalizations, 18 to 42 million outpatient visits, and 20 to 47 million other illnesses. The economic impact of such an outbreak would be between $71.3 and $165.5 billion. The American Diabetes Association (2008) estimated the direct and indirect costs of diabetes in 2007 at $174 billion, with $116 billion attributed to direct medical costs and the remaining $58 billion to indirect expenses such as lost work days and permanent disability. The American Heart Association set the cost of cardiovascular disease and stroke at $448.5 billion in 2008. Finally, Sobocki et al. (2006) estimated the cost of depression in Europe at 118 billion euros with direct medical costs accounting for 42 billion euros. Cost identification studies like these are enlightening because they provide a sense of the total costs associated with various medical conditions or health behaviors. However, they provide little guidance for decision making. For example, what is the best, or most efficient, method to treat Alzheimer’s disease? To answer questions like this, we must turn to other types of decision-making techniques, such as cost-benefit and cost effectiveness analysis.
Cost-Benefit Analysis As we learned in Topic 1 with the introduction of PPC analysis, resource scarcity forces society to make choices. For example, an entire economy must collectively decide how much medical care to produce and who will receive it, while each health care provider must determine the most appropriate method to produce health care services. Even the consumer who has complete medical insurance coverage faces scarcity and choices because time is a finite commodity. The consumer must decide whether the time needed to make a doctor’s appointment, travel to the physician’s office, and receive medical services is worth the value of foregone activities. Thus, scarcity necessitates choice. Economics is the social science that analyzes the process by which society makes these choices. Economists treat people as rational decision makers. Rationality means people know how to rank their preferences from high to low or best to worst. It also means that people never purposely choose to make themselves worse off. Consequently, it stands to reason that people will make choices based on their self-interests and choose those activities they expect will provide them with the most net satisfaction. Pursuing self-interest does not mean people are always selfish, however. For example, giving money to a charity, or volunteering one’s time at a local hospital, gives even the most devout good samaritan a considerable amount of pleasure. The decision rule people follow when choosing activities is straightforward and involves an assessment of the expected benefits and costs associated with each choice. If expected benefits exceed expected costs for a given choice, it is in the economic agent’s best interest to make that choice. In formal terms, the optimizing rule looks like this: (3-1) NBe(X) = Be(X) - Ce(X) where × represents a particular choice or activity under consideration, Be stands for the expected benefits associated with the choice, Ce equals the expected costs resulting from the choice, and NBe represents the expected net benefits. If NBe is larger than zero, the economic agent’s well-being is enhanced by choosing the activity. The fact that you are reading this course indicates the book’s expected benefits outweigh its expected costs (unless, of course, your professor forced you to buy and read it). That is, you expect this book to provide benefits in excess of the money you spent on it, plus the forgone use of your time. Nonreaders of this book obviously believe the costs outweigh the benefits, or that NBe is negative. Formal cost-benefit analysis utilizes the same net benefit calculus to establish the monetary value of all the costs and benefits associated with a given health policy decision. Such information is invaluable to policy makers who are under pressure to utilize scarce resources to generate the most good for society. To illustrate this point, let’s suppose that an all-knowing benevolent dictator, called the “surgeon general,” is responsible for ensuring the economic happiness of the people in some hypothetical society. The surgeon general realizes that people possess unlimited wants and that numerous goods and services, such as food, clothing, housing, medical care, and automobiles, provide them with satisfaction. The surgeon general also knows that scarcity of resources involves tradeoffs; that is, more of one good means less of the others. The surgeon general’s task is to maximize the social utility of the population by choosing the best aggregate mix of goods and services to produce and consume. In the context of the production possibilities curve, the surgeon general is trying to find the specific point that maximizes the collective wellbeing of the population. The surgeon general is assumed to accept the current distribution of income. To accomplish this objective, the surgeon general has the power to allocate land, labor, and capital resources to any and all uses. Consistent with the maximization of the social utility received from all goods and services, we can think of the surgeon general as trying to maximize the total net social benefit (TNSB) from each and every good and service produced in the economy. The TNSB derived from a good or service is the difference between its total social benefit (TSB) in consumption and its total social cost (TSC) of production. The difference represents the net benefit, or gain, that the society receives from producing and consuming a particular amount of some good or service. The TSB can be treated as the money value of the satisfaction generated from consuming the good or service. The TSC can be looked at as the money value of all the resources used in producing the good or service. For example, the TNSB from medical services can be written as (3-2) TNSB(Q) = TSB(Q) - TSC(Q). Equation 3-2 allows for the fact that the levels of benefits, costs, and net social benefit depend on the quantity of medical services, Q. The surgeon general maximizes TNSB by choosing the quantity of medical services at which the difference between TSB and TSC reaches its greatest level. Figure 3-1 presents a graphical representation of this maximization process. Notice in the figure that total social benefits increase at a decreasing rate with respect to the quantity of medical services. This shape reflects an assumption that people in society experience diminishing marginal benefit with respect to medical services and indicates that successive incremental units generate continually lower additions to social satisfaction. TSCs increase at an increasing rate and reflect the increasing marginal costs of producing medical services. The slope of the TSB curve can be written as (3-3) MSB(Q) = ΔTSB/ΔQ, where MSB stands for the marginal social benefit from consuming a unit of medical services. Obviously, MSB decreases with quantity since the slope of the TSB curve declines due to diminishing marginal benefit. Similarly, the slope of the TSC curve is (3-4) MSC(Q) = ΔTSC/ΔQ, where MSC represents the marginal social cost of producing a unit of medical services. MSC increases with output as the slope of the TSC curve gets steeper due to increasing marginal cost.
The TSB curve represents the monetary value of the total social benefit generated from consuming medical care. The curve is positively sloped to reflect the added monetary benefits that come about by consuming more medical care. The curve bows downward to capture the fact that society experiences diminishing marginal benefit with regard to medical care. The TSC curve represents the TSC of producing medical care and is upward sloping because total costs increase as more medical care is produced. The curve bows toward the vertical axis because the marginal cost of producing medical care increases as more medical care is produced. TNSB is maximized when the vertical distance between the two curves is greatest and that occurs at Q 0 level of medical services.
TNSB is maximized where the vertical distance between the two curves is the greatest at distance AB. A common principle in geometry is that the distance between two curves is maximized when their slopes are equal. That condition holds at output level Q0 and implies that allocative efficiency, or the best quantity of medical services, results where (3-5) MSB(Q) = MSC(Q). Thus, the surgeon general chooses output Q0 because it maximizes TNSB. To illustrate this point in a slightly different manner, Figure 3-2 graphs the MSB and MSC curves. Notice that the negatively sloped MSB and the positively sloped MSC reflect diminishing marginal benefit and increasing marginal costs, respectively. The efficient amount of medical services is at Q0 in Figure 3-2 because MSB equals MSC. Let us consider why Q0 is the efficient or best level of medical services by examining the figure more closely. In the figure, units of medical services to the left of Q0, such as QL, imply that too few medical services are being produced because MSB (point E) is greater than MSC (point F). At Ql, an additional unit of medical services generates positive additions to TNSB because the net marginal social benefit, the difference between MSB and MSC, is positive. Society is made better off if more medical services are produced. At Q0, where MSB equals MSC, the net marginal social benefit is equal to zero and TNSB is maximized. The MSB curve stands for the marginal social benefit generated from consuming medical care and is downward sloping because of the notion of diminishing marginal benefit. The MSC curve stands for the marginal social cost of producing medical care and is upward sloping because of increasing marginal costs. TNSB is maximized at Q0 level of medical care where the two curves intersect. At that point, the MSB of consuming medical care equals the MSC of production. If QL amount of medical care is produced, then the MSB exceeds the MSC and society would be better off if more medical services were produced. If Qr amount of medical care is produced, then the MSB is less than the MSC and too much medical care is produced. In contrast, output levels to the right of Q0 suggest that too many medical services are being produced. For example, at QR, MSC (point G) exceeds MSB (point H) and net marginal social benefit is negative, subtracting from maximum total net social benefits. The cost of producing unit QR exceeds the benefits at the margin, and society could be made better off by not producing this unit. This same argument applies to all units of medical services to the right of Q0. TNSB is represented by the area below the MSB curve but above the MSC curve in Figure 3–2. This is because TNSB is equal to the sum of the net marginal social benefits, or the difference between MSB and MSC for every unit of medical services actually produced. Thus, in Figure 3–2, the area ABC represents the maximum TNSB that society receives if resources are allocated efficiently. (Conceptually, this area is equal to the vertical distance AB in Figure 3-1.) If the surgeon general decides to produce QL instead of Q0 units of medical services, society fails to receive the part of the TNSB indicated by area ECF. In economics, the lost amount of net social benefits is referred to as a deadweight loss. In this example, it measures the cost associated with an underallocation of resources to medical services. Similarly, if the surgeon general chooses to produce QR units of medical services, a deadweight loss of area GCH results. Area GCH indicates the net cost to society from producing too many units of medical services and therefore too few units of all other goods and services. The preceding discussion can be easily couched in terms of the net benefit calculus in Equation 3-1. For example, if we solve Equation 3-5 for the difference between the marginal social benefit and the marginal social cost, we get (3-6) NMSB(Q) = MSB(Q) - MSC(Q), where NMSB equals the net marginal social benefit the society derives from consuming a unit of the good. If NMSB is larger than zero, total net social benefit increases if an additional unit of the good is consumed. Naturally, if NMSB is negative, the society is made worse off if an additional unit of the good is produced and consumed.
The Practical Side of Using Cost-Benefit Analysis to Make Health Care Decisions Public policy makers concerned with formulating health policies that affect the overall well-being of society, or TNSB, must wrestle with the problem of operationalizing Equation 3-6. That is no easy task, as it requires they establish the monetary value of all the costs and benefits associated with a given health policy decision. The problem is complicated by the fact that some of the costs and benefits may be of an indirect nature and therefore difficult to quantify. For example, suppose you are responsible for estimating the net benefits associated with a rehabilitation program that requires one hour of exercise a day for people who recently had a heart bypass operation. One of the costs you will have to measure is the opportunity cost of the patients’ time. Your first inclination may be to base your estimate on the average hourly wage of the people in the program. But what if the people conduct their daily exercise regime on their own time rather than while at work? You now face the problem of determining the opportunity cost of leisure time. Although no hard-and-fast rule exists, the opportunity cost of leisure time is most often estimated at some fraction, usually one-half, of the average hourly wage. As you can see, indirect costs or benefits may be hard to quantify. The benefits, or diverted costs, of a medical intervention fall into four broad categories: 1. The medical costs diverted because an illness is prevented. 2. The monetary value of the loss in production diverted because death is postponed. 3. The monetary value of the potential loss in production saved because good health is restored. 4. The monetary value of the loss in satisfaction or utility averted due to a continuation of life or better health or both. The first benefit is usually the easiest to calculate and involves estimating the medical costs that would have been incurred had the medical treatment not been implemented. The next two benefits involve projecting the value of an individual’s income that would be lost due to illness or death. The last benefit is the most subjective and therefore the most difficult to quantify, because it involves estimating the monetary value of the pleasure people receive from a longer life and good health. For example, how does one attach a dollar value to the decrease in pain and suffering an individual may experience after hip replacement surgery? Or what is the monetary value of the satisfaction a parent receives from watching a child grow up? Given the difficulty involved in measuring the pleasure of life, many studies simply calculate the other three types of benefits. The resulting figure is considered to reflect a lower-bound estimate of total benefits. In this simple example, we considered the costs and benefits associated with a new medical treatment where one never existed before. As a result, we considered the total costs and benefits experienced by society. In some instances, however, that approach is not appropriate. Consider a new medical treatment that potentially displaces, or complements, an existing one. In this situation, the appropriate practice is to focus on the incremental, or marginal, costs and benefits associated with the new treatment rather than the total costs and benefits. As such, only the added costs and benefits of the new treatment are considered.
Discounting The costs and benefits of any medical decision are likely to accrue over time rather than at a single point in time. For example, the benefits of a polio vaccination are felt primarily in terms of allowing children who might otherwise have been afflicted with polio to lead normal, healthy, active lives. The benefits in this case accrue over many decades. Therefore, an adjustment must be made to account for the fact that a benefit (or a cost) received today has more value than one received at a future date. That is, the net benefit of an activity yielding a stream of future returns must be expressed in present value (PV) terms before proper comparisons can be made. In simplest terms, PV means that an individual prefers $100 today rather than a year from now. Even if the individual wants to spend the money a year from now, he or she is still made better off by accepting the money today. For example, $100 deposited in a savings account offering a 4 percent annual return yields $104 a year later. We say that the PV of $104 to be received a year from now at a 4 percent rate of interest equals $100. In more formal terms, we can state PV using the following equation:
where F equals a fixed sum of money and r represents the annual rate of interest, or the rate at which the sum is discounted. In our example, F equals $104 and r is 4 percent, or 0.04, so PV equals $100. Notice that a higher interest rate means the PV of a fixed sum falls. For example, if the rate of interest increases to 5 percent, the PV of $104 decreases to $99.05. Thus, the PV of a fixed sum is inversely related to the rate at which it is discounted. When referring to sums of money received over a number of periods, the PV formula becomes slightly more complicated. If different sums of money, or net benefits, are to be received for a number of years, n, at the close of each period, the formula looks like the following:
where Ft (t = 1, 2, 3, . . . , T) equals the payment, or net benefit, received annually for T years. For simplicity’s sake, we normally assume the discount rate is fixed over time. Each annual payment is expressed in today’s dollars by dividing it by the discounting factor. The discounting factor equals 1 plus the rate of interest raised to the appropriate power, which is the number of years in the future when the payment is to be received. The sum total, or PV, represents the present value of all annual payments to be received in the future. If Equation 3-8 is rewritten in summation form and specifically in terms of benefits and costs over time, it looks like the following:
where NB equals the PV of net benefits. In every cost-benefit study in which the effects of a medical treatment or project occur over time, careful consideration must be given to choosing the discount rate. That is because the rate at which future payments are deflated can profoundly affect the present value of a project, especially when the costs or benefits do not accrue until far into the future. The earlier polio vaccination example is a case in point. A cost-benefit analysis of a polio vaccination project involves taking the PV of benefits potentially received 70 years into the future (the average American can expect to live about 75 years). Selecting an interest rate that is too high results in the choice of medical interventions that offer short-term net benefits. Conversely, choosing an interest rate that is too low leads to the choice of medical projects that provide long-term net benefits. Theoretically, the chosen interest rate should equal the rate at which society collectively discounts future consumption, or society’s time preference. In an industrial economy, however, there are many interest rates to choose from, including the prime business lending rate, the residential mortgage rate, and the U.S. government bond or T-bill rate. So naturally, the “correct” interest rate is open to interpretation. Most studies choose a discount rate of between 3 and 5 percent or look to private financial markets for guidance. In the latter instance, the interest rate on government bonds is the typical choice. The T-bill interest rate is chosen because it supposedly represents a risk-free rate of return and therefore reflects the rate at which the private sector discounts future streams of income in the absence of risk. Some studies circumvent this problem by presenting a range of estimates based on alternative rates of interest. It is then left to the ultimate decision maker to choose the appropriate rate of discount.
The Value of Life To properly estimate the total benefits of a medical intervention, we must be able to measure the value of a human life, because many medical interventions extend or improve the quality of life. The most common method used to determine the monetary worth of a life is the human capital approach. Economists view expenditures on education and health as personal investments that enhance an individual’s ability to command a higher salary in the marketplace; hence the term human capital. The human capital approach essentially equates the value of a life to the market value of the output produced by an individual during his or her expected lifetime. The technique involves estimating the discounted value of future earnings resulting from an improvement in or an extension of life.
Figure 3-3 provides some average estimates of the PV of lifetime earnings (including fringe benefits) by age and gender, discounted using a 3.0 percent discount rate. Notice that the discounted value of lifetime earnings initially increases with age and then decreases. The PV figures increase at first because as an individual ages beyond infancy, the value of lifetime earnings that accrue mainly in the middle adult years are discounted over a shorter period of time. For both males and females, the discounted value of lifetime earnings peaks in the between the ages of 20 and 24, $1.52 million for males and $1.09 million for females. Eventually, lifetime earnings decrease with age as productivity and the number of years devoted to work decrease. The estimates are also sensitive to the discount rate. For example, if the figures were discounted at a 5 percent rate rather than a 3 percent rate (earnings figures not shown in Figure 3-3), the present value of lifetime earnings for males between the ages of 20 and 24 falls to $1.06 million and for women in the same age group they fall to $775, 711. Naturally, the higher the discount rate the lower the discounted value of lifetime earnings. Although the human capital approach is the most widely accepted method for determining the value of a life, the technique is not without shortcomings. One concern is that the approach is unable to control for labor market imperfections. For example, from Figure 3-3, it is apparent that the discounted value of lifetime earnings for males is substantially greater than that for females. Gender discrimination in the workplace may account for part of the difference. As a result, women may be penalized and assigned a lower value of life because of their gender. Also, racial and other forms of discrimination may result in an inappropriate estimate of the value of life when the human capital approach is used. The human capital approach can also be criticized because it fails to consider any nonmarket returns the individual might receive from other activities, such as leisure. As such, it does not take into account the value of any pain and suffering averted because of a medical treatment, nor does it consider the value an individual receives from the pleasure of life itself. For example, take an extreme view. According to the human capital approach, a chronically unemployed person has a zero or near-zero value of life. An alternative approach used to measure the value of a life is the willingness-to-pay approach. The willingness-to-pay approach is based on how much money people are willing to pay for small reductions in the probability of dying. This kind of information is revealed when, for example, people install or fail to install smoke detectors in their homes, wear or do not wear automobile seat belts, or smoke or do not smoke cigarettes. For example, assume that people in society choose to spend $100 per person per year on some device that improves environmental quality and reduces the probability of a person dying by 1 in 10,000. In this case, the imputed value of the average person’s life equals $1 million ($100 4 1/10,000). To understand how the willingness-to-pay approach works, consider a person who is deciding whether to purchase a potentially life-saving medical service. The benefit of the life-saving medical service equals the reduced probability of dying, p, times the value of the person’s life, V Using a cost-benefit approach, the “marginal” person purchases the medical service if the benefit, p x V, just compensates for the cost, C, or (3-10) × V = C, although “inframarginal” consumers might perceive greater benefits because they value their lives more highly. Dividing both sides by p results in (3-11) V = C/. Equation 3-11 implies that a lower-bound estimate can be calculated for the value of a human life by dividing the cost of a life-saving good or service by the reduced probability of dying. The advantage of the willingness-to-pay approach is that it measures the total value of life and not just the job market value. The imputed value of life generated by the willingness-to-pay approach includes the value of forgone earnings plus the nonmarket value received from life and good health. As a result, the willingness-to-pay approach generally estimates the value of a life to be higher than that generated by the human capital approach. For example, based on a survey conducted in 1999, Alberini et al. (2002) estimated the mean value of a statistical life to equal $933,000 in Canada and $1.5 million in the United States for a 5 in 1,000 reduction in risk. The mean estimates jumped to $3.7 million in Canada and $4.8 in the United States for a 1 in 1,000 reduction in risk. Viscusi (1993) found the willingness-to-pay estimates to range between $3 and $7 million in 1990 dollars, while Mrozek and Taylor (2002) reviewed more than 40 studies and found the statistical value of a life to be between $1.5 and $2.5 million in 1998 dollars. All indications are that the willingness-to-pay estimates are higher than the human capital estimates. Keeler (2001) recently illustrated how the human capital approach can be reconciled with the willingness-to-pay approach by estimating the discounted value of life and considering the monetary value of all time, not just work time. As such, he estimated the value of life by assuming that all time is valued at the market wage rate and controlling for the total number of hours remaining for an individual at a given age, rather than simply the remaining number of work hours. Given that the average worker under 50 years of age is likely to spend only between one-tenth and one-fifth of future hours working, you can imagine how this increased the discounted value of a remaining life. For a 30-year-old male, Keeler estimated the value of all future hours to equal slightly more than $2.6 million in 1990 dollars, which is more than five times the discounted value of future earnings, and in line with willingness-to-pay estimates. While his figures are crude, they illustrate that people place a significant monetary value on the amount of time spent outside work, and that researchers need to consider that when estimating the value of a life.
Summary Economic theory suggests that the demand for medical care represents a derived demand because it is but one input in the production of health. As a result, the utility received from consuming medical care is in the form of the satisfaction that accrues from improvements in the stock of health. Utility analysis also indicates that the quantity demanded of health care is inversely related to price because improvements in health are subject to diminishing returns. The demand for medical care, like the demand for many other services, depends on the out-of-pocket price, income, the prices of substitutes and complements, and time costs, along with a host of noneconomic factors, such as tastes and preferences, quality of care, and the state of health. Economists use the concept of elasticity to measure the degree to which an economic agent, such as a consumer, adjusts to a change in the value of an independent variable. The most common elasticity is the own-price elasticity of demand, which measures the extent to which consumers react to a change in the price of a good or service. In mathematical terms, it equals the percentage change in quantity demanded divided by the percentage change in price. If the demand for a product is elastic, the consumer’s willingness to purchase the product is very sensitive to a price change. On the other hand, if the demand for the product is inelastic, price changes play a less significant role in determining overall demand. From a graphical perspective, the more elastic the demand for a product, the flatter the demand curve. Additional types of elasticities, such as the income elasticity of demand, have also been employed to assess how demand reacts to changes in variables other than own price. The empirical evidence indicates that the demand for medical care is inelastic with respect to price. Medical care also appears to be a normal good in that the demand for medical care increases with real income. In addition, time costs along with many noneconomic variables, such as age, gender, severity of illness, education, and consumer knowledge, influence demand. The evidence from the RAND HIS study verifies that health insurance plays a major role in determining the demand for medical care. As economic theory suggests, when the level of health insurance rises, the amount of medical care demanded increases while the price elasticity of demand becomes more inelastic.
Review Questions and Problems 1. In your own words, use utility analysis and production theory to explain why the demand curve for medical care is downward sloping. 2. After reading the topic on demand theory, a classmate turns to you and says, “I’m rather confused. According to economic theory, people demand a good or service because it yields utility. This obviously does not apply to medical services. Just last week I went to the dentist and had a root canal, and you can’t tell me I received any utility or satisfaction from that!” Explain to your classmate how utility analysis can be used to explain why he went to the dentist. 3. Use a graph to illustrate how the following changes would affect the demand curve for inpatient services at a hospital in a large city. A. Average real income in the community increases. B. In an attempt to cut costs, the largest employer in the area increases the coinsurance rate for employee health care coverage from 10 percent to 20 percent. C. The hospital relocates from the center of the city, where a majority of the people live, to a suburb. D. A number of physicians in the area join together and open up a discount-price walk-in clinic; the price elasticity of demand between physician services and inpatient hospital services is -0.50. 4. In recent years, many elderly people have purchased Medigap insurance policies to cover a growing Medicare copayment. These policies cover some or all of the medical costs not covered by Medicare. Use economic theory to explain how the growth of these policies is likely to influence the demand for health care by elderly people. 5. If you are covered by a private or a public insurance plan, obtain a pamphlet outlining the benefits provided and the cost of the plan. Are there any copayments or deductibles? If so, use economic theory to explain how they may influence your demand for medical care. 6. In your own words, explain what a fuzzy demand curve is. Why does it exist? What are its implications? 7. In reaction to higher input costs, a physician decides to increase the average price of a visit by 5 percent. Will total revenues increase or decrease as a result of this action? Use the concept of price elasticity to substantiate your answer. 8. You have just been put in charge of estimating the demand for hospital services in a major U.S. city. What economic and noneconomic variables would you include in your analysis? Justify why each variable should be included in the study, and explain how a change in each variable would likely affect the overall demand for hospital services. 9. Define own-price elasticity of demand, and explain how it is related to the demand curve. Provide four reasons why the demand for medical services is likely to be inelastic with respect to its price. 10. You are employed as an economic consultant to the regional planning office of a large metropolitan area, and your task is to estimate the demand for hospital services in the area. Your estimates indicate that the own-price elasticity of demand equals -0.25, the income elasticity of demand equals 0.45, the cross-price elasticity of demand for hospital services with respect to the price of nursing home services equals -0.1, and the elasticity of travel time equals -0.37. Use this information to project the impact of the following changes on the demand for hospital services. A. Average travel time to the hospital diminishes by 5 percent due to overall improvements in the public transportation system. B. The price of nursing home care decreases by 10 percent. C. Average real income decreases by 10 percent. D. The hospital is forced to increase its price for services by 2 percent. 11. According to Whitney et al. (1997), the price of dental services “decreased by $4.86 per day wait for a new-patient appointment and by $5.20 per minute wait in the reception room” (p. 783). Based on these findings, what would happen to the position of the demand curve for dental services if patients had to wait even longer for an appointment with a dentist? 12. A study estimates the demand for over-the-counter cough and cold medicines to be: Log Q = 0.885 - 0.744 log(P) - 0.50 log(INC) + 0.253 log(ADV) - 0.30 log(PHYSP) (5.52) (4.92) (1.40) (6.64) (0.99)
Adj. R2 = 0.30 N = 243 where Q = Annual dosages demanded of cough and cold medicines P = Price per dosage of cough and cold medicines INC = Average income of buyers ADV = Advertising expenditures on cough and cold medicines PHYSP = Market price of a physician visit t-statistics shown in parentheses below the estimated coefficient
All variables expressed in logarithms so the coefficient estimates can be interpreted as elasticities. A. Which of the estimated coefficients have signs contrary to theoretical expectations? Explain. Be specific in your explanation. B. Which coefficient estimates are statistically significant from zero at the 5 percent level or better? Explain. C. What percentage of the variation in dosages demanded remains unexplained? Explain. D. Suppose the price per dosage increased by 10 percent. By how much would dosages demanded change? Explain. Would total revenues to cold medicine producers increase or decrease? Explain.
References Aaron, Henry J. Serious and Unstable Condition: Financing America’s Health Care. Washington, D.C.: The Brookings Institution, 1991. Acton, Jan Paul. “Nonmonetary Factors in the Demand for Medical Services: Some Empirical Evidence." Journal of Political Economy 83 (June 1975), pp. 595-614. Bhattacharya, Jayanta, William B. Vogt, Aki Yoshikawa, and Toshitak Nakahara. “The Utilization of Outpatient Medical Services in Japan." Journal of Human Resources 31 (1996), pp. 450-76. Chiswick, Barry. “The Demand for Nursing Home Care." Journal of Human Resources 11 (summer 1976), pp. 295-316. Cockx, Bart, and Carine Brasseur. “The Demand for Physician Services: Evidence from a Natural Experiment.” Journal of Health Economics 22 (2003), pp. 881-913. Contoyannis, Paul, Jeremiah Hurley, Paul Grooten-dorst, Sung-Hee Jeon, and Robyn Tamblyn. “Estimating the Price Elasticity of Expenditure for Prescription Drugs in the Presence of Non-Linear Price Schedules: An Illustration from Quebec, Canada.” Health Economics 14 (2005), pp. 909-23. Cromwell, Jerry, and Janet B. Mitchell. “Physician-Induced Demand for Surgery.” Journal of Health Economics 5 (1986), pp. 293-313. Davis, Karen, and Louise B. Russell. “The Substitution of Hospital Outpatient Care for Inpatient Care.” Review of Economics and Statistics 54 (May 1972), pp. 109-20. Eichner, Matthew J. “The Demand for Medical Care: What People Pay Does Matter.” American Economic Review Papers and Proceedings 88 (May 1998), pp. 117-21. Feldman, Roger, and Bryan Dowd. “Is There a Competitive Market for Hospital Services?” Journal of Health Economics 5 (1986), pp. 272-92. Feldstein, Martin S. Economic Analysis for Health Services Efficiency. Amsterdam: North-Holland, 1967. Feldstein, Paul. Health Care Economics. New York: Wiley, 1988. Frank, Richard G., et al. “The Demand for Childhood Immunizations: Results from the Baltimore Immunization Study.” Inquiry 32 (summer 1995), pp. 164-73. Freiberg, Lewis, Jr., and F. Douglas Scutchfield. “Insurance and the Demand for Hospital Care: An Examination of the Moral Hazard.” Inquiry 13 (March 1976), pp. 54-60. Fuchs, Victor R., and Sarah Rosen Frank. “Air Pollution and Medical Care Use by Older Americans: A Cross Areas Analysis.” Health Affairs 21 (November/ December 2002), pp. 207-214. Gerdtham, Ulf-G, et al. “An Econometric Analysis of Health Care Expenditure: A Cross-Section Study of the OECD Countries.” Journal of Health Economics 11 (1992), pp. 63-84. Gold, Marsha. “The Demand for Hospital Outpatient Services.” Health Services Research 19 (August 1984), pp. 384-412. Gravelle, Hugh, Mark Dusheiko, and Matthew Sutton. “The Demand for Elective Surgery in a Public System: Time and Money Prices in the UK National Health Service.” Journal of Health Economics 21 (May 2002), pp. 423-49. Headen, Alvin E. “Economic Disability and Health Determinants of the Hazard of Nursing Home Entry.” Journal of Human Resources 28 (1993), pp. 80-110. Hsieh, Chee-ruey, and Shin-jong Lin. “Health Information and the Demand for Preventive Care among the Elderly in Taiwan.” Journal of Human Resources 32 (1997), pp. 308-33. Keeler, Emmett B., and John E. Rolph. “How Cost Sharing Reduced Medical Spending of Participants in the Health Insurance Experiment.” Journal of the American Medical Association 249 (April 22-29, 1983), pp. 2220-22. Kenkel, Don. “Consumer Health Information and the Demand for Medical Care.” Review of Economics and Statistics 72 (1990), pp. 587-95. Lamberton, C. E., W. D. Ellingson, and K. R. Spear. “Factors Determining the Demand for Nursing Home Services.” Quarterly Review of Economics and Business 26 (winter 1986), pp. 74-90. Leibowitz, Arleen, et al. “Effects of Cost-Sharing on the Use of Medical Services by Children: Interim Results from a Randomized Controlled Trial.” Pediatrics 75 (May 1985a), pp. 942-50. Leibowitz, Arleen, Willard G. Manning, and Joseph P. Newhouse. “The Demand for Prescription Drugs as a Function of Cost-Sharing.” Social Science and Medicine 21 (1985b), pp. 1063-69. Leu, Robert E. “The Public-Private Mix and International Health Care Costs.” In Public and Private Health Services, eds. A. J. Culyer and B. Jonsson. Oxford: Basil Blackwell, 1986. Manning, Willard, et al. “How Cost Sharing Affects the Use of Ambulatory Mental Health Services.” Journal of the American Medical Association 256 (October 1986), pp. 1930-34. Manning, Willard G., et al. “Health Insurance and the Demand for Medical Care: Evidence from a Randomized Experiment.” American Economic Review 77 (June 1987), pp. 251-77. Manning, Willard G., Howard L. Bailit, Bernadette Benjamin, and Joseph P. Newhouse. “The Demand for Dental Care: Evidence from a Randomized Trial in Health Insurance.” Journal of the American Dental Association 110 (June 1985), pp. 895-902. Manning, Willard G., and Charles E. Phelps. “The Demand for Dental Care.” Bell Journal of Economics 10 (autumn 1979), pp. 503-25. Martin, Stephen, and Peter C. Smith. “Rationing by Waiting Lists: An Empirical Investigation.” Journal of Public Economics 71 (January 1999), pp. 141-64. McCarthy, Thomas. “The Competitive Nature of the Primary-Care Physicians Service Market.” Journal of Health Economics 4 (1985), pp. 93-118. McMenamin, Peter. “What Do Economists Think Patients Want?” Health Affairs 9 (winter 1990), pp. 112-19. Miller, Lisa. “Medical Schools Put Women in Curricula.” The Wall Street Journal, May 24, 1994, p. B1. Mueller, Curt D., and Alan C. Monheit. “Insurance Coverage and the Demand for Dental Care.” Journal of Health Economics 7 (1988), pp. 59-72. Murray, C. J. L., R. Govindaraj, and P. Musgrove. “National Health Expenditures: A Global Analysis.” Bulletin of the World Health Organization 74 (1994), pp. 623-37. Newhouse, Joseph P. “Medical-Care Expenditures: A Cross-National Survey.” Journal of Human Resources 12 (winter 1977), pp. 115-24. Newhouse, Joseph P., et al. “Some Interim Results from a Controlled Trial of Cost Sharing in Health Insurance.” New England Journal of Medicine 305 (December 17, 1981), pp. 1501-7. Newhouse, Joseph P., and the Insurance Experiment Group. Free for All? Lessons from the RAND Health Insurance Experiment. Cambridge, Mass.: Harvard University Press, 1993. Nyman, John A. “The Private Demand for Nursing Home Care.” Journal of Health Economics 8 (1989), pp. 209-31. O'Grady, Kevin F., Willard G. Manning, Joseph P. Newhouse, and Robert H. Brook. “Impact of Cost Sharing on Emergency Department Use.” New England Journal of Medicine 313 (August 22, 1985), pp. 484-90. Parkin, David, Alistair McGuire, and Brian Yule. “Aggregate Health Care Expenditures and National Income: Is Health a Luxury Good?” Journal of Health Economics 6 (1987), pp. 109-27. Phelps, Charles E. “Diffusion of Information in Medical Care.” Journal of Economic Perspectives 6 (summer 1992), pp. 23-42. Phelps, Charles E., and Joseph P. Newhouse. “Coinsurance, the Price of Time, and the Demand for Medical Service.” Review of Economics and Statistics 56 (August 1974), pp. 334-42. Rosett, Richard N., and Lien-fu Huang. “The Effect of Health Insurance on the Demand for Medical Care.” Journal of Political Economy 81 (April 1973), pp. 281-305. Smith, Dean G., “The Effects of Copayments and Generic Substitution on the Use and Costs of Prescription Drugs.” Inquiry 30 (summer 1993), pp. 189-98. Van Vliet, Rene C. J. A. “Effects of Price and Deductibles on Medical Care Demand, Estimated from Survey Data.” Applied Economics 33 (October 10, 2001), pp. 1515-24. Whitney, Coralyn W., et al. “The Relationship between Price of Services, Quality of Care, and Patient Time Costs for General Dental Practice.” Health Services Research 31 (February 1997), pp. 773-90. Winslow, Ron. “In Health Care Low Cost Beats High Quality.” The Wall Street Journal, January 18, 1994, p. B1. Woolhandler, Steffie, and David U. Himmelstein. “Paying for National Health Insurance – and Not Getting It.” Health Affairs 21 (July/August 2002), pp. 88-98.