GSFch9 - RT 11 Supply and Demand 9. 10. 50 U.) A...

Info iconThis preview shows pages 1–9. Sign up to view the full content.

View Full Document Right Arrow Icon
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 2
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 4
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 6
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 8
Background image of page 9
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: RT 11 Supply and Demand 9. 10. 50 U.) A profit-maximizing firm finding that its demand is inelastic will necessarily find it profitable to increase its price; therefore, its equilibrium price elasticity will neces- sarily be greater than 1.0 in absolute value. Are the market— and firm-specific elas- ticity data reported here consistent with this theory? The consumer’s indifference curves in Figure 8.1 indicate substitutability between visits and other goods.What will the indifference curves look like if the consumer per- ceives no substitutability? What will happen to the elasticity of demand in this case? Suppose that Martha’s income is $20,000 per year. She can spend it on health care visits, which cost $40 per visit, or on groceries (standing for all other goods), which cost $50 per bag of groceries. Draw Martha’s budget constraint. Using indifference curves, show Martha’s optimum, if she buys 300 bags of groceries per year. Suppose that Martha’s income rises to $21,000 per year, and that she increases her consumption of health care visits by five Visits. Using the graphs for Exercise 1, draw the new equilibrium. What is her income elasticity of demand for health care visits? Consider the following information on Alfred’s demand for visits per year to his health clinic, if his health insurance doesnot cover (100 percent coinsurance) clinic visits. 7 P 5 10 15 20 25 30 35 40 hmoxuooxokoko {Q Alfred has been paying $30 per visit. How many visits does he make per year? Draw his demand curve. What happens to his demand curve if the insurance company institutes a 40 per- cent coinsurance feature (Alfred pays 40 percent of the price of each visit).What is his new equilibrium demand? Suppose that a consumer makes V0 physician visits each year at a price of P0. If the price elasticity is -0.4, what will happen to the number of visits, if the price rises by 10 percent? What will happen to total physician expenditures? Why? . If the price elasticity of demand is —0.5 and the income elasticity is +0.3, then what will be the effect of a simultaneous 10 percent increase in price, and 10 percent in- crease in income on health expenditures? Draw a diagram for hospital care that reflects the income-elasticity estimates found empirically. As income increases, what happens to the proportion of income spent on hospital care? . Would the opportunity cost of waiting time be higher for higher-income people or lower—income people? Given your answer, for which income group would money price tend to be a smaller portion of the full price? PA R T III: INFORMATION ll“I‘I‘!U‘lgV‘l‘lclul‘lnl‘l‘fl‘l‘l DIFI’IBIVIEIDIDI aaaemeawaeuwaemm C H‘ 1B1bw31§13131915 n—u—u—a--—-—-—-— Ilai_l1l‘lul‘lfl‘l‘l‘ ‘ aH‘I‘IJa l ulvlvlvlvlvlvvl E R unnameneawawaulmm «B1BVQVBVBVQVE43 .-.--—n-u—u—u——— Asymmetrc lntormoton and Agency Overview of Information Issues H Asymmetric Information . 3 Application of the Lemons Principle: Health Insurance The Agency Relationship E Consumer Information, Prices, and Quality 5 Conclusions that individuals are fully informed about prices, quantities, and the relationships of medical care and other inputs to their levels of health. We examined decision making within a model that assumed perfect information. Depending upon the purpose of the model, such an assumption may be justified even if it is not realistic. However, a more complete understanding of the health economy requires particular insight into the effects of various informational problems in health care markets.1 Though imperfect information has long been regarded as prevalent throughout the health economy, until recently, insight into its specific effects has been lacking.'Ihe eco- nomics of information has emerged as a distinct specialty, and needed tools to study in- formational problems have been developed. Health economics has been greatly enhanced by the study of asymmetric information and agency relationships. The former The traditional theory of demand, as we have seen, begins with the assumption lThe emergence of health economics as a distinct field is often traced to Kenneth J. Arrow’s (1963) seminal article, “Uncertainty and the Welfare Economics of Medical Care." Arrow emphasized the role of imperfect information and uncertainty, especially the features of health care markets due to the “imperfect mar- ketability of information.” 187 RT 1]] Infonnation encompasses situations where buyers and sellers have different levels of information; the latter concerns situations where, for lack of information, buyers or sellers rely on other parties to help make decisions. ew of information issues The markets for many health care services and for insurance in particular are marked by significant degrees of asymmetric information and agency relationships. For ex- ample, adverse selection, a phenomenon in which insurance attracts patients who are likely to use services at a higher—than—average rate, results from asymmetric informa- tion. It seems plausible, and most analysts agree, that a potential beneficiaries have bet— ter information than the insurer about their health status and expected demand for health care. As a result, premiums for higher-risk patients will be underpriced, encour- aging such patients to overinsure while the opposite holds for lower-risk patients. This phenomenon, which will be discussed further, reduces the efficiency of health insurance markets while redistributing income from the healthy to poorer risks. Information and agency problems account for many other important characteris— tics of health care markets. The possible preference for health care delivery by nonprofit hospitals and nursing homes (chapter 18) has been attributed to patients’ lack of infor- mation and inability to discern quality. For some patients, a nonprofit status might be taken as reassurance of higher quality because decisions are made independent of a profit motive. Lack of quality information also is thought to be an important motive for licensure and other regulatory measures (chapters 16 and 20). The present chapter has three goals. First, we introduce information asymmetry, de- scribe its relative prevalence, and determine its consequences, especially for insurance markets. It will quickly become clear, however, that adverse selection in insurance is only one consequence of asymmetric information. Asymmetric information, as when a patient is less well informed about appropriate treatments than the attending physician, typically leads to an agency relationship be— tween the patient and provider. The second goal of this chapter is to describe the agency relationship and examine some of the problems arising in health care markets from im- perfect agency. The special and controversial case of supplier—induced demand (SID) is revealed as an asymmetric information/agency problem in the following chapter. Our final goal is to examine the effects of imperfect consumer information on the price and quality of health care services. Despite informational disadvantages for con- sumers, they often influence markets. in predictable ways. Here and throughout the chapter, we identify arrangements that commonly evolve to reduce the disadvantages for the less well-informed parties. Thus, the ultimate consequences of asymmetric in- formation and imperfect agency on the efficient functioning of markets are often not as severe as one would assume initially. netric Information Basic microeconomic theory usually includes an assumption that the market being ana- lyzed exhibits perfect information. Under conditions of perfect information, all decision makers, meaning all consumers and producers, have complete information on all prices, CHAPTER 9 Asymmetric Information and Agency 189 as well as the quality of any good or service available in the market. Consumers also Will be as well informed about the product as the seller. Obviously, information is never perfect in the real world. However, theories in many disciplines are necessarily developed with concepts that are rarely or never actu- ally obtained in practice. Examples from other sciences include the idea of a perfect vacuum, frictionless motion, or the idea of a geometric point that has no dimensions. In economics, perfect information is often a useful starting point because the properties and predictions of the standard models relying on this assumption are so well under- stood. The researcher’s task is then to assess whether the actual functioning of a given market is close enough to the predicted behavior that the application of the model is useful, even though all of the assumptions of the model are not fully satisfied. On the other hand, cases do arise where imperfect information does seem to mat- ter. During the past three decades, economists have developed new insights into the ef- fects of imperfect and asymmetric information. This section examines some of that work, including contributions by health economists to the specific problems of the health sector. ON THE EXTENT OF INFORMATION PROBLEMS IN THE HEALTH SECTOR Before investigating several contributions to the economic theory of information, we begin by asking how prevalent information problems are in the health sector. It is ob- vious not only that imperfect information is prevalent in health care markets, but also that there is asymmetry of information. Levels of information will diflfer among partici— pants in many transactions, such as between physicians and patients. Often, the patient is poorly informed compared to the provider about his or her condition, the treatment available, expected outcomes, and prices charged by other providers. Furthermore, it is our presumption that information problems exist in the health sector that are promi- nent enough to require the special analysis of the economics of information. Although we can agree that information problems arise in health care markets, we must avoid the temptation to overgeneralize or to draw the point too far. To say that information problems exist in the health sector does not mean that these problems are necessarily worse than in any other market. For example, markets for insurance, other professional services, automobile and appliance repairs, and many other goods and services also exhibit asymmetries. We should not necessarily conclude that infor- mation asymmetries in health care markets make it impossible for corrective institu- tions, practices, or products to evolve; nor does it necessarily preclude the possibility of competition. _ . Pauly (1978) notes that half or more of physician visits customarily are made for services such as general checkups or chronic care for which the patient has some if not considerable experience. From data on the portion of medical expenditures attributable to ambulatory physician care, we can estimate that if half of this care is reasonably well informed, then about 8 percent of total medical care is informed. Reasoning in this man- ner about all sorts of medical care and products, Pauly concluded that plausibly “one- fourth or more of total personal health-care expenditures might be regarded as ‘reasonably informed” (p. 16). By adding nursing home services and chronic conditions, Pauly (1988a) subsequently argued that this ratio is about one third. 0 RT DI Information The second branch of this argument notes that the information gap in several med- ical care issues is shared by the provider. As we will emphasize when describing the “small area variations” literature in chapter 10, the physician often is uncertain if not uninformed about the outcomes of many medical procedures. In such cases, informa- tion asymmetry does not necessarily arise even though it may well be correct that the patient is ill informed. Finally, economic analysts of information asymmetry problems have been able to show that markets may perform well in the face of some degree of information asym- metry provided a sufficient portion of the consumers are reasonably well informed.2 For example, perhaps a majority of consumers who use personal computers are relatively poorly informed about their technical aspects and relative qualities and prices. How- ever, a significant minority of consumers tends to be highly informed. In markets like this, it is possible that the informed minority is sufficient to provide the economic disci- pline it takes to make the market perform well so that the rest of us will tend to find re- liably that the higher-priced computers also tend to be of higher quality. We conclude this section by summarizing its main point. Certainly information gaps and asymmetries exist in the health sector. They'are perhaps more serious for health care than for other goods that are important in household budgets. This makes it useful for the student of health economics to investigate the theory of information asymme- tries and its application to health care. However, it may be unwise to overgeneralize to the point of asserting that information gaps preclude the possibility of having a high de- gree of competition. In particular, mechanisms to deal with information gaps should not be overlooked. These mechanisms include Iicensure, certification, accreditation, threat of malpractice suits, the physician—patient relationship, ethical constraints, and having some informed consumers. Will a state of relative consumer ignorance preclude high levels of competition? Will health care markets be characterized by a high degree of price dispersion and the provision of unnecessary care or care that is not in the patient’s best interests? Can some of the characteristics of health care markets and the evolution of their institu- tional arrangements be related to asymmetric information? The following sections ad- dress these and other questions by beginning with the pioneering work on asymmetric information. ASYMMETRIC INFORMATION IN THE USED—CAR MARKET: THE LEMONS PRINCIPLE Akerlof (1970) is often credited with introducing the idea of asymmetric information through an analysis of the used-car market. Though seemingly unrelated to health care, his classic article is worth studying for two reasons. First, it tells us much about adverse selection and the potential unraveling of health insurance markets. Adverse selection is one of the keys in our understanding of some major contemporary issues such as the uninsured and the performance of health maintenance organizations or other delivery systems. Second, Akerlof’s example leads right into the issue of agency. In Akerlof’s example, used cars available for sale vary in quality all the way from cars that are still in mint condition to some that are complete lemons. Information asym- 2See Salop (1976) and Grossman and Stiglitz (1976). CHAPTER 9 Asymmetric Information and Agency 191 metry arises if, as is often plausible, the sellers know better the true quality of their cars than do the potential buyers. Akerlof devised an example that showed that cases may arise Where such information asymmetry causes the market for used cars to perform poorly or even to disappear entirely. To illustrate the nature of the used-car example, consider a somewhat simpler ex- ample, but one that retains the essential features. In particular, suppose nine used cars are to be sold (potentially) that vary in quality from 0, meaning a lemon, to a high of 2, meaning a mint—condition used car. In fact, suppose that the nine cars have respectively quality levels (Q) given by the cardinally measured index values of 0, i, a, i, 1, 1%, 1%, i, and 2. Under a cardinal index, a car with a value of 1 has twice the quality of a car with an index of %. The distribution of these cars is shown in Figure 9.1, where the hori- zontal axis shows the quality level and the vertical axis shows the uniform probability, %in this case, of randomly picking a car of each given quality. Suppose further that the owner of a car knows its quality level exactly but that the nonowners (potential buyers) know only the distribution of quality. It is known that the owners have a reserve value on their cars, so that Reserve Value to the seller = $1,000 X Q. That is, the owners would sell their car only if they could get at least $1,000 for every unit of quality that the car has. On the other hand, the nonowners are more eager for used cars and value them so that the Value to Nonowners = $1,500 X Q. To make this experiment a complete market, suppose that an auctioneer is hired to call out market prices; sales take place when the auctioneer finds a price that successfully equates quantity demanded with quantity supplied. ‘ Does a Market Exist? . Consider what would happen here under asymmetric information. Suppose the auc- tioneer calls out aninitial trial price of $2,000 per car.At this price, all owners know it is worth— while to sell their cars, so all nine cars will be supplied. However, nonowners, knowing only the distribution of quality but not the quality of each individual car, will make a best guess that a given car is of average quality; that, is Q = 1.They would not buy any cars at a price of $2,000, because they are willing to pay only $1,500 per unit of quality.They guess that all cars have a quality of 1, for a product of $1,500 X 1 = $1,500, which is less than the $2,000 asked. They would be willing to buy cars only if the price were less than or equal to $1,500. rig Probability Quality | | l l .3. 4 2 4 RT Ill Information So the auctioneer, perhaps trying to accommodate the potential buyers, tries a IOWer price, say $1,500. Unfortunately, at this price, the owners of the two best cars will with- draw from the market. Why? The owner of the car with 2 units of quality is receiving only $1,500/2, or $750 per unit of quality; the owner of the car with 1% units will act the same way. The withdrawal of the two best cars causes the average quality of the seven remaining cars to fall. With nine cars, the average of the distribution was Q = 1. Now at a price of $1,500 per car, the best car offered will have a quality level of %, and the aver- age quality will now be %. Potential buyers would now be willing to pay only $1,500 per unit quality X (% unit of quality per average car) or $1,125 for any car. Just as the previ- ous price of $2,000 per car was too high, the new price of $1,500 is too high for buyers. Will an equilibrium price ever be found? Surprisingly, in this example, an equilibrium, one that satisfies both buyers and sellers, will not be found. The reader can discern this by trying several successively lower prices. In the end, the cars will not be sold even though nonowners value the cars considerably more than their current owners. Akerlof saw the problem this way: When potential buyers know only the average quality of used cars, then market prices will tend to be lower than the true value of the top-quality cars. Owners of the top-quality cars will tend to withhold their cars from sale. In a sense, the good cars are driven out of the market by the lemons. Under what has become known as the Lemons Principle, the bad drives out the good until, in some cases such as this one, no market is left. Imperfect Versus Asymmetric Information To see that the problem is asymmetric rather than imperfect information for both buyers and sellers, consider what would have occurred if information had been sym- metric. Suppose, in particular, that both owners and nonowners were uncertain of the quality, that they knew only the average quality of used cars on the market. Again, let the auctioneer start with a trial price of $2,000 per car. All owners, at their best guess, may presume that their car is of average quality, and that the average will again be Q = 1. Thus, at a price of $2,000 per car, all nine cars would be offered for sale. How- ever, the nonowners would be willing to pay, at most, $1,500 based on their guess that a given car is of average quality (that is, $1,500 per unit of quality, multiplied by expected quality of Q = 1). Again, suppose the auctioneer tries to accommodate the potential buyers by offering a lower trial price, say $1,500. If the owners have imperfect informa- tion rather than better information, they will guess that their cars are of average qual- ity, and thus worth (to them) $1,000 per unit of quality, multiplied by average quality of 1. So the owners are willing to supply nine cars at a market price of $1,500, and the buy- ers are willing to purchase them at that price. The market thus exists, and clears (supply equals demand) if the information is symmetric—in this case equally bad on both sides. This example is extreme in several respects. The assumption of an auctioneer, the assumption that there is only one price for the used cars, the implicit assumption that the parties are not influenced by risk, and even the assumption that the quality of the cars is exogenously given could each be modified to add more realism. Since the lemons example was published, several analysts have worked on models that modify these as- sumptions. In some cases, this changes the result significantly. However, Akerlof’s main point remains illuminating. Phlips (1988) notes: Obviously, the example is extreme. It is meant to make the point that the pres- ence of unidentifiable lemons makes it difficult to sell cars of better quality at a good price. . . . The more general phenomenon of asymmetric information CHAPTER 9 Asymmetric Information and Agency 193 coupled with a reduction of average quality of the goods and services traded in the market has received the name “adverse selection” (p. 70). Application of the Lemons Principle: Health Insurance Adverse selection is a problem for markets involving health insurance, as well as for an- alysts studying the relative merits of alternative health care provider arrangements. The application of the Lemons Principle to the problem of health insurance can be seen di- rectly with the help of the previous example, which is a mirror image of the insurance problem. In Figure 9.2, let the horizontal axis measure the expected health expenditure levels of a population of n potentially insured people, instead of measuring the quality of used cars. Assume that they have the same demographic characteristics and that their expected health expenditure levels for the insm'ed period range from a low of $0 up to an expenditure level of $M. The vertical axis represents the probability with a uniform disuibution (so that the probability of any level of spending is 1 The insurer must at least break even, which means that the premium (or price) received from each insured must cover the insured population’s average expenditure and other expenses (includ- ing marketing and overhead). Information asymmetry is likely to occur here because the potential insureds know more about their expected health expenditures in the coming period than does the insurance company. To illustrate, assume a potential insured knows his or her future expenditure exactly but that the insurance company knows only the distribution of expenditures for all insured persons. Again, use the device of the auctioneer to illustrate the point. Suppose this time that the auctioneer attempts a first trial price of $0! All potential beneficiaries would cer- tainly demand coverage at this price. Just as certainly, the insurance company, expect— ing an average expenditure of $ (%)M, would require a premium of at least $ (%)M. ‘9 Following Akerlof’s analysis, suppose the auctioneer tries a higher price, say $ (%)M, hoping that this will clear the market. In this case, all potential beneficiaries who expect an expenditure level below $ (%)M will choose to self-insure, that is, leave the insurance market altogether, because this premium is higher than their privately known levels of health expenditure.When these healthier people leave the market, the average expected 3|-L Health expenditure ($) M—L rel—x .q ,\’ ‘7“ 44“ - 33“", 1T III Information expenditure level of'the remaining insured persons, those with expected expenditures from :3 (%)M to $M, rises to $ (3M .Thus, the higher health risks tend to drive out the lower health risk people, and a functioning market may even fail to appear at all for some otherwise—insurable health care risks. Observe again that it is the asymmetry of information rather than the problem of incomplete information that leads to this result. If patients were no better at pre- dicting their health expenditures in our example than the insurer, adverse selection would not take place. That is, all potential beneficiaries would have expected expendi- tures of $ (%)M and would be willing to purchase insurance at the premium of 58 (%)M. INEFFICIENCIES OF ADVERSE SELECTION This example illustrates the effects of adverse selection. Analysts of the health insur- ance industry recognize that even in its less extreme forms, adverse selection will ap- pear. Even if functioning health insurance markets do evolve in the presence of information asymmetry of this kind, the resulting adverse selection tends to result in economic inefficiencies. . What are the inefficiencies? Unlike the example, few people can know exactly their future level of expenditures. Risk is the main reason for insurance. However, if the lower risks are grouped with higher risks and all pay the same premium, the lower risks face an unfavorable rate and will tend to underinsure. They sustain a welfare loss by not be- ing able to purchase insurance at rates appropriate to their risk. Conversely, the higher risks will face a favorable premium and therefore overinsure; that is, they will insure against risks that they would not otherwise insure against. This, too, is inefficient.3 If information asymmetry threatens to lead to inefficiency and even to the elimina- tion of fimctioning markets in some cases, we would expect consumers and providers to resort to other economic devices and institutions to help overcome the problem. To illustrate, reconsider the used-car example. While the lemons problem in used-car mar- kets is real, many economic devices have evolved to help counter it. For example, a buyer may hire a mechanic to examine the car of interest, the seller may offer a warranty, and agencies or consumer unions may arise to provide quality information. Similar features have evolved in health care markets. In health insurance markets, beneficiaries often are not covered for preexisting conditions. Premiums for individual policies may be based on other information that insurers use to predict expenditures. They may consider fac- tors such as age, employment status, and occupation. EXPERIENCE RATING AND ADVERSE SELECTION Group insurance can be a more useful mechanism to reduce adverse selection. Most em— ployees and their families in the United States are insured through employer group plans rather than through individual policies. Group plans enable insurers to implement experience rating, a practice where premiums are based on the past experience of the group, or other risk-rating systems to project expenditures. Because employees usually 3Adverse selection is not merely a theoretical prediction. It appears in the markets for supplemental Medicare insurance (Wolfe and Goddeeris, 1991) and individual (nongroup) insurance (Browne and Doerpinghaus, 1993). In addition to inefficiency, income is redistributed from consumers who are lower risks to those who are higher risks. CHAPTER 9 Asymmetric Infornmtion and Agency 195 have limited choices both within and among plans, they cannot fully capitalize on their information advantage. Van de Ven and Van Vliet (1995) show that adverse selection can be reduced substantially if insurers use just a few risk factors to determine premiums. Although experience rating can reduce adverse selection, it has come under in- reasing attack with the rapid growth of managed care plans such as health maintenance organizations (HIVIOs). HMOS, described more fully in chapter 12, are organizations that receive a predetermined premium to provide the contracted health care for its en- rollees. In contrast to traditional insurance Where providers are independent of insur- ers, HIVIOs integrate insurance with the provision of health care. I-llVlOs and similar entities have been promoted heavily on the belief that they have a powerful self-interest in eliminating the inappropriate care that might be recom- mended by agents in traditional fee-for—service systems. However, the intense competi- tion to enroll healthy populations in managed care plans, known as “cream skimming” or “cherry picking,” has led to concerns that insurers are more interested in finding fa- vorable groups than providing quality care. Experience rating also redistributes income toward healthy populations, and the perceived inequity of such redistributions has be- come a public issue. ‘ To deal with these concerns, several states have introduced some degree of man- dated community rating, a practice in which an insurer charges all groups within an area the same premium. In a less rigid form, upper and lower limits on premiums are estab- lished through rate bands. The effects of these changes have not yet been determined, but Goldman et a1. (1997) predict some serious redistributional consequences of the ,1. ‘ community rating schemes considered for California. In particular, because health care spending in wealthy, urban areas is relatively high, their model predicts large regional transfers of income from poorer, rural communities to those areas. The Agency Relationship An agency relationship is formed whenever a principal (for example, a patient) dele— gates decision—making authority to another party, the agent. In the physician—patient re- lationship, the patient (principal) delegates authority to the physician (agent), who in many cases also will be the provider of the recommended services. The motive behind this delegation of authority is that the principals recognize that they are relatively un- \ informed about the most appropriate decisions to be made and that the deficiency is on? best resolved by having an informed agent. Thus, asymmetric information and agency are closely related phenomena. AGENCY AND HEALTH CARE What would the perfect agent do? Following Culyer (1989), the perfect agent physician chooses as the patients themselves would choose if only the patients possessed the in- formation that the physician does.4 This is in line With the medical code of ethics to the extent that the patient’s own interest focuses on his or her health. When any conflict arises, the perfect agent focuses on the patient’s preferences, not his or her own. 4See also Clark and Olsen (1994) for a model and analysis that makes use of the assumption of perfect agency. RT III Information The problem for the principal is to determine and ensure that the agent is acting in the principal’s best interests. Unfortunately, a divergence of interests may arise, and it may be difficult to introduce arrangements or contracts that eliminate conflicts of interest. As an example, Dranove and White (1987) ask why physicians are not reimbursed on the basis of improvements in patient health. More simply, why are they not reim- bursed only if the patient is cured? It would appear that such a contract would be a natu- rally arising device that would merge the interests of both principal and patient. The authors suggest that the unavailability of such contracts in health care lies in the prob- lem of asymmetric information, although in this case it is the physician who may lack information about the patient’s well-being. To illustrate, take patients with low back pain. Regardless of their improvement, the patients have a financial incentive to understate the extent of their improvement. The provider also has an incentive to overstate the difficulty in treating the patients and in improving their health in order to increase the payment (which, let us assume, is based on the difficulty of the case). Further, it is these information problems and not other special characteristics of health care delivery that preclude payment based on the de- gree of improvement of the patients’ conditions. Dranove and White further apply agency theory to explain other features in the or- ganization of health care delivery. A feature we commonly find is that patients often es- tablish a long-term relationship with a physician afli pay that physician on a fee-for—service basis. As discussed earlier, such an arrangement would appear prone to lead to conflict between patient and provider. It is thus natural to ask the following questions: Why does this particular physician—patient arrangement arise and why is it so common? Dranove and White explain this phenomenon with an argument based on infor- mation theory. They argue that a continuous relationship between patient and physi- cian provides the patient with increasing information with which to monitor the physician. This information places constraints on the extent to which the provider is able to deviate from an agency responsibility. Monitoring also encourages the physician to make appropriate referrals to other providers when he or she is unable to provide the services alone. We also can add to their argument by further pointing out that a continuous rela- tionship reduces the cost of transferring information about medical history, circum- stances, and preferences from patient to provider. These advantages of the usual physician—p atient relationship would be eroded if patients and providers were to switch to limited-period contracts under which providers are reimbursed on a different basis. CHAPTER 9 Asymmetric Information and Agency Contingency Fees and Medical Malpractice Guarantees of satisfaction are common in the market for many consumers’ goods. Most‘ large retailers will accept returns without question if the customer is not satisfied. How- ever, guarantees also have found their way into some services, such as for car time<ups and even hair styling. One of the most inter- esting and controversial examples of a service guarantee is the contingency fee system in tort cases, particularly for malpractice suits. Under the contingency fee system, the plain- tiff (party that sues) pays only if he or she wins damages (at a rate that is usually at the legal maximum, about one-third of the award in many states). Contingency fees are possible because, unlike with many medical services, the outcome is clear. Opponents of this system charge that it leads to unnecessary litigation and frivolous suits. Supporters, by contrast, claim that it pro- vides access to the courts to those with more ' limited financial means.They also suggest that the client and attorney share the common goal of maximizing the award. Although at first glance this appears to be an obvious proposi- tion, there remain significant problems of agency and asymmetric information under the contingency fee system. An inherent conflict arises because the ra- tional client would want an attorney to spend more and more time on the case, so long as each additional hour would yield some posi— tive benefit. In Figure 9.3, the marginal bene- fit curve, (33-)MB, represents the benefit of each additional hour (in terms of additional mone- tary gain) to a client who keeps % of the court- awarded damages or settlement and pays an attorney the remaining a as a contingency fee. Thus, the fully informed client wants an attor— ney to spend a total time of 0B hours on the case, the point at which (§)l\/[B becomes zero. The attorney, on the other hand, wants to maximize his or her own income net of costs. The attorney’s marginal benefit curve is given as (%)MB and the marginal cost is represented by MC. The attorney maximizes net income at 0A hours where (%)MB equals MC. He or she has a different utility flmction and will recognize that there is an opportunity cost to the time spent. Thus, the attorney will pre- fer to spend less time on the case than would be desired by a fully informed client. The extent to which attorneys deviate from client responsibilities is not known. Nevertheless, some critics of contingency fees have recog- nized this conflict and have thus charged that the system promotes early settlement of cases that would be further pursued if at- torneys acted solely in the best interests of their clients. 197 CONSUMER INFORMATION AND PRICES One of the most novel approaches to issues of consumer information and competition was introduced by Satterthwaite (1979) and Pauly and Satterthwaite (1981). The au- thors identify primary medical care as a reputation good—«that is, a good for which con- sumers rely on the information provided by friends, neighbors, and others to select from the various services available in the market. Each firm offering physician services pro- vides a service that is differentiated from others, that is, the firms do not offer identical services. As such, the market can be characterized as monopolistically competitive. mer Information, Prices, and Quality A further objective of this chapter is to examine the effects of imperfect information on the price and quality of medical services. Would relatively poor consumer information reduce the competitiveness of markets? Does increasing physician availability increase competition and lower prices as traditional economics suggests? What happens to qual— ity? How do consumers obtain and use information? Although research on these issues is still limited, several studies provide helpful insight. RT III Information MB, MC Total MB MB = Marginal benefit MC = Marginal costs Dollars MC Reputation Goods Under these conditions, the authors show that an increase in the number of providers can increase prices. The reasoning behind this surprising prediction is logical. Recall that a typical consumer relies on other consumers for information regarding their experiences with physicians. Thus, when physicians become numerous, the average number of friends who see any provider diminishes; this, in turn, diminishes the average level of information available. The consumer’s responsiveness to prices and other practice characteristics de- pends on his or her knowledge of—that is, information about—the available alternatives. Thus, this reduced information reduces the price responsiveness (i.e., the elasticity) of the firm demand curves, causing the equilibrium prices to rise. The economic idea is that re duced information tends to give each firm some additional monopoly power. That reduced information enhances monopoly power and reduces the elasticity of the firm demand curves is consistent with standard theory. That such a situation may arise as the number of sellers increases is an unconventional idea. It is understandable if the results run counter to one’s intuition. The authors have, however, provided em- pirical support for their theory, and the interested reader is referred to their work for further study of the issue. The Role of Informed Buyers The degree to which imperfect price information contributes to monopoly power should not be overemphasized. Recall that it is not necessary for every buyer of a com- modity to have perfect price information to elicit relatively competitive pricing condi- tions. Realistically, most consumers lack complete price information about many of the goods and services they buy (that is, they don’t know what alternative sellers are charg- ing). Yet, despite variations in the prices of individual items among, for example, gro- cery stores, the average charges for a set of items across similar types of stores arelikely to be similar and close to competitive pricing. A growing body of literature shows that CHAPTER 9 Asymmetric Information and Agency 199 it is sufficient to have enough buyers who are sensitive to price differentials to exert dis— cipline over the marketplace. This is likely to hold especially where the damaging threat exists of having any systematic differentials publicized by consumer organizations or the low-priced merchants themselves. The thrust of these arguments is to suggest that while imperfect price information is likely to produce higher prices, this phenomenon may be substantially limited. In health care markets, where many services are fully or partially covered by insurance, there are added considerations. While a patient may become less sensitive to price lev- els and price differentials in the choice of providers, third-party payers, such as insurers, have assumed a monitoring function.Through selective contracting and other fee agree- ments, the actual reimbursement is often lower than the provider’s charges. Price Dispersion The distinction between the effective transaction price and a provider’s charge also obscures evidence of dispersion of fees as distinct from the average level of fees. Under conditions of imperfect consumer information, Nobel Laureate George Stigler (1961) argued that variation in prices will increase. . Building on Stigler’s insight, Gaynor and Polachek (1994) have developed mea- sures of the degree of both buyer and provider ignorance by using frontier regression methods. These authors separated price dispersion into measures of incomplete buyer information, incomplete seller information, and random noise. They found that both pa— tients and physicians exhibited-incomplete information with the measure of ignorance being one and a half times larger for patients than for physicians. In addition, Juba (1979) found that the variations in physician fees appear to be larger than those found in other relatively competitive markets. It is not clear, though, that variations in effective transactions pricing resulting from insurance agreements lie outside of competitive norms. CONSUMER INFORMATION AND QUALITY Because quality is not easily monitored, the search for information regarding it can be costly to the consumer. At the same time, the consequences of poor quality care can be severe. Thus, as previous discussion suggested, despite asymmetric information, patients rely on a variety of countervailing arrangements that are intended to reduce their search costs. These include licensure and certification, the threat of malpractice suits, codes of ethics, and quality assurance schemes that are either mandated or volun- tary. The public’s demand for regulation and Iicensure, the role of accreditation such as that offered by the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), and other arrangements are taken up in later chapters. At this point, we pursue the consumer’s direct role through the Dranove and White argument that the physician—patient relationship enables patients to monitor providers and to encourage physicians to make appropriate referrals. To the extent that many specialists rely on referred patients, it would seem that these specialists have incen- tives to maintain quality. Are they also rewarded with higher prices when they provide higher-quality services? Theory suggests that if consumers have the ability to distinguish between quality, then the demand for higher-quality providers and thus price should be greater than for lower-quality providers. Haas-Wilson (1990) examined this propo- sition using data from the psychotherapy services market. She investigated whether XRT III Information the prevalence of referrals from informed sources affects the price of social workers’ psychotherapy services. Informed sources include other health providers and other professionals such as school counselors and clergy. Regression analysis of a sample of social workers’ fees indicated that fees are pos. itively and significantly affected by the percentage of clients who were obtained through informed referrals. The evidence thus shows that patients rely on informed sources (agents) for information and that higher quality, as measured by informed referrals, is rewarded by higher fees. Although more research is needed on other services, this work indicates that asymmetric information does not necessarily lead to market failure. OTHER QUALITY INDICATORS Luft and colleagues (1990) examined the influence of quality indicators and other vari- ables on the choice of hospitals for patients with seven surgical and five medical condi- tions. Various quality indicators—such as teaching status, transfer and referral patterns, medical school affiliation, and various indexes of outcomes——were examined. The au- thors found that medical school affiliation increased the probability of selection for seven of the twelve conditions studied and that poorer—than—expected outcomes index measures were associated with a lower probability of selection. Studies such as the previous one raise interesting questions about quality indicators and patients’ perceptions of quality. The medical literature has been strongly influenced by the structure, process, and outcome distinctions established by Donabedian (1966) to ' represent quality. Owing to its relative ease of measurement, the focus is often on struc- tural indicators, which refer to the availability of resources such as personnel per bed. However, an extensive marketing literature focuses on expressive quality as mani- fested through caring and communication skills. The marketing discipline also has stressed the credence properties of health care services.5 These are dimensions of qual- ity that cannot be evaluated, even after experience, and that are thought to be especially relevant to high-skill services. Considerable interest centers on the extent to which price, advertising, and other sources of information serve as signals for credence properties. Finally, despite licensure requirements and other schemes to promote quality, the media has disseminated claims by those within the profession that approximately 10 per- cent of physicians are incompetent or seriously impaired by drug and alcohol addiction. Consider also reports that some medical doctors are not actually medical doctors but have falsified credentials, work under the assumed names of deceased physicians, or prac- tice without licenses. Though the extent of these problems of incompetence, impairment, and fraud are not known with any accuracy, their acknowledged existence raises trou— bling questions about patients’ perceptions of quality and their ability to monitor qual- ity. The role of licensing boards and other organizations is similarly called into question. lusions There is little doubt that information gaps, asymmetric information, and agency prob- lems are prevalent in provider—patient transactions. However, for some health care ser- SSee Bopp (1990) and Lynch and Schuler (1990). Health economists also are beginning to examine the at— tributes of the physician—patient relationship. Vick and Scott (1998) found that the most important attribute is “being able to talk to the doctor." CHAPTER 9 Asymmetric Information andAgency Health Plan Report Cards The National Committee for Quality Assur— ance (NCQA), a private accreditation body for I-IMOs, issues report cards based on about 50 standardized measures of a plan’s perfor- mance (such as childhood immunization rates, breast cancer screening, and asthma inpatient admission rates). Newsweek, US. News & World Report, and various consumer groups also regularly rate HMOs. A key assumption behind these efforts is that informatiOn about quality will, like price information, help discipline providers through patient choices. Low-quality HMOs will not survive or at least they will not be able to change high—quality prices. Initial evidence on the intended effects of plan performance ratings questions the report card strategy.Through a consumer survey,Tum— linson et al. (1997) found that independent plan ratings are relatively unimportant/to their choices. Only 17 percent of respondents indi- cated that such ratings are essential, compared with 72 percent for specific plan benefits and 62 percent for out-of—pocket costs. Chernew and Scanlon (1998), employing multivariate sta— tistical methods on consumer choice of plans, confirm that ‘l‘employees do not appear to re- spond strongly to plan performance measures, even when the labeling and dissemination were intended to facilitate their use” (p. 19). 201 r vices, the problems are not necessarily great or larger than those for other goods. Patients are likely to be relatively poorly informed about treatment for conditions that they have not previously experienced and about care involving newer technologies. The informational asymmetries and reliance on provider—agents are likely to be most pro- nounced in these situations. Although there is a potential lack of competition, even wide information gaps do not necessarily lead to market failure. Leaving aside the role of licensure and regula- tion, arrangements have evolved to enable patients or their insurers to monitor the quality and prices of providers. Furthermore, higher—quality producers Ere rewarded by greater demand and higher prices. Evidence suggests that the variation in physician fees is greater than what is found in more competitive markets. Price dispersion data indi- cate some ignorance on both parts, physician and patient, with the patient being the lesser informed. Nevertheless, because of the use of referrals, accreditation, and other arrangements, the provider’s ability to raise prices above those charged by others and to sell low-quality services at high-quality prices is significantly constrained. Summary 1. Health care markets tend to be characterized by both imperfect information and asymmetric information.Asymmetric information describes a situation in which those on one side of a transaction have better information than those on the other side. I 2. Often, providers are relatively well-informed (e.g., about the patient’s illness and possible treatments). In other cases, buyers are relatively well-informed (e.g., the purchaser of insurance knows more about his or her health status and pertinent habits than the insurer does). RT 111 10. E Infommtion . The extent of consumer information problems should not be exaggerated. Con- sumers are reasonably well-informed on about one-fourth to one-third of their health care spending. One possible consequence of asymmetric information is that a market will not ex- ist. Even if it exists, a general reduction may occur in the quality of goods available (the “Lemons Principle”). . The Lemons Principle appears as the problem of adverse selection in health insur- ance and other health care markets. Adverse selection results from asymmetric information, not equally imperfect in- formation. Adverse selection in insurance results in inefficiencies through higher- risk consumers overinsuring, relative to the amounts they would purchase at actuarially fair rates, and lower risks correspondingly underinsuring. An agency relationship tends to be formed when a party (principal) delegates decision making to another party (agent). The problem for the principal is to de- velop a contract or relationship to ensure that the agent is acting in the principal’s best interests. . Various agency relationships have evolved to mitigate the problems associated with asymmetric information between patient and provider. These include the continu- ous physician—patient relationship and the health maintenance organization. Other constraints, such as licensure and accreditation, codes of ethics, and the threat of litigation, limit the ability of providers to deviate from their agency responsibilities. Many health care services are reputation goods. In markets for reputation goods, an increase in the number of providers can lead to an increase in monopoly power and higher prices. The existence of informed buyers helps exert discipline over the market by limit- ing price increases and price differentials among sellers. Higher quality tends to be rewarded with higher prices. Patients also respond to quality indicators in selecting a hospital. However, they rely relatively little on ob- jective plan ratings in their selection of HMOs. :ion Questions 1. U.) The market for higher education is another example a high degree of information asymmetry is likely. What mechanisms have evolved to help students in their choice of schools and classes within schools? Do you have confidence that higher-priced institutions provide higher—quality education? The situation in which an individual is interviewing for a job also exhibits infor- mation asymmetry. Explain why. How does the relatively poorly informed party deal with this? . The use of professional and independent buyer—agents to help individuals purchase automobiles or houses is becoming a more common phenomenon. Given the con- flict of interest facing the physician-agent, why do we not see greater use of a buyer- agent who is retained by the patient? The used-car market has publications that provide information on the quality and prices of used cars. Are similar avenues of information available to health con- sumers? What kind of information do they provide? Is it more or less effective than Exercises [0 U.) CHAPTER 9 Asymmetric Information andAgency 203 the information available on used cars? How would you, as a patient, find informa— tion about a provider’s quality or prices? How would you assess the confidence you have in that information? . What is a reputation good? What are examples of reputation goods outside the health care sector? Show what Pauly and Satterthwaite predict will happen to the demand curve for health services as a result of an increase in the number of providers. . Stigler argued that the variation in fees increases as buyer information decreases. Suppose you observe that each seller in a market is charging the identical price. What potentially conflicting inferences can you draw? . Why don’t physicians guarantee their work as do many auto repair shops? . Suppose that in the Akerlof exampl\e, there are only eight cars ranging in quality from .1} to 2 (i.e., there is no complete lemon). Hence, the mean quality level is 1.125. Determine Whether the market disappears completely and, if not, how many cars will be sold. ' . Using Figure 9.3, explain how the two MB curves shift if the attorney receives % of the expected gain. What will happen to the number of hours the attorney would pre- fer to spend on the case? What will the client prefer? . Give three examples of asymmetric information in which the health consumer has information that is unavailable to the health provider. Give three concrete exam- ples in which the health provider has information that is unavailable to the health consumer. In the Akerlof example, the individuals are treated as indifferent to risk. What would you expect to see in these markets if individuals wanted to avoid risk? What if there were some “risk lovers?” ...
View Full Document

Page1 / 9

GSFch9 - RT 11 Supply and Demand 9. 10. 50 U.) A...

This preview shows document pages 1 - 9. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online