Health Affairs Reimbursement To Private Payers ?

Leaders of health insurance companies, hospital systems, and physician organizations believe that when Medicare and Medicaid reduce payment rates to hospitals or physicians, these providers respond by raising prices to private insurers to offset a portion of the loss in revenue. This would mean that payment reductions in public programs contribute to increasing premiums for private insurance. But on both theoretical and empirical grounds, economists have been skeptical about the existence of this “cost shifting.” I show that more realistic models of the behavior of hospitals and physicians than exist in basic economics texts provide a conceptual basis for cost shifting. R e c ent d e v e l o pm ents in health care financing have rekindled the interest of policymakers in provider cost shifting. The Balanced Budget Act (BBA) of 1997 led to a substantial decline in Medicare hospital payment rates and a delayed (to 2002) decline in physician payment rates. Sharp drops in states’ revenues raise concerns about the prospect for sizable reductions in Medicaid payment rates to hospitals, physicians, and other providers. At the same time, private insurers are complaining loudly about the higher payments demanded by hospitals and some physician organizations. Policymakers need to incorporate estimates of the magnitude of cost shifting into making decisions on Medicare and Medicaid payment rates. Most executives in hospitals, physician organizations, health plans, and businesses have long been convinced that reductions in rates paid to Medicare and Medicaid lead directly to higher payment rates charged to private payers. But most economists who have published on the topic express strong skepticism about the possibility that cost shifting can and does occur.1 Not only do they point to empirical analyses that fail to obtain results supporting the existence of cost shifting, they also argue that cost shifting is conceptually impossible. The crux of their argument is the question of why providers with the ability to increase revenue through increases in prices to private payers would not have already exhausted such capacity prior to reductions in payment rates from public programs. Indeed, these economists maintain that even if objectives other than profits are maximized, such as providing uncompensated care, cost shifting is still not possible. H e a l t h T r a c k i n g W 3 4 7 2 8 O c t o b e r 2 0 0 3 DOI 10.1377/hlthaff.W3.472 ©2003 Project HOPE–The People-to-People Health Foundation, Inc. Paul Ginsburg ([email protected]), an economist, is president of the Center for Studying Health System Change in Washington, D.C. on A uust 4, 2017 by H W T am H ealth A fairs by http://conealthaffairs.org/ D ow nladed fom This paper attempts to reconcile this striking divergence between the perspectives of health care executives and those of many economists. After a discussion of what exactly is meant by the term cost shifting, the paper turns to explaining how cost shifting can occur. It shows that when one leaves the Economics 101 textbook and confronts the challenges of managing complex health care organizations, especially large nonprofit ones, maximization of profits or other objectives is not that useful a working assumption. The paper concludes with analysis of the conditions under which cost shifting is likely to be more pronounced, pointing out that present conditions are much more favorable to cost shifting than was the case eight to ten years ago. Defining ‘Cost Shifting’ There is no “correct” answer to the question of what cost shifting is, but some answers are more useful than others. Using the criterion of usefulness to policymakers, my preferred definition describes cost shifting as the phenomenon in which changes in administered prices of one payer lead to compensating changes in prices charged to other payers. An example would be if hospitals raised prices to private payers in response to Medicare payment rate reductions. This definition is consistent with the one used by Michael Morrisey, who has probably written the most on this subject.2 Another author has referred to such a definition as “dynamic cost shifting.”3 In this definition, the word cost is somewhat of a misnomer, since the focus is really on prices. Indeed, much of the discussion of cost shifting is really focused on the difference between prices charged private payers and the costs of treating the patients they cover. “Margin shifting” might be more descriptive of the phenomenon that concerns policymakers. But changes in administered prices also could affect the costs of providing services to privately insured patients. Hospitals and physicians usually find it impractical and otherwise undesirable to provide care in a very different manner for patients covered by different payers. Thus, if hospitals work harder to reduce costs in response to Medicare payment rate reductions, such as by increasing productivity or providing smaller wage increases, private payers may indeed gain some benefit from lower Medicare payment rates. Of course, such a change would be in the opposite direction to cost shifting that is based on changes in margins between price and costs. Such responses tend to reduce the magnitude of the total change in price to private payers, although the shifting of margins could be even greater. These impacts on the costs of providing care (as opposed to the costs faced by the payer) are often labeled “spillover effects” rather than “cost shifting.” This definition concerning the impact of a change in administered prices is narrower than one used by some others—differences in prices charged different payers that do not reflect cost differences. This alternative has a number of problems. First, the phenomenon of different prices for services sold to different buyers is common throughout the economy. Economists call it “price discrimination.” It is based on the notion that if a business perceives variation in its customers’ (or suppliers’) willingness to pay more (accept less) for a service, it will attempt to set different prices for different categories of customers or suppliers. The practice is limited by the ability to keep those who are targeted for the higher prices from shifting into the lower-price market. The airline industry has particularly elaborate mechanisms to charge people different prices for the same service. Identifying a phenomenon in hospital or physician services that is in fact the norm throughout the service sector of the economy would not have much significance for health policy. M a r k e t W a t c h H E A LT H A F F A I R S ~ W e b E x c l u s i v e W 3 4 7 3 “Identifying a phenomenon in hospital or physician services that is the norm throughout the economy would not have much significance for health policy.” on A uust 4, 2017 by H W T am H ealth A fairs by http://conealthaffairs.org/ D ow nladed fom Second, the broader definition does not align with policymakers’ specific options, such as whether or not to change payment rates. Medicaid rates are always lower than Medicare rates or negotiated rates for private insurance, but ending this practice has never been a realistic policy option. However, tracking changes over time in ratios of public program prices to prices paid by private insurance could be a useful test of the importance of cost shifting.4 The definition of cost shifting is relevant to policymakers setting administered prices because they need to know the full impact of their decisions, not just the budgetary impact. Consider the implications of a reduction in payment rates under two polar-opposite assumptions: (1) no induced change in private payment rates, and (2) increases in private payment rates that completely offset the reduction in administered prices. In the absence of cost shifting, a cut in administered prices will reduce profits or incomes to those who own hospitals or medical practices, limit providers’ ability or willingness to provide uncompensated care, and, over time, reduce providers’ capacity to provide services. The reduction in payment rates also will spur efforts to reduce costs and change the nature of services provided. Under the opposite assumption, complete cost shifting, a reduction in administrative prices will result in higher private health insurance premiums. This, in turn, will be reflected in higher product prices, lower wages, and lower profits in the general economy. Some view such a situation as similar to an unfunded mandate—government’s failure to fund the full costs of public programs and its reliance on private insurers to make up the difference. A problem specific to health care is that higher private insurance premiums will lead to higher numbers of uninsured people.5 The impact of a reduction of administered prices differs so much under these two assumptions that the question of the magnitude of cost shifting holds great importance for policy.6 Market Power And Cost Shifting The ability to shift costs is dependent on providers’ having “market power.” Although this term is often used imprecisely, economists would agree that market power for a seller is the ability to raise prices without losing all of its sales. A seller has a lot of market power if raising prices drives away few customers. So, when hospitals and physicians have a lot of market power, there is more potential for cost shifting because there is more discretion over price. But a provider with a great deal of market power will not necessarily shift costs. It all depends on whether a decrease in administered prices finds providers in a situation in which some ability to raise prices further was not used when administered prices were higher. Hospitals and physicians probably have at least some market power; the critical question is how much. In the era before managed care, both hospitals and physicians set charges for private patients. In the case of physicians, insurers screened these charges for “reasonableness.” The most important screen compared the charges with those of other physicians in the community. For charges above the screen, patients were responsible for the difference. So physicians had a great deal of market power to raise fees up to the screens, but less to raise fees further. To set prices higher than the screens, physicians tested patients’ willingness to pay more than was necessary to see other physicians. For hospitals, payment of full charges was closer to the norm, although Blue Cross plans often had contracts with hospitals that called for reimbursement of costs; some commercial insurers used an indemnity schedule or required patients to pay a percentage of the bill. Hospitals likely had substantial market power. Under managed care, hospitals and physicians must agree to a rate schedule with a health plan to be included in a network. Large hospitals or physician organizations tend to negotiate these rates with insurers. Smaller entities are offered rates and decide whether or not to accept them. But even in the latter case, the providers have market power. They can agree to some plans’ rate schedules but reject others. Rejecting one plan’s contract can still leave the provider with patients from other H e a l t h T r a c k i n g W 3 4 7 4 8 O c t o b e r 2 0 0 3 on A uust 4, 2017 by H W T am H ealth A fairs by http://conealthaffairs.org/ D ow nladed fom health plans and patients who are willing to see the physician on an out-of-network basis. Clearly, the degree of providers’ market power has varied over time. In general, they have less market power under managed care than before, in the sense that raising prices would likely lead to the loss of more patients. Indeed, the bottom line of Morrisey’s 1995 review is that while there may have been cost shifting in the past, it could no longer happen, because providers had so much less market power as a result of managed care.7 But some providers, such as hospitals and some large single-specialty physician group practices, appear to have regained some of the market power that they had lost by the mid-1990s, so even if Morrisey was correct in 1995, it may not be the case today.8 Hospital Cost Shifting Whether there can be cost shifting by hospitals and physicians depends on whether these providers maximize profits (or some other objective). If prices are often set below the profit-maximizing level, then there is room to raise them in response to a cut in administered prices. The stories for hospitals and physicians are quite different, so they are treated separately. Nonprofit status. Most hospitals are organized as charitable, private, nonprofit entities. Since they have no shareholders, profits are useful only to provide capital for plant and equipment, working capital, and reserves to get the institution through difficult periods. Profits certainly are valued, but they are not the objective of the organization. The theoretical literature on nonprofit hospitals offers the possibility that nonprofit hospitals will price their services lower than those of for-profit hospitals.9 Indeed, the courts often have permitted mergers among nonprofit hospitals that have greatly increased market concentration, under the belief that monopoly power will not be exploited and that the cost savings from scale economies will be passed on to consumers.10 Mission. Nonprofit organizations all have stated missions, which are intended to guide them in their decisions. But hospital missions tend to be vague. Hospitals often seek to “meet the needs of the community.” This can mean many things, including providing care to the poor, offering the latest medical technology, or avoiding delays in physicians’ obtaining care for patients. The latter might include adequate emergency department capacity. These different objectives can be at odds, in that pursuing one might be at the expense of others. Different members of hospital boards are likely to have different priorities, so that the direction of the organization could reflect a political compromise. The nature of such compromises means that it is difficult for hospital management to maximize one or more objectives. Governance. Looking at the governance of a nonprofit hospital reinforces the notion that behavior is likely to be different from profit maximization. Hospital boards are often made up of doctors who practice there and a variety of community leaders. These might include political leaders, religious leaders, business leaders (both large employers and businesses that view the hospital as a customer), and people with the ability and willingness to raise money for the organization. Such boards are likely to guide hospitals in directions other than profit maximization. Many leaders are likely to perceive that higher health insurance premiums would have negative effects on the community, and they would not be in favor of setting prices at the profitmaximizing level. Of course, the potential of prices’ being less than the profit-maximizing level depends upon the hospital’s having sufficient market power. Hospitals with limited market power might need to set prices at a profit-maximizing level even if their objective is to meet the needs of the community. M a r k e t W a t c h H E A LT H A F F A I R S ~ W e b E x c l u s i v e W 3 4 7 5 “If prices are often set below the profit-maximizing level, then there is room to raise them in response to a cut in administered prices.” on A uust 4, 2017 by H W T am H ealth A fairs by http://conealthaffairs.org/ D ow nladed fom An additional perspective on the likelihood that nonprofit hospitals might not maximize profits comes from the general literature on corporate governance. In 1932 Adolf Berle and Gardner Means pointed out that the separation of ownership and management often leads to corporations’ not operating efficiently or maximizing profits.11 Corporations with the most concentrated ownership can limit these problems. Hostile takeovers of the 1980s were perceived as a market response to the problem of separation of management from ownership. If management failed to realize the company’s potential, an outside group would purchase the organization and replace the management. Fear of hostile takeovers likely began to motivate management to seek opportunities to realize the organization’s potential. The 1990s saw increased use of an additional way to motivate management: stock options. Motivations. Compare the threat of hostile takeovers and stock options to the situation faced by managements of nonprofit hospitals. With vague missions and boards comprising community leaders, nonprofit hospitals are likely to have serious problems of separation between their boards and their management. Furthermore, unlike a corporate board, where all agree on the need for long-term profit maximization, hospital boards likely have major differences of opinion about what are the most important long-term objectives for hospitals. Edmund Becker and Sharyn Potter discuss how nonprofit hospitals with social ties have difficulties in managing diverse stakeholders and achieve lower degrees of organizational efficiency.12 With this limitation in governance, management could find setting prices below the profit-maximizing level to be more comfortable. For hospitals, raising prices will have uncertain effects. If all insurers agree to the higher rates demanded by a hospital, then revenues will rise proportionally. But if some do not, then the hospital might face a sizable loss of volume. These uncertainties are more formidable when fewer important insurers operate in the market. Hospital managers are more likely to take these risks if the organization’s viability is at risk than if surpluses are substantial. In the course of Community Tracking Study (CTS) site visits, I recall an interview in which the chief executive officer (CEO) of a prominent hospital indicated that the hospital was full but losing money and that raising prices to private payers was the only option to maintain financial viability. The implication is that these risks might not have been taken if Medicare and Medicaid rates had been higher. For-profit status. Not all U.S. hospitals are nonprofits. Investor-owned hospitals account for 15 percent of beds in community hospitals.13 Most are part of corporate entities that own many hospitals as their primary business. Should we expect investor-owned hospitals to shift costs? Two factors could lead to cost shifting. First, investor-owned hospitals are competing with nonprofit hospitals. If the latter keep prices low, then the former will have to adjust to this. Second, investor-owned hospitals have traditionally done things that one would not consider profit maximization to be accepted in the community. For example, these hospitals generally provide some charity care, even though it reduces their short-term profits. In the same way that investor-owned hospitals want to avoid the perception that they do nothing for health care for the poor, they also want to avoid a perception that they are “gouging” payers by setting prices higher than are needed to make a “reasonable” profit. Physician Cost Shifting The framework for cost shifting by physicians differs from that for hospitals. Most physicians practice in small practices, which are often partnerships of physicians. Such partnerships are likely to make decisions that reflect the partners’ desire to earn high incomes as well as labor-supply decisions. These considerations are quite different from those of large nonprofit organizations, whose mission is to serve the community. Economists have argued for decades about whether neoclassical models are useful in explaining physicians’ behavior.14 This debate has been driven by empirical results and inforH e a l t h T r a c k i n g W 3 4 7 6 8 O c t o b e r 2 0 0 3 on A uust 4, 2017 by H W T am H ealth A fairs by http://conealthaffairs.org/ D ow nladed fom mal assessments of real-world behavior that suggests that physicians (1) often do not set their prices as high as the market will bear, and (2) induce demand for their services to varying degrees. The notion is that physicians find it distasteful to set prices high or induce demand, so either is done only when an external event, such as a cut in administered prices, threatens to lower their income. Such behavior also could be seen as a realistic accommodation to the complexity of decision making by medical professionals, who would rather spend their time practicing medicine than running a small business. This approach often is ref lected in anecdotes from physicians, which focus on what practices need to do to break even (after the physician has been paid his or her target income). Such informal modeling is often referred to as “supplier-induced demand” or the “target-income hypothesis.” These models provide a basis for cost shifting. If an administered price is reduced, then physicians will work to offset the effects on their incomes by both raising prices to privately insured patients and inducing additional demand for services from patients. Thus, private payers (insurers and privately insured patients) will pay higher prices and pay for additional services. Economists tend to be uncomfortable with models of behavior that do not involve maximization of something. A particular problem with the model sketched above is the absence of an explanation for what determines the target income—at least beyond the short run. Economists have responded by developing maximization models that are more elaborate. For example, James R izzo and David Blumenthal have developed a model that incorporates the notion of relative incomes.15 Physicians have expectations about what their income should be at a particular stage of their career (based on their perception of what their peers earn). Thus, the utility of additional income depends on how physicians’ incomes compare with those of their peers. Rizzo and Blumenthal’s model shows how a reduction in administered prices leads to an increase in prices that physicians have discretion over. Some of the more recent literature has noted that under managed care, physicians have less ability to set prices or induce demand than in the past.16 But this could be changing, as physicians have gained more experience in navigating the managed care market and have more bargaining power as a result of tighter capacity and customers’ demands for broad provider networks.17 But physician cost shifting is not the issue of greatest interest to policymakers in connection with changes in Medicare and Medicaid physician payment rates. Reduction in beneficiaries’ access to care is of greater immediate concern. Whatever the model of physician behavior, theory suggests that reductions in public program payment rates induce physicians to reduce the proportion of time devoted to these programs’ beneficiaries. Of course, with all practices responding to these incentives, the ability of many to increase the proportion of patients covered by private insurance is necessarily limited. Nevertheless, physicians’ ability to change the payer mix of patients distinguishes them from hospitals.