Responding to Structural Change: A Call for a Review of the Competitive Consequences of Hospital Mergers
Comments by Richard J. Gilbert(*)
before the Federal Trade Commission Hearings on Global Competition Policy
November 14, 1995
Thank you for the opportunity to address for a second time the Commission's Hearings on Global and Innovation-Based Competition. Susan DeSanti and Debra Valentine originally invited me to participate in the forthcoming session on network competition in high technology markets. I responded to their invitation with a remark that during my tenure at the Antitrust Division, health care markets posed even more perplexing antitrust issues than high technology markets. While I offered no solutions, my concerns were enough to earn me a place at the table today. I will limit my comments to hospital mergers, although the antitrust challenges extend more generally to various joint ventures and other arrangements in the health care industry.
Structural change in health care
The forces of managed care, the declining per-patient subsidies for Medicare and Medicaid, and the substitution of outpatient services for traditional inpatient procedures have imposed tremendous financial pressures on the nation's hospitals. Hospitals have responded to these forces by searching for new ways to lower their costs and increase revenues. For many hospitals the solution has been to close their operations or to merge with other institutions. The number of U.S. community hospitals declined from over 5,700 in 1984 to under 5,300 in 1993.(1)
The sheer number of hospital mergers is a major challenge to antitrust enforcement. When I was at the Antitrust Division, it seemed that we could exist on a diet of nothing but hospital mergers. It is likely that many of the structural changes in this industry are driven by a pro-competitive desire to reduce costs and improve service quality. In some cases the structural changes may be an anticompetitive reaction to the economic power of managed care. The dilemma, of course, is how to sort proposed mergers into the appropriate categories.
The Federal Trade Commission and the Antitrust Division deserve credit for providing excellent guidance to the health care industry in their joint Statements of Enforcement Policy and Analytical Principles Relating to Health Care and Antitrust. The statements provide a clear and practical discussion of the types of transactions that are likely and unlikely to warrant antitrust scrutiny.
However, the Health Care Statements provide more accurate antitrust advice with respect to network joint ventures than to mergers and acquisitions. The merger safe harbors are important contributions, but they provide little comfort for many of the hospital mergers and acquisitions that have occurred and will continue to occur. The geographic scope of hospital markets is local for many services. Outside of urban areas where patients and third-party payers may choose among several alternative hospitals, most hospital mergers occur in geographic markets that, on their face, appear to raise antitrust concerns. The Department of Justice and the Federal Trade Commission have responded to the wave of hospital mergers by abstaining from challenging many mergers and acquisitions that would suggest antitrust concerns under a literal application of the 1992 Horizontal Merger Guidelines. But the Agencies have not articulated the economic or social policy reasons for their enforcement decisions.
I am not here to suggest a re-writing of the Merger Guidelines for hospital markets. A re-writing of the Merger Guidelines is not necessary because the Guidelines are flexible enough to allow mergers in concentrated markets to proceed without challenge when those mergers pose no substantial threat to competition. Instead, I wish to urge the Commission to use its investigatory powers to learn more about the competitive effects of hospital mergers. The changes in hospital markets are dramatic and profoundly important for our quality of life. The Commission could do a great service by undertaking a critical review of the effects of antitrust enforcement in this industry.
Hospitals pose challenges for competition analysis
Antitrust analysis typically begins with the determination of the relevant product and geographic markets. The analysis proceeds to consider competitive effects, and, if an arrangement may adversely affect competition, considers efficiency justifications. These steps do not require modification for analysis of hospital mergers. However, at each step, hospitals pose considerable challenges.
Consider the definition of relevant product and geographic markets. The typical approach in the evaluation of hospital product markets is to employ a "cluster" of acute care services. At a general level, competition is clearly increasing for many components of the cluster of health care services. Procedures that had required a hospital stay are now being performed at physician offices and out-patient clinics. Regional hospitals, such as the Mayo clinic, have developed centers of excellence in tertiary care and compete for patients over a large geographic area. What is left in the cluster? Obstetric care is probably still a local market. But even obstetric care is feeling the pressures of competition from maternity clinics and from regional hospitals that provide intensive neonatal services.
Antitrust policies have been developed for an economy that is comprised largely of proprietary firms. Most hospitals are not-for-profit. Neither the Agencies nor the courts have proffered an antitrust exemption for not-for-profit organizations and such an exemption probably is not warranted. Some studies, such as Becker and Sloan (1985) using 1979 data, find that costs and profitability are similar for proprietary and not-for-profit hospitals, especially after adjusting for not-for-profit benefits such as tax status and tax-exempt financing. Not-for-profits have financial objectives that may closely coincide with the objectives of a proprietary hospital. Both not-for-profit and proprietary hospitals need to raise funds to invest in new facilities to provide quality care and attract physicians and patients. Not-for-profit hospitals may use surplus revenues to support other programs, in effect turning the hospital into a profit center.
Although not-for-profit and proprietary hospitals are similar in some respects, several studies suggest that not-for-profits behave differently with respect to the determination of prices and other dimensions of service. For example, based on a 1980 survey, Renn et. al. (1985) conclude that average net patient care revenues per admission at investor-owned chain hospitals were 12 percent higher than revenues per admission at not-for-profit chain hospitals and 9 percent higher than revenues per admission at not-for-profit free-standing hospitals. Noether (1988) found that prices charged by proprietary hospitals for Medicare patients were on average 13 percent higher than prices charged by comparable not-for-profit hospitals. Based on a survey of financial reports from 1983 to 1988, Hoerger (1991) found that net income earned by not-for-profits was less variable than the net income earned by proprietary hospitals and was less responsive to changes in reimbursement policies. This suggests that pricing by not-for-profit hospitals is focused more on providing a steady source of revenue than on maximizing profits at each point in time.(2)
These studies do not establish that a merger of not-for-profit hospitals will have significantly different effects on the price and quality of service than an otherwise comparable merger of proprietary hospitals. Yet they do suggest differences that warrant consideration in an antitrust analysis.
Recent antitrust evaluations of hospital mergers have focused on the likelihood that the merged institution will eliminate or reduce discounts to managed care payers. It is reasonable to expect a proprietary hospital to set different profit-maximizing prices for patients that are covered by managed care and patients with full indemnity insurance. It is also reasonable to expect that a merger in a concentrated market would have a greater impact on the profit-maximizing price for managed care payers than on the profit-maximizing price for relatively price-insensitive indemnity patients. Consequently, in most cases it is an appropriate simplification to confine the antitrust analysis of a merger of proprietary hospitals to the competitive impacts on managed care payers.
In contrast to a merger of proprietary hospitals, the competitive effects of a merger of not-for-profit hospitals may not be confined to managed care payers. Suppose a not-for-profit hospital has a self-imposed revenue constraint that is sufficient to cover costs and provide a comfortable margin for investment. If competition for managed care patients causes the hospital to provide discounts for these payers, the result may require higher prices to other customers to meet the hospital's revenue target. Thus, an analysis of the competitive effects of a merger of not-for-profit hospitals may necessitate tradeoffs of competitive impacts for different customers.(3)
There are other important factors that complicate an assessment of competitive effects in hospital markets. Hospitals are partially regulated by government reimbursement policies. The competitive objectives of hospitals are derived from a complex mix of public and private concerns. It is generally recognized that prior to the growth of managed care, hospital competition created a medical arms race that raised costs to consumers with questionable quality benefits.(4) Managed care and cost containments in federal reimbursement programs have slowed the arms race. But hospital competition is still influenced by decision-makers other than the final consumer and the outcomes of that competition are still complicated by important non-price characteristics.
Finally, merging hospitals have become adept at demonstrating large promised efficiencies from their proposed combinations. If these efficiencies are real and if they are passed on to consumers, they likely would justify a conclusion that most hospital mergers are not anticompetitive. The validity of the claimed efficiencies from hospital mergers and the likelihood that they will be passed on to customers are both open question. However, I do not believe that they are easily dismissed.
Another look at the market structure-price relationship for hospitals
A recent study of California hospitals provides some limited support for the view that the risk to competition from hospital mergers may be low relative to the possible efficiency benefits. Using 1992 data collected by Dranove and White (1995) for 203 private (not-for-profit and proprietary) hospitals in California, my colleague Carlo Cardilli and I analyzed the correlation between hospital price-cost margins and several explanatory variables. These included the number of other hospitals in each hospital's estimated geographic market, managed care penetration in the geographic market, the percentage of Medicare and Medicaid patients, the size of the hospital, and several other factors, such as its status as a teaching hospital, technology mix, and whether the hospital was a proprietary organization.
As we expected, the results showed a significant, negative correlation between the proportion of hospital patients covered by managed care and the average price-cost margin of the hospital. The data support the conclusion that managed care is effective in limiting hospital prices. We were more surprised to find no statistical relationship between price-cost margins and the level of market concentration. Specifically, holding managed care penetration constant, there was no significant difference in price-cost margins for markets with one, two, three, or more hospitals.
These results are incomplete. The geographic market definition is based on fairly rudimentary patient flow data. Many hospitals did not provide sufficient data to be included in the analysis. The causal relationship between price-cost margins and managed care penetration is complicated. The analysis accounted for the level of managed care penetration in different markets and various estimates were made employing standard statistical methods to account for interactions between managed care penetration and expected price-cost margins. But these are only first steps in a larger project. Clearly, there is much that we do not know and these results are merely suggestive.(5)
Summary -- a research agenda
Let me emphasize that I am not advocating special antitrust rules for health care. I am not advocating an expansion of the failing firm or the failing industry defense. I am advocating a search for knowledge. The stakes are high enough to justify a major investigation by the Commission of the competitive effects of structural changes in hospital markets and perhaps other health care organizations. I urge you to support such an investigation and to give particular consideration to the following questions:
(1) Do hospitals in more concentrated markets charge higher prices?
(2) Does the structure-conduct-performance relationship differ for not-for-profit and proprietary hospitals?
(3) Is there a close relation between promised efficiency measures and the actual measures that are implemented after a hospital merger or acquisition?
(4) Is there evidence that mergers in concentrated hospital markets have led to higher prices?
(5) Is there evidence that mergers in concentrated hospital markets have led to changes in the breadth and quality of services?
(6) Is there evidence that mergers in concentrated hospital markets have slowed the rate of managed care penetration?
(7) Have discounts secured by managed care payers resulted in higher prices for other payers?
Thank you for the opportunity to address the Commission on this important subject.
American Hospital Association (1194), Hospital Statistics, annual, reported in The Economic Report of the President.
Becker, E. and F. Sloan (1985), "Hospital Ownership and Performance," Economic Inquiry, vol. 23, pp. 21-36.
Congressional Budget Office (1993), Responses to Uncompensated Care and Public Program Controls on Spending: Do Hospitals Cost-Shift?, Washington.
Dranove, D. and W. White (1995), "Government Cutbacks in Hospital Reimbursement: Who Bears the Burden?," Northwestern University working paper.
Hoerger, T.J. (1991), "'Profit' Variability in For-Profit and Not-For-Profit Hospitals," Journal of Health Economics, vol. 10, pp. 259-289.
Joskow, P. (1980), "The Effects of Competition and Regulation on Hospital Bed Supply and the Reservation Quality of the Hospital," Bell Journal of Economics, vol. 11, pp. 421-447.
Noether, M. (1988), "Competition Among Hospitals," Journal of Health Economics, vol. 7, pp. 259-284.
Robinson, J. and H. Luft (1985), "The Impact of Hospital Market Structure and Patient Volume on Average Length of Stay and the Cost of Care," Journal of Health Economics, vol. 4, pp. 333-356.
Robinson, J., H. Luft, S. McPhee, and S. Hunt (1988), "Hospital Competition and Surgical Length of Stay," Journal of the American Medical Association, vol. 259, pp. 696-700.
Renn, S.C, C.J. Schramm, J.M. Watt and R.A. Derzon (1985), "The Effect of Ownership and System Affiliation on the Economic Performance of Hospitals," Inquiry, vol. 22, Fall, pp. 219-236.
U.S. Department of Health & Human Services (1992), Effects of Hospital Mergers on Costs, Resources and Patient Volume, Office of the Inspector General.
Weisbrod, B.A. and M. Schlesinger (1986), "Public, Private, Nonprofit Ownership and the Response to Asymmetric Information: The Case of Nursing Homes," in S. Rose-Ackerman (ed.), The Economics of Nonprofit Institutions, Oxford University Press.
(3) Evidence of cost-shifting behavior by not-for-profit hospitals is mixed. Hoerger (1991) and the Congressional Budget Office (1993) concluded that such cost-shifting behavior is common. However, Dranove and White (1995) found little evidence of cost-shifting. After a cut in hospital Medicaid reimbursements, they found that hospitals did not raise prices to privately-insured patients. Dranove and White's analysis suggests that even not-for-profit hospitals set prices based on the individual characteristics of each patient group and do not shift costs between patient groups.