Full opinion text
OPINION and ORDER MELLOY, Chief Judge. The United States has brought this action under the antitrust laws. The defendants, Mercy Health Services and Finley Tri-States Health Group, Inc. own Mercy Health Center (Mercy) and Finley Hospital (Finley), respectively. Mercy and Finley are the only two general acute care hospitals in Dubuque, Iowa. The two hospitals have agreed to form a partnership, Dubuque Regional Health Services (DRHS), which all parties acknowledge constitutes a merger for purposes of antitrust analysis. On June 10, 1994, the government filed a complaint seeking injunctive relief against the merger as a violation of § 7 of the Clayton Act, 15 U.S.C. § 18, and of § 1 of the Sherman Act, 15 U.S.C. § 1. The parties agreed to waive a preliminary injunction hearing in order to proceed to trial on the matter. After expedited discovery, the matter was tried to the bench. Two weeks of testimony was heard and several hundred items were moved into evidence. I. Findings of Fact A. Hospital Background Dubuque is situated within Dubuque County, Iowa, which in 1993 had a population of 86,403. Dubuque lies on the Mississippi River, at a point where Iowa adjoins the southwestern portion of Wisconsin and the northwestern portion of Illinois. Mercy is an acute care hospital which, in 1994, had approximately 320 staffed beds, 9980 acute care patient discharges and an average daily census of 127. Mercy’s acute care commercial discharges for 1994 were estimated to be 3622 and the average daily acute care commercial census was 44. Finley is also an acute care hospital and, in 1994, was estimated to have 124 staffed beds, 5247 acute care patient discharges and an average daily census of 63. Finley’s acute care commercial discharges for 1994 were estimated to be 2175 and the average daily acute care commercial census was 21. There are seven rural hospitals in the area: Galena-Stauss Hospital in Galena, Illinois, 15 minutes from Dubuque, has 25 licensed beds and an average daily census of 3. Southwest Health Center in Platteville, Wisconsin, 30 minutes from Dubuque, has 35 licensed beds and an average daily census of 11. Lancaster Memorial Hospital in Lancaster, Wisconsin, 30 minutes from Dubuque, has 35 licensed beds and an average daily census of 10. Delaware County Memorial Hospital in Manchester, Iowa, 40 minutes from Dubuque, has 58 licensed beds and an average daily census of 12. Jackson County Public Hospital in Maquoketa, Iowa, 30 minutes from Dubuque, has 99 licensed beds and an average daily census of 12.4. Guttenberg Memorial Hospital in Guttenberg, Iowa, 40 minutes from Dubuque, has 37 licensed beds and an average daily census of 7. Finally, Central Community Hospital in Elkader, Iowa, 60 minutes from Dubuque, has 29 licensed beds and an average daily census of 3-4. These hospitals primarily serve patients who are closer to the rural hospital than to any other hospital. The rural hospitals mainly provide primary care services and do not provide the breadth of services Mercy and Finley offer. Guttenberg provides the largest array of services in that it offers 68% of the same Diagnosis Related Groups (DRGs) Mercy and Finley provide. Galena-Stauss has the smallest array of services, providing care for only 11.5% of the same DRGs Mercy and Finley provide. There are several regional hospitals within 70 to 100 miles of Dubuque which generally offer the same or greater range of services as provided by Mercy and Finley. Allen Memorial Hospital in Waterloo, Iowa has 194 staffed beds and an average daily census of 145. Covenant Medical Center in Waterloo, Iowa has approximately 322 staffed acute care beds with an average daily census of 173. St. Luke’s Methodist Hospital in Cedar Rapids, Iowa has 441 staffed beds with an average daily census of 284. Mercy Medical Center in Cedar Rapids, Iowa has 353 staffed beds and an average daily census of 193. St. Mary’s Medical Center in Madison, Wisconsin has 345 staffed beds and an average daily census of 265. Freeport Memorial Hospital in Freeport, Illinois has 143 staffed beds and an average daily census of 70. University of Wisconsin Hospital and Clinics in Madison, Wisconsin has 496 staffed beds and an average daily census of 386. Meriter Hospital in Madison, Wisconsin has 429 staffed beds and an average daily census of 258. The University of Iowa Hospitals and Clinics in Iowa City, Iowa (UIHC) has 868 staffed beds and an average daily census of 677. Mercy and Finley’s patient bases are composed of individuals covered by government insurance programs, traditional indemnity insurance and managed care payers. Managed care payers include health maintenance organizations (HMO’s), and preferred provider organizations (PPO’s). HMO’s generally charge a set fee which covers all of an enroll-ee’s health care needs, including hospitalizations. HMO’s generally restrict the doctors and hospitals from which an enrollee can receive care to those physicians and hospitals providing a discounted rate to the HMO. HMO’s often work with the hospitals when an enrollee is hospitalized to insure that the costs of the hospitalization remain as low as possible. HMO’s may have their own clinics to which enrollees are obligated to go for office visits and most outpatient needs. HMO’s generally stress preventative care and require preapproval prior to being hospitalized in order to keep the rate of hospitalizations low. PPO’s generally negotiate discounted rates with physicians and hospitals and then require their enrollees to receive their care from the discounted care providers or risk being denied reimbursement. PPO’s generally do not have their own clinics and do not stress preventative care. In contrast to the managed care payers, the indemnity health insurers have not traditionally attempted to gain discounted rates from physicians or hospitals. Instead, these traditional insurers cover a percentage of the health care costs with the remainder being paid by the insured. HMO’s and PPO’s have only recently established themselves in Iowa, but already 25% of Mercy and Finley’s patients are covered by managed care payers. Of the remaining 75% of Mercy and Finley’s inpatients, 50% are covered by Medicare and Medicaid and the other 25% are covered by traditional indemnity insurance. A hospital will discount its stated charges to managed care payers in order to entice the managed care entity to send more enroll-ees to their hospital for inpatient care. Mercy currently gives discounts to the following managed care entities: Medical Associates, Heritage National Health Plan, Self-Insured Systems Corporation (SISCO), Alliance Select PPO (a Blue Cross plan), HMO of Wisconsin, Wisconsin Education Association Insurance and the Affordable Health Plan. Finley contracts with Blue Cross/Blue Shield of Iowa, Alliance Select PPO Program (a Blue Cross plan), Heritage National Health Plan, Inc., Heritage Preferred, Medical Associates HMO, HMO of Wisconsin, Blue Cross of Illinois PPO, BeefAmerica Operating Co., Pro America Network, Inc., American Health Care Advisory Association, Hospice of Du-buque County, and Collaborative Medical for Religions (a charitable contribution). Medical Associates is a physician-owned medical practice which operates the Medical Associates HMO in the Dubuque area. The HMO has approximately 30,000 enrollees. The Medical Associates HMO accounts for 11-12% of Mercy’s net revenues. Physicians employed by Medical Associates refer patients to Mercy for inpatient and outpatient care constituting approximately 80% of Mercy’s yearly revenue. Heritage National Health Plan operates a PPO which covers 5,000 lives in the Dubuque area. Nationally it covers 280,000 persons. It accounts for 2-3% of Mercy’s net revenues. SISCO is a third-party payer which administers self-insured employee health plans and also contracts for reduced rates for some of its employer-customers. Blue Cross has a state-run managed care plan in which it offers the hospital a contract on a “take it or leave it” basis. The hospital can either agree to the stated rates or not be on the plan’s preferred provider list. It accounts for 8-9% of Mercy’s net revenues. Affordable accounts for less than 1% of Mercy’s revenues. HMO Wisconsin accounts for less than 0.5% of Mercy’s revenues. Medicare and Medicaid account for 46-47% of Mercy’s net revenues. Traditional indemnity insurance accounts for 26% of Mercy’s revenues. Medical Associates has a clinic in Dubuque in which it sees patients for regular office visits. It also operates an outpatient clinic in Dubuque which directly competes with Mercy and Finley for outpatient procedures. In addition to its Dubuque Clinic, Medical Associates has permanent facilities in Bellevue, Iowa; East Dubuque, Illinois; and Galena, Illinois. It operates outreach clinics in Belle-vue, Iowa; Dyersville, Iowa; Lancaster, Wisconsin; Maquoketa, Iowa; Platteville, Wisconsin; Boscobel, Wisconsin; Elkader, Iowa; Guttenberg, Iowa; Manchester, Iowa; Oel-wein, Iowa; Darlington, Wisconsin; Clinton, Iowa; and Montieello, Iowa. These outreach clinics are primarily conducted in order to obtain secondary referrals for Medical Associates’ specialists. B. Health Care Market Trends Traditionally, hospitals competed on the basis of amenities and perceptions of quality. Only in the last ten to fifteen years have hospitals begun to compete on the basis of price. To a large degree, this competition has occurred because of the arrival of managed care. These large purchasers shop on the basis of price and are able to induce hospitals to discount stated charges in return for the managed care payer’s promise to direct a larger volume of patients to the hospital. Hospitals must employ cost containment measures in order to become more cost efficient so that they can afford to give discounts to the managed care payers. Competition for Medicare, Medicaid and traditional indemnity patients still occurs on the basis of amenities and perceptions of quality. Managed care entities are capable of inducing their enrollees to receive inpatient care at a specific hospital. The entity does this in one of the following ways: (1) by only paying the hospitalization costs if the enroll-ee is treated at a specified hospital, or (2) by requiring the enrollee to pay a higher percentage of the enrollee’s hospitalization costs (a co-pay) if the enrollee chooses to be hospitalized at a non-preferred hospital. Managed care entities are gaining sophistication in their ability to induce patients to be hospitalized at a preferred hospital. However, the managed care entities do not always show price sensitivity when they establish requirements. For example, Medical Associates may be charged a lower rate by one hospital for a certain DRG, however, the physicians who admit the patients may not be aware of the rate differential and consequently do not encourage patients to go to the cheaper hospital. Traditionally, when people decided where to receive inpatient care, they considered the following factors: where their physician had hospital privileges and could therefore take care of them during the hospitalization; what hospital was convenient; the reputation of the hospital; and the amenities the hospital offered. Patients have only recently added out-of-pocket expenses to this list. (Traditional indemnity insurance covered a specific portion of the fees no matter which hospital was utilized and patients generally had little knowledge of whether hospital prices varied for their specific hospitalization needs.) The advent of managed care has resulted in individuals also considering what their out-of-pocket expenses will be. Under traditional indemnity insurance a certain percentage of a hospitalization is covered without regard to the hospital used. Managed care enroll-ees are much more aware of their out-of-pocket expenses resulting from a hospitalization because the policies generally state that expenses incurred at certain hospitals will either not be covered or will require a higher co-pay. People who continue to be covered by either traditional insurance or by Medicare and Medicaid remain largely insensitive to price differentials. The government dictates hospital fees for the care of Medicare and Medicaid patients, and such fees are often below the cost of providing the services. This causes “cost shifting” of the costs of these services to non-Medicare and non-Medicaid patients. Patients covered by traditional indemnity insurance have been charged full charges which take into account the need to cost shift. Even though managed care groups negotiate with the hospitals for rates below full charges, the rates are greater than those charged Medicare and Medicaid patients and include some cost shifting. Price competition is important among the managed care groups. Generally, the managed care payer will negotiate the best rates and the greatest discounts possible with area hospitals and physician groups. Based on the rates obtained, the managed care payer determines the premium it must charge en-rollees for services. The managed care group next approaches businesses to try to convince them to offer their managed care plan to their employees. Thirty-five to 40% of the managed care premium is due to inpatient hospital charges, so the rates negotiated with the hospitals significantly impact the premium charged by the managed care payer and affects the cost competitiveness of the managed care entity receiving the discount. Hospitals contract at discounted rates with the managed care entity when they feel the managed care payer can induce its enrollees, through financial incentives or otherwise, to use the contracting hospital. A hospital will generally only contract with the managed care entity if it feels that (1) by giving a discount, the managed care entity will shift a higher volume of patients to the hospital, or (2) by refusing to give a discount, the managed care entity will shift a significant volume of patients away from the hospital. Managed care plans influence which physicians and hospitals an enrollee chooses by offering incentives to use certain care providers (i.e., a reduction in co-pay or a total waiver of co-pay), by requiring a co-pay if the enrollee uses a different care provider or hospital, or by simply requiring the enrollee to use designated hospitals and care providers to receive reimbursement. Along with the rise of managed care has come the increased use of outpatient services for procedures which traditionally required overnight hospitalization. The length of hospital stays are also becoming significantly shorter because of the increased use of home care and because managed care payers closely monitor the length of time an enrollee is hospitalized. This results in less hospital utilization and a decrease in the average daily census. Because hospitals have high fixed costs, the loss of patient volume requires hospitals to broaden their service area in order to maintain patient census and resulting cash flow. Prior to the trend in shorter and fewer hospitalizations and to the advent of cost competition, hospitals had their own “catchment area” from which they drew their patients and from which other hospitals generally did not try to encroach upon the residing population. The main competition between regional hospitals was in the “fringe area,” the area half-way between two regional hospitals. However, the competitive market has changed and is continuing to change at a rapid pace. Hospitals, needing to maintain their patient volume, have begun to establish outreach efforts in the fringe areas as well as deep in the heart of what used to be considered another hospital’s catchment area. The efforts include the establishment of hospital owned clinics and the purchase of physician practices in towns up to 50-60 miles from the hospital. These clinics have succeeded in capturing patients and referring them to their associated hospital for inpatient care. Outreach efforts also include climes established by private physicians’ groups. The physician clinics are used to broaden the referral base for the clime’s various specialties. The clinics refer patients back to the hospital where the physicians have admitting privileges for inpatient and outpatient services. Consequently, when a physicians’ group extends its market area by opening an outreach clinic, it also extends the market area of the hospital with which the group is associated. Once an outreach clinic is established, the hospital and specialty referral patterns from the community where it is located change significantly within a short period of time (often within two months). Clinics can generally be established in a matter of weeks as they are often opened in existing hospitals or clinics. Both hospital and physician-owned clinics have proven they induce outreach patients to receive hospital care at the associated hospital, even if another regional hospital is closer. There is a great deal of evidence showing that these national trends exist in the Du-buque area. Mercy’s inpatient volume decreased by approximately 15% from 1990 to 1994. In order to maintain its present patient volume and revenues, Mercy must gain market share in what it considers its “secondary market,” the area outside of Dubuque County. This can only be done on a significant level by establishing outreach clinics. Mercy has not established any of its own outreach clinics, but depends on the managed care payers which use Mercy to do so— primarily Medical Associates. C. Other Findings of Fact Additional findings of fact are made under the Conclusions of Law section and throughout the legal analysis. In addition, attached as Appendix A, are the facts to which the parties have stipulated. II. Conclusions of Law A. Anti-trust Analysis Section 7 of the Clayton Act provides, in part, that “no person ... shall acquire the whole or any part of the assets of another person ... where ... the effect of such acquisition may be substantially to lessen competition.” 15 U.S.C. § 18 (1973 & Supp.1995); United States v. E.I. du Pont de Nemours & Co., 353 U.S. 586, 589, 77 S.Ct. 872, 875, 1 L.Ed.2d 1057 (1957) (Section 7 is “designed to arrest in its incipieney ... the substantial lessening of competition....”). To meet the requirement of Section 7, the government must show a reasonable probability that the proposed transaction would substantially lessen competition in the future. FTC v. University Health Inc., 938 F.2d 1206, 1218 (11th Cir.1991); FTC v. Warner Communications, Inc., 742 F.2d 1156, 1160 (9th Cir.1984). Section 1 of the Sherman Act requires the same analysis. 2 Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 304c., at 9 (1995); United States v. Rockford Mem. Corp., 898 F.2d 1278, 1281-86 (7th Cir.), cert. denied, 498 U.S. 920, 111 S.Ct. 295, 112 L.Ed.2d 249 (1990). To determine whether there is a reasonable probability of a substantial lessening of competition, the courts have focused on whether the merger will cause the merged entity to have enough market power such that it could profitably increase prices. Whether the merger will cause an anticompetitive effect is not “the kind of question which is susceptible of a ready and precise answer in most cases.” United States v. Philadelphia National Bank, 374 U.S. 321, 362, 83 S.Ct. 1715, 1740, 10 L.Ed.2d 915 (1963). Generally, the plaintiff attempts to prove its case by showing that the merged entity will have a large percentage of the relevant market such that, either individually or through collusion, its hold over the market will enable it to profitably raise prices above competitive levels. The government may make a prima facie showing that the merger will result in anticompetitive effects by showing that the merged entity will have an undue share of the relevant market. Id., at 363, 83 S.Ct. at 1741; University Health, Inc., 938 F.2d at 1218. Thus before a court can determine the effect of a merger on competition, it must first define the relevant market. The relevant market is defined by identifying those “producers that provide customers of defendants]’ ... firms with alternative sources for the defendants]’ product or service.” 2A Phillip E. Areeda, Herbert Hovenkamp & John L. Solow, Antitrust Law ¶ 530a, at 150 (1995). A properly defined market excludes other potential suppliers (1) whose product is too different (product dimension of the market) or too far away (relevant geographic market) and (2) who are not likely to shift promptly to offer defendants’ customers a suitably proximate alternative. Id. “Those who can readily shift into offering such a product are in the market.” Id. Although a great deal of emphasis is placed on market share statistics, they are not conclusive indicators of anticompetitive effects. United States v. General Dynamics, 415 U.S. 486, 498, 94 S.Ct. 1186, 1194, 39 L.Ed.2d 530 (1974). If the government is able to make a prima facie case, the defendants can overcome the presumption of illegality by showing that the market-share analysis gives an inaccurate reflection of the acquisition’s probable effect on competition within the relevant market. United States v. Citizens & Southern Nat’l Bank, 422 U.S. 86, 120-21, 95 S.Ct. 2099, 2118, 45 L.Ed.2d 41 (1975). To rebut the government’s case, the defendants may produce “ ‘[njonstatistical evidence which casts doubt on the persuasive quality of the statistics to predict future anti-competitive consequences.’ ” University Health Inc., 938 F.2d at 1218 (quoting Kaiser Aluminum & Chem. Corp. v. FTC, 652 F.2d 1324, 1341 (7th Cir.1981)). Factors which courts have previously considered to be relevant include “ ‘ease of entry into the market, the trend of the market either toward or away from concentration, and the continuation of active price competition.’ ” Id. When determining whether a merger will have an anticompetitive effect, the court must view the merger functionally within the context of its industry’s “ ‘structure, history, and probable future.’ ” General Dynamics, 415 U.S. at 498, 94 S.Ct. at 1194 (quoting Brown Shoe v. United States, 370 U.S. 294, 322 n. 38, 82 S.Ct. 1502, 1522 n. 38, 8 L.Ed.2d 510 (1962)). If the defendant is successful in rebutting the government’s statistical evidence, the government then has the burden of presenting additional evidence that the merger will lessen competition and this burden “ ‘merges with the ultimate burden of persuasion, which remains with the government at all times.’ ” University Health Inc., 938 F.2d at 1219 (quoting United States v. Baker Hughes Inc., 908 F.2d 981, 983 (D.C.Cir.1990)). B. Relevant Product Market The' parties have agreed that the relevant product market is acute care inpatient services offered by both Mercy and Finley. This definition excludes inpatient psychiatric care, substance abuse treatment, rehabilitation services, and open-heart surgery. This limits the product market to those services for which Mercy and Finley currently compete for inpatients. C. Relevant Geographic Market The burden is on the government to prove the relevant geographic market. Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447, -, 113 S.Ct. 884, 890, 122 L.Ed.2d 247 (1993). The relevant market is hotly disputed by the parties. The government asserts that the geographic market includes Du-buque County, Iowa and a half-circle with a 15 mile radius extending from Dubuque County’s eastern edge into Illinois and Wisconsin. This geographical area would include Mercy, Finley and Galena-Stauss Hospital. Currently, 88% of the patients within the market use the three hospitals within the market. Only about 2% of the 88% use Galena-Stauss; the remainder use Mercy and Finley. Approximately 76% of the patients at the three hospitals come from within the market. The defendants assert that the proper geographic market includes Mercy, Finley, the seven closest rural hospitals and the regional hospitals situated in Cedar Rapids, Waterloo, Iowa City, Davenport, and Madison. Mercy and Finley’s market share of the patients within this geographical area is approximately 10%. The government does not contest that if this is the relevant market, the merger is not illegal. The parties dispute whether the government has properly defined the relevant geographic market by trying to show whether DRHS could (1) profitably raise prices by 5% to managed care entities, or (2) profitably raise prices by completely eliminating present discounts to managed care entities. In the event the court finds the government has not met its burden as to the proposed geographic market, the government has proposed an alternative market encompassing only the City of Dubuque. The government defined its market based on (1) the views of the major purchasers of health care services in the Dubuque area, (2) where physicians have privileges, (3) the views of physicians, and (4) patient flow data. Looking at the views of the major managed care providers in the Dubuque area, the government finds it significant that Medical Associates HMO and the Heritage Plan both feel their health care plans must include one of the Dubuque hospitals in order for the plan to be marketable to Dubuque employers. The government argues this is evidence that there is no market substitute for the Dubuque hospitals. The government then looked at the overlap of physicians who were currently on staff at the area hospitals. The government showed that there is a great deal of physician staff overlap between the staffs of Mercy and Finley, but very little between the rural hospitals and Mercy or Finley. Seventy-six percent of Mercy’s physicians were shown to be on staff at Finley and over 90% of Finley’s physicians were on staff at Mercy. The government concludes from these statistics that (1) those patients who currently use a physician with privileges at Mercy and Finley cannot switch to a rural hospital due to their loyalty to their physician and their physician’s inability to admit them at a rural facility, and (2) Mercy and Finley are seen by physicians as being comparable hospitals as physicians are willing to work out of either of them. By contrast, Dubuque physicians are not willing to work out of the rural hospitals, hence the rurals are not comparable and therefore they are not substitutes for Mercy and Finley. The government also asserts that Dubuque physicians would not seek privileges at the rural hospitals as they do not see the rural hospitals as being able to provide the broad range of services which the majority of their patients require and because the rurals are not adequately staffed and do not have the equipment to do many inpatient surgical procedures. The government then goes on to look at the inflow and outflow of patients from the market. The “Elzinga-Hogarty” test was employed by the government to argue that the relevant geographical market was quite limited. This test looks at the empirical data to determine (1) the area from which the hospitals draw their patients and (2) where the residents in that area go for hospital care. When both numbers are at 90%, the market definition is said to be “strong”. Numbers at 75% are said to constitute a “weak” market definition. Robert Pitofsky, New Definitions of Relevant Market and the Assault of Antitrust, 90 Columbia Law Rev. 1805 (1990); Elzinga & Hogarty, The Problem of Geographic Market Delineation Revisited: The Case of Coal, 23 Antitrust Bull. 1, 2 (1978). Looking at the government’s proposed market definition — Dubuque County and a fifteen mile radius into Illinois and Wisconsin — 76% of Mercy’s and Finley’s inpatients come from within this defined market, and 88% of those residing within the market seek care within the market. Shrinking the market to include only the City of Dubuque, 55% of'Mercy and Finley’s inpatients come from within the market and 91% of those residing in Dubuque seek hospital care at either Mercy or Finley. Defining the market as a twenty-five mile radius around Dubuque, 85% of Mercy’s and Finley’s patients come from within the market and 89% of those persons within the market seek care inside the market. Broadening the market to a 35 mile radius around Dubuque, 89% of Mercy and Finley’s inpatients come from within the market and 82% of the persons within that market seek care at Mercy or Finley. Finally, the government asserts that the regional hospitals are not in the relevant geographic market as they only compete for patients at the “fringes,” the area located at the midpoint between Dubuque and a particular regional hospital. The government reasons that people will not travel greater distances than necessary for inpatient care and, therefore, only those patients at the midway point between two regional hospitals could be induced to switch hospitals due to price concerns. The government concludes that DRHS would have the ability to raise prices as the rurals do not act as a competitive alternative to DRHS and the regional centers are only a competitive force at the fringes of Mercy’s and Finley’s service areas and therefore could not induce enough people to switch hospitals to defeat a price rise by DRHS. The defendants criticize the government’s market definition as being dependent on incorrect assumptions, and as not accounting for current market trends. The defendants also assert that the government’s geographical market is too small because a 5% price increase would be easily defeated by patients who would choose to travel to other hospitals for cheaper care. In reaching this result, the defendants challenge the government’s assumptions of strong doctor-patient loyalty, that outreach clinics only have an effect on hospital use at the fringes of a hospital’s service area, and that there are significant barriers to entry into this market. 1. Analysis of the Relevant Geographic Market The court finds that the government’s case rests too heavily on past health care conditions and makes invalid assumptions as to the reactions of third-party payers and patients to price changes. The government’s case also fails to undergo a dynamic approach to antitrust analysis, choosing instead to look at the situation as it currently exists within a competitive market. In determining the relevant geographic market, the court must determine “the market in which the seller operates and to which the purchaser can practicably turn for supplies.” Tampa Elec. Co. v. Nashville Coal Co., 365 U.S. 320, 327, 81 S.Ct. 623, 628, 5 L.Ed.2d 580 (1960); Philadelphia National Bank, 374 U.S. at 359, 83 S.Ct. at 1739. The analysis must focus not merely on where patients have gone for acute inpatient services, but where they could practicably go should DRHS become anticompetitive. Bathke v. Casey’s General Stores, Inc., 64 F.3d 340, 345 (8th Cir.1995); Morgenstern v. Wilson, 29 F.3d 1291, 1295-1296 (8th Cir. 1994), cert. denied, — U.S. -, 115 S.Ct. 1100, 130 L.Ed.2d 1068 (U.S.1995). This requires a fluid analysis. It is not sufficient to take a snapshot of the current situation and define the relevant geographic market to be synonymous with the current service areas of the defendant hospitals. See, Bathke, 64 F.3d at 345; Morgenstern, 29 F.3d at 1295; In the Matter of Hospital Corporation of America, 106 FTC 361, 466 (1985), aff'd, Hospital Corporation of America v. FTC, 807 F.2d 1381 (7th Cir.1986), cert. denied, 481 U.S. 1038, 107 S.Ct. 1975, 95 L.Ed.2d 815 (1987); Drs. Steuer & Latham v. Nat. Med. Enterprises, 672 F.Supp. 1489, 1511 (D.S.C.1987), aff'd, 846 F.2d 70 (4th Cir.1988). In order to determine the actual geographic market, current market behavior must be put into a dynamic analysis which looks at possible competitive responses from other hospitals, third party payers and consumers. In this light, it is important to note that the government’s reliance on the Elzinger-Ho-garty test is merely a starting point — it states where Mercy and Finley are currently attracting patients and the area from which most patients seek services from either Mercy or Finley. This is the snapshot of the market as it exists under current conditions and does not pretend to answer the question of what would happen if there was an attempt to exercise market power by one of the market participants. See Robert Pitof-sky, New Definitions of Relevant Market and the Assault of Antitrust, 90 Co-lum.L.Rev. 1805 (1990). The government’s case in general makes the mistake of relying too heavily on past conditions in the hospital industry and discounts the effects of outreach efforts, the strong emphasis that regional hospitals place on expanding their service areas and the willingness of managed care enrollees to make changes in their health care for financial reasons. The government’s case relies on several assumptions and conclusions that are not supported by the evidence. One of the assumptions is that there is strong doctor-patient loyalty. This assumption supports the argument that patients cannot be enticed to use a non-Dubuque hospital because the patient already has a Dubuque doctor to whom the patient is very loyal. There are several problems with this assumption. First, the evidence shows that' many hospital patients do not have an established physician relationship, or, are admitted by a specialist who is not the patient’s regular physician. Secondly, the evidence does not support a finding of strong patient-physician loyalty that prevents patient shifting for financial reasons. The medical professionals testified there is a large amount of patient shifting that occurs when an employer changes or modifies a health insurance plan so as to encourage usage of a particular physician or group of physicians. Testimony also shows that managed care entities have been very successful in enrolling patients in HMO’s that require switching physicians or agreeing to be seen by any available physician. The government’s assertion that the outreach clinics have little effect on hospitalization patterns is also erroneous. There are several communities where referral practices have been dramatically altered due to outreach efforts. Most notably Manchester, Independence and Clarion, Iowa. While the effect of the outreach clinics has not been quantified, the fact that they are established is significant as it is only through the generation of inpatient and outpatient referrals that these clinics become profitable. Outreach efforts have been characterized by the regional hospitals as being pivotal in their efforts to expand patient volume at a time when all other trends are causing them to lose volume. In other words, without growth in these areas most hospitals acknowledge they would incur a significant decrease in revenue. The start-up of outreach clinics cannot be justified based solely upon the direct business they obtain on site. They are economically justified due to the increased referrals they bring to the physicians and hospitals sponsoring them. Outreach clinics generally attract patients from a fifteen mile radius and these clinics have been established within twenty-five miles of Dubuque. 18.8% of Mercy and Finley’s patients come from within fifteen miles of an outreach clinic associated with another regional hospital. The government attempts to argue that outreach clinics are only effective in shifting patients from the fringe areas to other hospitals as people will not agree to be hospitalized at a hospital a significant distance away from their home and family. The government argues that most hospitalizations require two pre-operative visits to the hospital, the actual hospitalization and two or three post-operative visits to the hospital and patients will not want to incur the expense and inconvenience of travel for these visits. However, the evidence was clear that one reason why the outreach clinics are so effective in altering hospitalization practices is because the pre- and post-operative visits take place at the outreach clinic and the patient is only required to travel to the hospital for the actual surgical procedure. In fact, people will use outreach clinics which are associated with hospitals that are a significant distance from their hometown. To illustrate, the Platteville and Lancaster areas are 25 miles from Dubuque and sixty to seventy miles from Madison, Wisconsin, yet Medical Associates, which refers persons back to Dubuque for inpatient care, meets strong competition in these communities from the Dean Clinic which refers its patients to St. Mary’s Hospital in Madison. The fact that the residents of southwest Wisconsin are willing to drive to Madison for their inpatient needs also shows that the government’s assumption that persons within twenty-five miles of Dubuque will only go to Dubuque for their inpatient needs is an incorrect assumption. The government has also failed to account for the fact that there are several zip codes within 25 miles of Du-buque which currently send over one-third of their inpatients to other hospitals. Looking at the zip codes encompassing Cuba City, Wisconsin; Galena, Illinois; New Vienna, Iowa; Potosi, Wisconsin; and Hazel Green, Wisconsin, there were 930 discharges in a six month period from these zip codes and only 560 were discharged from Mercy or Finley (60.2%). 179 of these discharges were to hospitals other than Mercy, Finley, the seven rurals or the University of Iowa (19.2%). Obviously, these persons are already being attracted to hospitals outside the immediate area. Discharges from these five zip codes account for 7.7% of Mercy and Finley’s total discharges. The main point made by the government which the court does agree with is the inability of the rural hospitals to prevent DRHS from acting in an anticompetitive manner. The government argues that it would not be feasible for more patients to use the rural hospitals as the rural hospitals are not equipped to handle anything but the least complicated hospitalizations. The defendants counter that the rural hospitals are a practicable alternative as they have a great number of empty beds and are capable of meeting the needs of a significant number of inpatients. Due to the high utilization rates of some of the rurals by those in the same zip code as the hospital, many of the rurals can apparently care for a wide range of diagnoses. For example, in the zip code encompassing Guttenberg, Iowa, where Guttenberg Hospital lies, 178 out of 269 hospitalizations were at the Guttenberg hospital (66.1%). Further, Guttenberg Hospital has cared for inpatients hospitalized for 68% of the DRGs for which Mercy and Finley admit patients. However, in the zip code encompassing Galena, Illinois, where Galena-Stauss lies, only 93 out of 340 hospitalizations were at Galena-Stauss Hospital (27.3%) and Galena-Stauss has admitted patients for only 27% of the DRGs for which Mercy and Finley admit patients. Therefore, the rural hospitals have a wide divergence as to each’s ability to care for more patients than are currently hospitalized there. This evidence shows that the rural hospitals are technically qualified to meet the hospitalization needs of more patients than currently choose to go there. However, other barriers exist to prevent greater utilization of the rurals, the largest barrier being individual attitudes toward the rurals. Persons living any distance from the rurals simply do not seek them out for care. For example, despite the high utilization rate by the population living within the same zip code as Guttenberg Hospital, in the rest of Clayton County, in which the Guttenberg hospital lies, only 92 out of 579 hospitalizations were at the Guttenberg Hospital (15.8%). In all of Jo Daviess County, in which Galena, Illinois is a part, there were 63 out of 702 hospitalizations at Galena-Stauss (8.9%). Further, a survey done by Jackson County Public Hospital in Maquoketa found they were significantly used by persons within 20 miles of the hospital, but not by persons of any greater distance. Another barrier to the rurals taking on a greater patient volume is the difficulty the rurals have in obtaining qualified doctors and nursing staff willing to live and work in a rural community. It is not at all apparent that the rurals could attract the staff necessary to expand patient volume. Further evidence that the rurals could not significantly expand their utilization rate is that Mercy and Finley do not consider the rurals as competitors. Both Mercy and Finley look to the rurals as a source of referrals for the services that the rurals do not provide and try to cooperate with the rurals in order to capture some of this referral business. While the rurals could conceivably capture some patients should the managed care payers give incentives for their use, it does not appear to be a large percentage of the current population using Mercy and Finley. 2. 5% Price Increase The government also argues that there are no competitors who could defeat a 5% price increase by DRHS and that therefore, the merger would result in an anticompetitive effect. The 1984 Merger Guidelines instituted the 5% test — suggesting that if a merged entity could sustain a 5% price rise for approximately one year, the merger should be enjoined as a violation of the antitrust laws. The government concluded that a 5% price increase would be profitable, in large part, due to the following assumptions: (1) people have strong loyalties to their doctors and will not switch hospitals as this would entail switching doctors, (2) the outreach clinics only influence behavior in the “fringe areas” between two regional hospitals, and (3) persons within twenty-five miles of Dubuque will only use Dubuque hospitals. Both parties agreed that for a 5% price increase to be unprofitable, DRHS would have to lose 8% of its present population. The government asserts that this will not happen as (1) at most, 4% of those within twenty-five miles of Dubuque would shift to a non-Dubuque hospital, and (2) at most, 3.2% of the 13% of Dubuque patients living outside this 25 mile radius might shift. The government argues that those within twenty-five miles of Dubuque would not shift as they would not travel outside of Dubuque for health care needs partly because they would find it inconvenient, and partly because a shift would necessitate changing physicians which most would be unwilling to do. The government broke down the 13% of Mercy and Finley’s patients who reside outside the 25 mile radius into categories and found that the following persons would be unlikely to switch in the event of a 5% price rise: (1) the .6% admitted who reside completely outside the area (e.g. California), (2) the 1.2% who are admitted through the emergency room whom the government assumes happened to be in the Dubuque area when their emergency arose, (3) the 1.1% who are admitted by Medical Associates physicians because these patients will not change hospitals due to their loyalty to their physicians and because their physicians only have privileges at Mercy and Finley, and (4) the 3.6% who were admitted by their Dubuque primary care physician, who the government assumes will not switch because they have ties to Dubuque and that particular doctor. The government asserts of the remaining 6.5% who do not fall in any of the above categories, only half, or 3.2% would switch. The court finds the government’s numerical conclusions to be incorrect, again due to the government’s assumption of strong patient-physician loyalty and the government’s erroneous belief that persons within twenty-five miles of Dubuque will not travel outside of Dubuque for hospital care. Adding in the percentages which the government excluded due to these assumptions puts the total of those likely to switch in the event of a 5% price rise at a number higher than the 8% necessary to make the price rise unprofitable. 3. Elimination of Managed Care Discounts The government contends the more likely scenario is that DRHS would reduce or eliminate all current discounts to managed care entities resulting in a 15-30% increase in prices. The evidence was that such a price increase would have to be countered by a 20-35% loss of patients to make such a price increase unprofitable. The government asserts that with a total abandonment of managed care discounts, to defeat a price rise, DRHS would have to lose all of its patients residing outside of the government’s proposed market area (24%) and perhaps a number of those residing within the market area. The government argues that it is inconceivable that such a large exodus from the Dubuque hospitals would occur. The government presumably relies on its assumptions that persons within twenty-five miles of Dubuque will not shift to other facilities, that outreach clinics are largely irrelevant to the analysis, and that there is strong doctor-patient loyalty preventing shifting. For the reasons stated above, the court has already found that these assumptions are erroneous. The government continues to fail to look at the merger within the context of current market trends. All evidence is that there is a great deal of competition for health care dollars — including stiff competition between the managed care entities. As hospitalization costs account for approximately 40% of an HMO’s fee, a 15-30% price increase in hospitalization rates would equate to a 6-12% increase in the cost of an HMO. HMOs which could avoid this increase would be much more cost competitive and presumably would be able to attract more enrollees. Therefore, if DRHS acted in a noncompetitive manner, an HMO that could successfully induce Dubuque area residents to use alternative hospitals would be at a significant cost advantage. The government asserts that those within twenty-five miles of Dubuque cannot be induced to travel outside of Dubuque for health care; however, the court finds to the contrary for several reasons: (1) a significant portion of the population already chooses to go outside Dubuque for health care, (2) managed care entities have successfully shifted patients in the past, and (3) managed care enrollees are increasingly willing to travel for financial savings. Turning to the first point, already, a significant number of Dubuque residents use the University of Iowa Hospitals and Clinics (“UIHC”), a ninety mile drive, due to quality concerns. The government asserts that these persons are only going to the University of Iowa for health care that cannot be obtained in Dubuque, and supports this assertion by testimony from doctors stating that they only send patients to the UIHC for care unavailable in Dubuque. However, 63% of Dubuque residents using the UIHC are self-referrals who are not going to the UIHC on the referral of a Dubuque doctors. Further, the zip code data shows a widely varied use of the UIHC across zip codes, indicative of the defendants’ position that persons are not going only for care unavailable in Du-buque. Presumably, if persons were only using the UIHC for care unavailable in Du-buque, the percentages of persons going to the UIHC from each zip code would be fairly equal. There was also testimony that a plastic surgeon from Cedar Rapids successfully induces Dubuque residents to travel to Cedar Rapids for plastic surgery by offering them lower prices on the surgery and hospitalization. Plastic surgery is not often covered by private health insurance and therefore persons wanting plastic surgery have to pay for the full cost out of their own pocket. A final example of persons traveling outside of Dubuque.for health care needs is that of a Dubuque physician who successfully directs cardiac patients needing catheterization, angioplasty and open heart surgery to Iowa City. Although this is reportedly done for political reasons and not for price considerations, it nevertheless does show that persons can be induced to travel from the Du-buque area for inpatient care available in Dubuque. There is also evidence that managed care entities can successfully induce Dubuque residents to use other regional hospitals for their inpatient needs. Heritage shifted patients in need of open heart surgery and substance abuse treatment from Dubuque to Waterloo and Davenport in a successful attempt to obtain lower prices from Mercy in these service areas. The government vehemently argues that this evidence is irrelevant as these are services outside the product market and that there is no evidence that people who would switch for these services would switch for the services within the relevant product market. The court finds, however, that it may rely on this evidence for two reasons: (1) the evidence is that it is generally harder to induce people to travel for the longer term hospital care necessitated by these services than for the type of hospital care within the relevant product market, and (2) there is no evidence that patients distinguish these types of care when considering whether to travel for hospital care. The main basis for the government’s assertion that people inside Dubuque will not travel for hospital care is that they have not done so in the past, they are not willing to go to the extra expense of having to travel and put their families through the expense and inconvenience of having to stay in a hotel to be near them during their hospitalization, and they do not want to switch physicians. Persons traveling for substance abuse treatment and open heart surgery would apparently have these same concerns, yet they are traveling. Finally, the court turns to the third point — that financial incentives can induce people to travel. Many witnesses testified that employees are more willing to travel now that they are covering a greater portion of their health care costs. Survey results indicate that, for less than $1000.00, Dubuque residents could be induced to travel to another hospital for care. The government disputes the reliability of the surveys, but the survey outcomes are confirmed by the testimony of officers within the managed care corporations and employers purchasing health care plans. For example, Doug Sesterhahn testified that SISCO has induced enrollees to travel an hour to an hour-and-a-half for health care by providing financial incentives. Mr. Sest-erhahn also testified that he felt SISCO could provide incentives that would induce Du-buque employees to go to Platteville and Southwest Medical Center for care. James Thompson felt that for some categories of patients — nonsurgical, cancer care and heart care patients — a $200 incentive would be enough to induce some Dubuque residents to travel to another regional center for their care. While he was unsure as to the numbers of persons who would travel for a $200 inducement, Mr. Thompson also testified that, with financial inducements of between $500-1000, Heritage could shift patients from inside Dubuque to the regional hospitals for many services in the relevant product market. The average full charge for inpatient hospitalization at Finley and Mercy is approximately $6,000. Thus, the current 20% discount to managed care payers averages $1,200. If that discount was eliminated at DRHS but available at one of the regional hospitals, the managed care payers would be able to offer a very significant financial incentive in order to get patients to travel. The managed care payers could include a Dubuque hospital in their health care package but also impose different co-pay requirements between the DRHS hospital and regional hospitals so as to induce travel by enough patients to defeat the price increase. The court must also consider the present trend in hospitalizations toward shorter and fewer hospital stays requiring hospitals to enter into intense competitions for an increased population base. The court finds it relevant that the regional hospitals are in an intense competition to increase the population base from which they draw patients due to the trend in shorter hospital stays and less frequent stays. Mercy has recognized, beginning with its long range plan of 1990, that, to maintain its average daily patient census, it is dependent on increasing its base of patients residing beyond Dubuque County. DRHS would be under similar pressures. The current state of hospital competition would require that DRHS strongly attempt to expand the use of its hospital by those persons outside Mercy and Finley’s traditional catchment areas. If DRHS raised its prices so as to become noncompetitive with other institutions, it would not be able to expand beyond its traditional service area. The court also notes that 23.7% of Mercy and Finley’s patients currently reside in a zip code in which 33% of the residents within that zip code currently use a hospital other than Mercy or Finley. In addition there are an additional 5.9% of Mercy and Finley’s patients coming from other zip codes which lie within 15 miles of a regional outreach location. The government makes no attempt to state why these individuals would not be able to switch hospitals in the event of a 15-30% price rise. 4. City of Dubuque as the Relevant Market The government has offered an alternative geographic market. It contends that the City of Dubuque is a geographic market within which DRHS would be able to exercise market power because Dubuque residents will not travel to the rurals or other regional hospitals for hospital care that is available in Dubuque. While this is a much smaller area than DRHS would service, it may be the relevant market if it is a significant market area and DRHS could successfully exert market power over this group. See, Rebel Oil Co. v. Atlantic Richfield Co., 51 F.3d 1421, 1434 (9th Cir.1995) (citing Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 518.1b, at 534 (Supp.1993) (geographic market can be shown by showing resulting market power over any group of buyers)). To be the relevant market, there must not be any other supplier of inpatient services who could constrain the price raising capability of DRHS to purchasers of inpatient services who reside within the City of Dubuque. See Phillip E. Areeda & Herbert Hovenkamp ¶518.^1, at 538 (Supp.1993). The validity of a horizontal merger must be determined based on the merger’s “effect on competition generally in an economically significant market.” Brown Shoe Co., 370 U.S. at 335, 82 S.Ct. at 1529. How large a market must be to be economically significant has never been fully addressed. See 2A Phillip E. Areeda & Herbert Hovenkamp ¶ 530e, at 154 (1995). However, the Supreme Court in Brown Shoe Co. found that, under some circumstances the relevant geographic area may be as small as a single metropolitan area. Brown Shoe Co., 370 U.S. at 337, 82 S.Ct. at 1530. The court went on to explain: The fact that two merging firms have competed directly on the horizontal level in but a fraction of the geographic markets in which either has operated, does not, in itself, place their merger outside the scope of § 7. That section speaks of “any ... section of the country,” and if anticompeti-tive effects of a merger are probable in “any” significant market, the merger — at least to that extent — is proscribed. Id., at 337, 82 S.Ct. at 1530. The issue is “where, within the area of competitive overlap, the effect of the merger on competition will be direct and immediate.” Philadelphia National Bank, 374 U.S. at 357, 83 S.Ct. at 1738. For purposes of this analysis, the court will assume Dubuque is large enough to be an economically significant geographic area for purposes of the antitrust laws. The government’s argument that hospital services are local and that persons residing within Dubuque will not go elsewhere for hospital services is largely based upon the government’s perception of inpatients’ desires for the convenience of a local hospitalization for themselves and for their family members who will be visiting them. The government analogizes the present case with that of Philadelphia National Bank. The government contends that hospital services, like the banking services at issue in Philadelphia National Bank are local in nature and that people want the convenience of a local hospital. Id. at 358, 83 S.Ct. at 1739 (“The factor of inconvenience localizes banking competition as effectively as high transportation costs in other industries.”). The government next argues that this geographical region is relevant as DRHS could raise prices to all managed care enrollees within the immediate Dubuque area while providing more favorable pricing to those persons that might possibly switch to other hospitals. It is the government’s contention that managed care entities cannot get residents of Dubuque to use hospitals outside of Dubuque, and therefore, the merger would substantially lessen competition to those persons residing within the immediate Dubuque area. The government relies heavily on its assertion that residents of Dubuque will not travel for medical care available in Dubuque. However, the court finds, as stated above, that Dubuque residents have in the past traveled for their health care needs and would travel in the future should DRHS prices become noncompetitive. Much was made of the “Ottumwa Situation” — in that the government argued that the two hospitals in Ottumwa, Iowa were allowed to merge and upon merger, promptly eliminated the discounts to the managed care entities in the area. While the court does not find that there was sufficient evidence presented on the Ottumwa merger to determine how closely the facts behind the Ottum-wa merger reflect those of DRHS, the court does note significant differences — the regional hospitals were not targeting areas around Ottumwa for expansion, to this date there has been little outreach activity in the Ottum-wa area, and Ottumwa is further from other regional hospitals than is Dubuque. The court does recognize that other courts have discounted the ability of third party payers to prevent price rises. See University Health, Inc., 938 F.2d at 1213; Hospital Corp. of America, 807 F.2d at 1391. These courts found that large insurers who “purchase” hospital services on behalf of their insured merely pass any increased cost on to the consumer. In University Health Inc., the court found that the third party payer could not refuse to reimburse its insured because it deemed the price of the hospital services was too high. University Health Inc., 938 F.2d at 1213. However, these cases did not focus on managed care entities, but focused instead on traditional indemnity insurers. In the case at hand, the defendants have shown that the third party payers of interest in this matter — the managed care entities — can cause their enrollees to shift to other hospitals if prices at one hospital are too high, thus avoiding having to pay the higher price and providing strong incentive for the anticompetitive hospital to lower its charges. Further, the competition has become fierce between HMO’s. If an HMO were merely to pass on increased hospitalization costs to en-rollees, it will lose enrollees to plans that offer a cheaper rate. This pattern has been evidenced in Platteville, Wisconsin, an area where Medical Associates HMO has met strong competition from competing plans. Even assuming Dubuque residents would not shift in the event of a price increase by DRHS, the government has also failed to establish that DRHS could successfully initiate a price discrimination program. Mr. Thompson testified in general terms that it would be possible for the hospitals to price discriminate on the basis of where a patient resided, but no one established the mechanics of how this would work. Generally, managed care payers contract with the hospitals at one price, but it is conceivable that they could either 1) establish different prices according to which employer the inpatient was insured through, or 2) establish different prices a