|
|
||||||
|
01/2006 |
||||||
|
John J. Miles Appeared in Antitrust Health Care Chronicle I. Introduction Few Federal Trade Commission ("FTC") initial decisions have generated as much interest among health-care and antitrust attorneys as FTC Chief Administrative Law Judge ("ALJ") Stephen J. McGuire's October 20, 2005 decision in Evanston Northwestern Healthcare Corporation ("ENH").1 That decision held that Evanston Northwestern Healthcare's ("ENH") January 2000 acquisition of Highland Park Hospital in the north-Chicago area violated Section 7 of the Clayton Act2 and orders ENH to divest the facility. Both sides have filed notices of appeal to the full FTC, although the Seventh Circuit will probably make the ultimate decision. The initial decision itself, however, (and even anticipation of the initial decision) has already generated articles3 and seminars. 4 The stakes in the case are high. Evanston Northwestern could lose about 28 percent of its hospital capacity as a result of divestiture. It states that it has already spent over $20 million defending the case. Nineteen FTC attorneys were on complaint counsel's post-trial brief, and ENH's brief listed 15 counsel. More important, the ultimate decision in the case, probably several years away, will likely have major implications for hospital-merger enforcement in the future. Keen interest in the initial decision is understandable for that reason and for other reasons as well:
II. Background Prior to the merger, Evanston, Highland Park's parent, and two other hospitals were part of the Northwestern Healthcare Network ("NHN"); NHN was the hospitals' sole member. That network was a loosely knit organization in the sense that there appeared to be very little integration among the hospitals through NHN. The hospitals operated autonomously, with NHN apparently serving as little more than a contracting network for the hospitals' risk contracts. The network did not direct the hospitals' operational activities; the hospitals dissolved NHN the day after the merger. Sometime after the FTC announced its hospital-merger retrospective-review program in August 2002, FTC staff began investigating the ENH merger (and several other consummated hospital mergers that were not, or have not been, challenged). Complaint counsel filed its complaint in February 2004, challenging the merger in two counts that alleged the merger was unlawful under two separate legal and economic theories. 15 Count I alleged a traditional Section 7 claim, delineating an alleged relevant market and claiming that the merger substantially increased concentration in that market. Count II, however, was a non-traditional — indeed unique — type of challenge, at least in a merger case. Because the merger had occurred over four years before, there was a track record of ENH's post-merger pricing and other market factors. According to the complaint, ENH, as a result of the merger, obtained sufficient market power to "raise[] prices more than the price increases implemented by other comparable hospitals" and "raise its prices to private payers above the prices that the hospitals would have charged absent the merger." 16 Count II alleged no relevant market, no market concentration figures, and no combined post-merger market share. Rather, complaint counsel alleged that the merger could be condemned under Section 7 based solely on ENH's post-merger pricing behavior. The FTC's administrative hearing began in February 2005 and lasted eight weeks. Forty-two witnesses testified in person, and approximately 1,600 exhibits were introduced into evidence. 17 The ALJ issued his opinion on October 20, 2005, and, as noted before, both sides have filed notices of appeal. In attempting to prove that the ENH merger created or enhanced ENH's market power as alleged in Count I, complaint counsel put on what can be called a "structural," or traditional circumstantial-evidence Section 7 case, introducing evidence of market definition, concentration statistics, and post-merger market share. In attempting to prove the same conclusion pursuant to its Count II theory, complaint counsel introduced what can be called a "performance" or direct-evidence case of ENH's post-merger market power — its post-merger pricing patterns — arguing that regardless of the structural variables, ENH, in fact, raised its prices by significantly more than comparable hospitals because of market power gained through the merger. Proof under either theory, complaint counsel argued, was sufficient to prove a violation of Section 7. III. The ALJ's Structural Analyis18 A. Relevant Product Market ENH, on the other hand, argued that outpatient services should be included in the product market for two reasons: First, payers contract for inpatient and outpatient services together, and thus those services should be "clustered" into a single market for purposes of antitrust analysis. Second, prices for outpatient services affect prices for inpatient services because payers often trade-off higher prices for one for lower prices for the other — that all the health plan cares about is the aggregate cost of the services it purchases from the hospital. The ALJ did not discuss ENH's arguments in great detail, relying rather on the simple fact that inpatient and outpatient services are not substitutes in use. The cluster-market argument would seem to have little merit here in any event because, where the facts dictate, the cluster of different services can be separated into separate product markets for purposes of antitrust analysis. For example, a merger of two hospitals may present little competitive problem with respect to all hospital services but might merge the only two obstetric services in the relevant geographic market. In that situation, the relevant product market of concern would be limited to obstetric services, even if payers purchase them as a package with other hospital services. 21 ENH's second argument — the price-trade-off argument between inpatient and outpatient services — deserves some thought, but the ALJ handled it easily on the facts, noting an absence of evidence that ENH's payers had, in fact, engaged in that type of trade-off. 22 Even had there been such evidence, it would seem difficult to conclude that outpatient services constrained the price of inpatient services in the relevant sense of a payer's having the ability to circumvent a price increase for inpatient services by switching its purchases to outpatient services. 23 Under the Merger Guidelines, demand substitutability is the lodestar for defining a relevant market. 24 B. Relevant Geographic Market In previous hospital-merger decisions lost by the agencies, the courts tended to define geographic markets broadly. 25 They examined, primarily, patient-flow data, looking first at where patients actually went for hospital services and then where, if the merging hospitals (in a unilateral-effects case) or all the hospitals in the tentative market (in a coordinated-interaction case) attempted to raise price, patients could go to circumvent the price increase. Then, typically applying critical-loss analysis, 26 the courts attempted to determine whether, if the hospitals increased price, they would lose so many patients to more distant hospitals that the price increase would be unprofitable. If so, the next-closest hospital would be added to the tentative market and the test repeated. The process would be reiterated until the hypothetical price increase just became profitable because too little further patient substitution of more distant facilities would occur to make the increase unprofitable. At that point, the increase in profits from the price increase to remaining patients just exceeds the loss in profits from the loss of patients who would patronize other facilities because of the price increase. Commentators and the agencies have stringently criticized the heavy reliance by courts on patient-flow data because, they argue, it leads to unduly broad relevant geographic markets; it falsely suggests that a hospital's patients are willing to switch to other, more distant hospitals should the hospital in question increase its prices. 27 Accordingly, it appears that complaint counsel theorized the ENH case in a way that would wean the ALJ from that reliance. Complaint counsel posited a two-stage hospital-competition model. 28 Succinctly stated, in the first stage ("Stage 1") — the one at issue here — hospitals compete to participate in the health plans' networks, and the plans engage in selective contracting to constrain the hospitals' market power. As the ALJ put it, "The ability of the managed care organization to exclude hospitals from its network is a powerful tool that defines each side's bargaining position." 29 The crucial point is that the hospital's customers at the Stage 1 level of competition are health plans, not hospital patients, because the health plans rather than patient set and pay the price for the hospital's services. 30 Competition among hospitals in Stage 1 is largely based on price. The second stage of competition, occurring after the health plan has chosen the providers for its network, is competition among the participating hospitals for the plan members' patronage ("Stage 2"). At this stage of competition, the patient is the hospital's customer, and competition occurs along nonprice dimensions such as quality and amenities. To shorten a long story, complaint counsel's argument, and the ALJ's finding, was that if, in Stage 1 competition, the health plan rather than the patient is the customer, patient-flow data identifying the hospitals that patients use and showing the hospitals that they could use to escape a price increase by the merging hospitals are irrelevant. This would mean, in particular, that the Elzinga-Hogarty test for defining geographic markets based on patient in-migration into and out-migration out of a tentative market is inappropriate. 31 Plans, not patients, are price sensitive, and so a price increase by the merged hospitals would not induce their patients to travel to a different facility for services. Rather, the analyst has to examine what the plan could do to circumvent the price increase. For example, the plan might attempt to induce or steer its members to use other more distant hospitals by providing financial incentives (e.g., lower co-payments and deductibles) in the plan's benefits. To bolster its position that patient-data information is worthless, complaint counsel retained Dr. Kenneth Elzinga, an economist and a creator of the Elzinga-Hogarty test. He testified, and the ALJ found, that use of the Elzinga-Hogarty test — and more generally, any test relying on patient-flow data — is not helpful in defining relevant geographic markets for hospitals in Stage 1 analysis for three basic reasons. 32 First, the test is never appropriate in markets for differentiated services rather than homogenous services. 33 Second, there is a so-called "payor problem" in using such data. As just noted, although patients consume the hospital services, health plans pay for them. Since patients do not pay for the services, how they may travel for services provides no aid in determining how hospital customers will react to a price increase because patient selection of hospitals is not a function of hospital prices or price increases. 34 Third, Dr. Elzinga posited a "silent majority problem"— i.e., that patient-flow analysis incorrectly assumes that if some patients are willing to travel to a more distant hospital to escape a price increase by the merging hospitals, other patients will do so as well. (In fact, this is an important assumption underlying the relevance of identifying the "contestable zip codes" when doing a critical-loss analysis. 35 ) But, according to Elzinga, there is a "silent majority" of patients who would not travel in light of the price increase, and the merged hospitals could raise their prices as to them. 36 It is not clear that the payor and silent-majority problems lead to the conclusion that patient-flow data is irrelevant in defining geographic markets. The silent-majority argument assumes one of two things: either (1) the hospital can price discriminate by charging silent-majority patients a higher price that it charges the rest of its patients; or (2) the greater profits from the price increase to silent-majority patients exceed the loss in profits resulting from other patients switching to more distant hospitals. As to the first possibility, the hospital would raise its prices to silent-majority patients ("inframarginal" patients in economics lingo), while not increasing prices to patients that might switch ("marginal" patients) so they would have no inducement to switch. If the hospital cannot discriminate in price against its inframarginal patients and attempts to increase prices to both groups, a sufficient number of marginal patients might switch from the hospital in question to a more distant hospital so that the price increase would be unprofitable and the hospital would have to rescind it. In this way, absent price discrimination, the marginal patients protect the inframarginal patients from the price increase. It would seem that the necessary price discrimination by the hospital would be very difficult if not impossible. The hospital, for example, would have to charge a health plan different prices for different enrollees, depending on their place of residence, even though many marginal and inframarginal patients likely would be in the same health plan. This seems a most unlikely scenario. If it did occur, however, the relevant market would not include sales to the marginal patients but would be limited to sales of hospital services to inframarginal patients. 37 Patient-flow information would be helpful in determining the marginal and inframarginal patients. The second possibility is that the hospital does not price discriminate but that the price increase is profitable because too few patients are marginal and choose a more distant hospital. The profit-enhancing effect of the price increase is greater than the profits lost from patients switching to more distant hospitals. But to determine whether this is the likely result of the price increase, patient-flow data is important to help estimate the likely number of patients continuing to use the hospital and the likely number switching to more distant hospitals. The payor problem would seem more valid, but it does not render patient-flow completely irrelevant. It is true that, since the plan is paying for the services, the analysis should focus, at least primarily, on the plan's reaction to the price increase rather than that of patients. But how a plan reacts to a price increase (for example, whether it seeks out more distant hospitals for its members and whether those hospitals would be adequate substitutes) depends importantly on its members' hospital preferences: "managed care organizations must take patient preferences into consideration in constructing their networks." 38 And if the plan believes that its members would be satisfied without the merging hospitals but with a more distant alternative — and patient-flow information would seem important in making this determination — then it can substitute the alternative. The ALJ noted that the plans' demand for hospitals is a "derived demand," derived from its members' preferences. 39 Patient-flow data should be a substantial aid in determining those preferences by showing which hospitals subscribers are using and which hospitals they could use as a substitute if the merged hospital raised its prices to the plan. Related to this, the plan may be able to steer its members to more distant hospitals (that is, increase their incentive to travel) through incentives in plan benefit-packages — tiered benefits, for example, by which health-plan members pay more out of pocket for using hospitals that cost the plan more than others. Again, patient-flow data helps identify those hospitals by indicating hospitals that might be reasonable alternatives (and thus potentially in the relevant geographic market). In fact, one would expect health plans to conduct much the same types of analyses regarding patient willingness to travel if their cost were reduced that courts have approved in defining relevant geographic markets. In ENH, however, the ALJ found that little steerage by payers occurred in the Chicago area, although it is not clear how hard payers tried to steer and why they were unsuccessful40 — both important questions. The bottom line seems to be that patient-flow data is not irrelevant in defining geographic markets in hospital-merger cases even under the assumption that health plans rather than patients are the hospital's customers. This is because just as patient-flow information can help identify alternatives that patients would choose if they were price sensitive, it serves the same function in identifying facilities with which health plans might contract in light of a price increase. It may be true that some of the previous hospital-merger decisions overemphasized its importance in the analysis, and certainly it is just one of several factors a court should consider. Significantly, in ENH, ENH's expert relied on several factors in addition to patient-flow information, and she did not rely on the Elzinga-Hogarty test at all. 41 One of the most interesting issues raised by the ENH decision is whether the relevant geographic market should have been limited to the three ENH hospitals as complaint counsel argued. There were apparently some 46 hospitals within 20 miles of ENH, and there was abundant testimony that the ENH hospitals competed with other hospitals. Thus, complaint counsel's argument would appear ridiculous — until one considers the ALJ's findings on ENH's price increases together with the geographic-market definition methodology of the Merger Guidelines. 42 Under the "hypothetical monopolist" framework of the Merger Guidelines, 43 the analyst starts with the merging hospitals by themselves and hypothesizes that one or both raises price a small but significant amount as a result of the merger. If the price increase would be unprofitable because a sufficient number of patients would switch to more distant facilities, the tentative geographic market must be expanded to include the next-best substitute. But if not — if the merged hospital's price increase would be profitable — those hospitals, by themselves, are the only firms in the relevant geographic market. 44 The key point is that the relevant geographic market includes only the merged entity and other sellers, if any, necessary for the sellers (through coordinated pricing actions) to raise price profitably. It does not include all firms that may have some constraining effect on the merged entity's pricing behavior — i.e. , all firms with which it competes. In ENH, accepting the ALJ's conclusions about ENH's post-merger price increases (i.e., that they were significant, resulted from the increase in market power resulting from the merger rather than from other causes, and were profitable), it would appear that complaint counsel was right: the relevant geographic market would include only the ENH hospitals. This would not mean that ENH was under no pricing constraints from other hospitals or that it was a monopoly or had monopoly power, as complaint counsel seemed to argue45 and the ALJ claimed would be true if he accepted complaint counsel's argument. 46 Rather, it would mean that the competitive constraints on ENH from other hospitals were insufficient to prevent it from exercising significant market power. The ALJ included the ENH hospitals and four other hospitals in the market. But under Merger Guidelines analysis, this would mean that coordinated interaction among seven hospitals was necessary for the exercise of market power and thus that ENH, post-merger, could not exercise market power by itself — i.e. , could not profitably raise its prices a significant amount. The ALJ's analysis seems internally inconsistent. According to the ALJ, the relevant factors to consider in defining a relevant geographic market are the views of market participants, the geographic proximity of different area hospitals to one another, patient travel times to hospitals, and the degree to which the hospital medical staffs overlap. Based on testimony from some (but not all) payers that they needed ENH in their networks, the ALJ (at least at first) concluded that payers could not establish viable networks without ENH47 (which, if true, suggests that the ENH hospitals, by themselves, constitute a relevant market). There seems to be an important inconsistency in the opinion, however, because the ALJ states later in the opinion, when rejecting complaint counsel's argument that the market should be limited to the three ENH hospitals, that "[i]t is highly probable that the four non-ENH hospitals in the geographic market would have the ability to constrain prices at ENH, either now or in the future, and could be utilized by managed care organizations to create alternate hospital networks" 48 (which, if true, suggests that the ENH hospitals do not constitute a relevant market and do not have market power). Courts in some previous hospital-merger decisions have discounted (perhaps too much) the testimony of market participants, particularly payers, 49 but the payer testimony in ENH seemed rather subjective. For example, the ALJ noted that "United believes it cannot satisfy its customers without ENH"; PHCS "believed . . . that customers did not want to ‘buy the network if they did not have [ENH in] it'"; "Aetna believed it ‘couldn't walk away' from post-merger ENH because it would have ‘devastated us,' ‘killed our marketing' and ‘shut down' Aetna's marketing to local employers"; and "Unicare would be in a bind without ENH." 50 Perhaps these health plans offered specific facts and bases to support these conclusions at the hearing; if not, the ALJ should have probed them or discounted their testimony. 51 Representatives of One Health testified, according to the ALJ, that "customers complained about not having access to ENH" and thus that it "returned to ENH prepared to accede ‘essentially regardless of what the ultimate price was,'" although even that health plan "pointed to [other hospitals] as substitutes." Apparently One Health actually lost members because ENH terminated it over price — probative evidence, the weight of which would depend on the size of the membership loss and its effect on the plan's viability. 52 Surely, payer testimony merely that "I believe I need that hospital in my network; other hospitals aren't good substitutes," without supporting detailed facts, is not sufficient for a court to conclude that there are no substitutes for "that hospital." The payer should present facts and grounds for its opinions and evidence about the degree of its need for the hospital in question. 53 Next, the ALJ appeared to inappropriately conflate the examination of the geographic proximity of area hospitals to the ENH hospitals and patient travel times to hospitals. Relying on a simple, seemingly unscientific customer-survey report by one hospital in the area indicating that patients are willing to travel up to 16 minutes for emergency care and up to 35 minutes for an overnight stay in a hospital, the ALJ presumed that patients would select a health plan that included a local hospital, "ideally one within 16 minutes of their home or work." 54 This study seems like an extremely thin evidentiary reed to begin with, but, in addition, why the ALJ selected the 16-minute figure rather than the 35-minute figure for his analysis is far from clear. 55 Then, he appeared to include in the relevant market any hospital that was around 20-minutes driving time from one of the ENH hospitals. Apparently, he saw some correlation between patient travel time to a hospital and the proximity of the ENH hospitals to other hospitals. Based on proximity to the ENH hospitals, the ALJ selected, for inclusion in the relevant market, four hospitals in addition to the ENH hospitals: (1) Lake Forest, 13 minutes from Highland Park; (2) Advocate Lutheran General, 21 minutes from Evanston; (3) Rush North Shore, nine minutes from Evanston; and (4) St. Francis, eight minutes from Evanston. He excluded two hospitals that ENH's expert claimed should be included: (1) Condell, 24 minutes from Highland Park; 56 and (2) Resurrection, 25 minutes from Evanston. He also excluded three other hospitals that were 23, 19, and 30 minutes from an ENH hospital. It seems far from clear what this type of analysis says about the scope of a relevant geographic market. Presumably, the closer other hospitals are to ENH hospitals, the more substitutable they would be for ENH hospitals, but the degree of substitutability, which is the important consideration, would depend on a plethora of other characteristics as well, especially in a market that both sides agreed is a differentiated services market — e.g., health-plan member and ENH-patient spatial dispersion throughout the area, hospital preferences, and willingness to travel to more distant hospitals if offered financial incentives to do so. It would seem that the ALJ should have considered the geographical proximity of area hospitals to ENH patients — another reason why patient-flow data would seem relevant. In sum, the discussion of the relevant geographic market is not particularly satisfying. Patient-flow data, particularly to the extent it suggests patient preference and willingness to travel, would seem a relevant consideration, although certainly not a dispositive one, 57 even when viewing health plans, rather than patients, as the hospital's customers in a Stage 1 competitive analysis. Moreover, the opinion seems to contradict itself on the question of whether ENH was a "must have" system for a viable health plan. On the other hand, the inclusion of hospitals other than ENH in the relevant geographic market seems inconsistent with the findings that ENH was able to increase its prices significantly because of market power gained through the merger. C. The ALJ's Competitive-Effects Structural Analysis The ALJ, curiously, appeared to place most reliance on the HHI figures — that is, the degree of market concentration — rather than on ENH's post-merger market share and importantly, as discussed later, whether Evanston and Highland Park were each other's most direct competitor. The heavy reliance on market concentration is curious because the competitive-effects concern from this merger was not coordinated interaction — i.e., concern that, as the market became more concentrated, it became more likely that ENH and its competitors would price interdependently as an oligopoly. Rather, the concern was that ENH would be able to exercise market power (in fact, actually had exercised market power) itself by unilaterally increasing its prices. The connection between the level of market concentration and the ability of a firm unilaterally to exercise market power is not clear. 61 Indeed, the ALJ's entire structural-analysis discussion is extremely brief (two-and-a-half pages of a 225-page opinion), even though this analysis was the primary basis for the opinion's conclusion that the merger was unlawful; according to the ALJ, the performance analysis only "confirmed" the conclusion of the structural analysis. 62 The ALJ merely concluded that, based on the market-concentration standards of the Merger Guidelines, 63 the high HHI statistics were sufficient to show that the merger was likely to result in market power and that ENH's post-merger market share met the test of the Supreme Court's 1962 Philadelphia National Bank decision, in which the Court indicated that a merger resulting in a firm with a 30 percent or higher market share would be highly suspect. 64 This analysis is fuzzy at best. Finally worth noting is that, with one exception, all the hospital-merger challenges litigated by the agencies up to ENH involved mergers resulting in significantly higher post-merger market shares than 40 percent (usually in the 60 to 80 percent range) and post-merger HHIs significantly above the Merger Guidelines' standard for a highly concentrated market. 65 Thus, one question that the ENH decision might raise is whether the lower statistics leading to a conclusion that the ENH merger was unlawful will energize the agencies to attack aggressively hospital-mergers resulting in substantially lower market shares than those challenged in the past. It seems unlikely that either federal enforcement agency would have attempted to block the ENH transaction based on the post-merger market share alone. The evidence in ENH, however, showing that a hospital merger resulting in a 40 percent market share permitted the merging hospitals to raise prices significantly may convince the agencies that they have been too conservative in not challenging hospital mergers with post-merger shares in the 35 to 40 percent range. This, and the fact that many hospital markets are already highly concentrated under the Merger Guidelines standards, suggests that should the initial decision be affirmed, the number of hospital mergers challenged may increase significantly. IV. The ALJ's Performance-Evidence Analysis The ENH opinion contains an extended discussion, especially when compared to its structural-analysis discussion, of the experts' testimony concerning ENH's post-merger price increases compared to the price increases of other Chicago-area hospitals at the same time. In a sense, much of this discussion seems moot since ENH's own expert conceded that ENH's price increases outstripped those of the hospitals placed into comparable control-groups for price-increase comparison purposes. 68 Rather, the question with regard to the pricing evidence was whether the differences between ENH's price increases and those of the control-group hospitals resulted from market power generated by the merger as complaint counsel argued, or from other benign or procompetitive factors as ENH claimed — its learning that Highland Park was supposedly obtaining better prices than it had prior to the merger or the alleged increases in the quality it provided as a result of the merger, as discussed below. Both sides and the ALJ agreed that a firm's large post-merger price increases and the fact that its increases were significantly higher than those of comparison hospitals, by themselves, are not sufficient to show market power69 because, in a differentiated services market, firms are expected to charge different prices. There may be different demands for their services, and they may incur different costs. 70 Because of the different non-price attributes of differentiated services, some consumers will favor one provider over others and be willing to pay more for that provider's services. Accordingly, the ALJ explained, the appropriate methodology compares the price changes over time between the hospital under examination and appropriate control groups of comparable hospitals, controlling for factors other than market power that might explain the differences. 71 The ALJ described this approach as the "differences in differences" methodology. After selecting a control group of comparable hospitals, step one in the analysis was to calculate the "ENH difference" — i.e. , the difference between the ENH premerger and post-merger prices. Step two was to compute the "control group difference" — i.e. , the price differences at the control-group hospitals for the same time period. The final step was to calculate each difference as a percentage of each hospital's premerger price and compare the ENH difference with the control-group hospital differences. 72 If the ENH difference was significantly greater (controlling for other factors), it would be reasonable to assume that the differences resulted from greater ENH market power. ENH's expert, based on his own analysis, concluded that ENH's non-quality adjusted price increases were significantly (nine to ten percent) greater than those at the control-group hospitals. 73 His emphasizing the non-quality-adjusted nature of the price differences left ENH with room to argue that its price increases resulted from increases in quality and thus that there were really no differences between its price increases and those of the control-group hospitals. The ALJ rejected that argument, as noted later, on the ground that there was no showing that ENH's quality increased more than that at the control-group hospitals. Because market power was not the only possible explanation for ENH's greater price increases, the parties agreed that for the pricing comparison to be valid, the experts had to control for factors other than market power resulting from the merger that might explain why the ENH difference was greater than the control-group differences. Thus, complaint counsel's expert controlled for some eight variables that she believed could explain the price differences: (1) cost increases affecting all hospitals; (2) changes in regulations affecting all hospitals; (3) increases in consumer demand for hospital services; (4) ENH's claimed increase in quality; (5) changes in the hospitals' payor mix; (6) changes in the hospitals' mix of customers; (7) changes in teaching intensity at the hospitals; and (8) decreases in outpatient prices. 74 It is not clear whether the expert considered that some of ENH's contracts were quite old when the merger was consummated in 2000; some, for example, had not been renegotiated since 1995 or 1996, and thus it would seem that ENH's making up for lost time also could be an explanation for its price-increase differences outstripping those of the control-group hospitals, depending on the age of the control-group hospitals' contracts. Ultimately, the ALJ, based on the testimony about post-merger pricing, determined that ENH's post-merger price increases exceeded those of the control groups by between 11 and 18 percent. These estimates were higher than those of ENH's expert, but, importantly, there appears to be no question that the ENH price increases were significantly greater than those of the control groups. Several other aspects of the performance-effects analysis are interesting. First, complaint counsel made no effort to estimate the "competitive price" in the market and then compare ENH's prices to it, relying rather on the comparison of ENH's price increases to those of other similar area hospitals. Estimating the competitive price is always an extremely difficult task by itself, which becomes even more complex in a differentiated services market. Moreover, because hospitals provide differentiated services, all the hospitals were probably charging supracompetitive prices. In a somewhat curious statement, the ALJ noted that "[t]he evidence indicates, but does not conclusively establish, that Respondent's prices were supra competitive," citing complaint counsel's not attempting to prove the competitive price. 75 But he went on to explain that the merger was unlawful because it "enhanced ENH's market power, regardless of whether ENH's prices have yet risen to a supra competitive level." 76 There would seem to be no doubt that ENH's prices were supracompetitive; the question would be to what extent. 77 Second, the ALJ noted that the evidence "strongly suggests that [ENH's] prices did rise to a supra competitive level without a reduction in output."78 This seems strange because restricting output is precisely the mechanism by which a firm increases prices through the exercise of market power — price rises because quantity supplied falls. 79 It would have been helpful if the ALJ had explained this seeming anomaly. If ENH did not increase prices by restricting output, by what mechanism did it increase prices? Was its demand curve perfectly inelastic such that it could charge almost any price without restricting output? Third, according to the ALJ, health-plan networks in ENH's area needed either Evanston Hospital or Highland Park in their networks (together with other area hospitals) to have a viable network. Prior to the merger, according to complaint counsel and the ALJ, health plans were able to play the two hospitals off against each another, forcing them to compete to become participating hospitals. Subsequent to the merger, they could not do so because ENH engaged in the forced bundling of its hospitals — i.e. , ENH forced health plans to contract with all three hospitals to contract with any one of them. 80 There seem to be two related problems with the theory. First, there was little, if any, evidence that area health plans engaged in selective contracting prior to the merger or that they played Evanston and Highland Park off against each other to obtain lower prices. Of course, that might have been because both hospitals recognized the health plans' implicit threat not to contract with one of the hospitals if its rate demands got out of line and this forced both to offer competitive or near-competitive rates. But it also could have been because health plans believed that they needed both in their networks. Some explanation of this would have been helpful. In addition, the two hospitals did not appear to be particularly close substitutes for each other. Evanston was much larger, offered a significantly broader array of services, was a teaching facility, and had a better reputation. This would have decreased a payer's ability to play the two hospitals off against each other prior to the merger and also lessened any effect on competition from the merger. Fourth, the Merger Guidelines' theories of competitive harm from unilateral effects seemed inapplicable. The Merger Guidelines explain that, in general, there are two mechanisms by which a merger might allow the merged entity unilaterally to raise prices anticompetitively. 81 First, if the merging firms and other firms in the relevant market sell identical or almost identical services (i.e. , the relevant market is homogeneous), the merged entity, if sufficiently dominant, may be able to raise price by restricting output because other firms have too little capacity to increase their output by the amount necessary to compensate for the loss in output. 82 In ENH, however, the parties and the ALJ agreed that the relevant market was a differentiated services, heterogeneous market. Moreover, a firm with a post-merger 40 percent market share is not usually considered "dominant." Where the market is comprised of differentiated rather than homogeneous services, the concern about the merger is that, post-merger, the merged entity may be able to raise prices profitably because if one of the merging parties — e.g. , Evanston — raises its prices and the merging firms are close substitutes for each other, customers of that party who switch suppliers because of the price increase will patronize the other merging party — e.g., Highland Park (called "localized competition" between the merging firms). Depending on contribution margins, if a sufficient number of those customers switch to the second party, the price increase will be profitable — i.e., the price increase will be profitable if the profit lost by the party raising its price is less than the profit gained by the party to which customers switch. 83 The merger results in higher prices merely because it eliminates competition between the merging firms regardless of competition from any other firms. The differentiated products unilateral-effects theory requires proof of several factors. 84 The most important is that the services of the merging firms be close substitutes. 85 Indeed, the most substantial unilateral price increases are expected when the services of the merging firms are the first and second choices of a substantial number of customers — when, premerger, the merging parties are each other's most direct competitors, although the most-direct-competitor condition need not hold for some price increase to result. 86 The Merger Guidelines provide that where the services of the merging firms are the first and second choices of a substantial number of customers of the merged entity, where post-merger concentration figures exceed the Merger Guidelines safe-harbor concentration standards, 87 and where the merged firm's post-merger market share is at least 35 percent, the merger may have significant anticompetitive effects. 88 In the Oracle differentiated products merger decision, however, the court held that the defendant's post-merger market share must approach the monopoly level for the theory to apply. 89 Neither complaint counsel nor the ALJ discussed the differentiated products unilateral-effects theory. ENH did raise it, arguing that complaint counsel failed to show that Evanston and Highland Park were each other's most direct competitor. 90 The facts were not clear on this point. Several payer representatives testified that they viewed Evanston and Highland Park as each other's most direct competitor, 91 but other witnesses testified that other hospitals were those hospitals' closest competitor. 92 The ALJ emphasized the differences between the two hospitals, and apparently complaint counsel's expert did not testify that Evanston and Highland Park were the first and second choices of a substantial number of consumers. Moreover, there was apparently no empirical work done on the issue, such as the calculation or estimation diversion ratios (i.e., the percentage of customers leaving a merging hospital that increases its prices that goes to the other merging hospital as the second choice) 93 or any analysis of cross-elasticity of demand between the two hospitals. 94 Finally, this was not a situation in which only one of the hospitals, Evanston, for example, raised its prices and the price increase was profitable because Evanston customers then substituted Highland Park. ENH raised prices at all three of its hospitals. Thus, the facts did not fit the usual theory explaining how a merger in a differentiated services market can result in unilateral anticompetitive effects. V. Is Performance or Direct Evidence of Anticompetitive Effects Sufficient To Prove a Violation of Section 7? No decision has held that proof of actual post-merger anticompetitive effects, without structural-analysis evidence, is sufficient to prove a violation of Section 7, 97 and the ALJ rejected the argument in ENH. Rather, he considered complaint counsel's post-merger pricing evidence simply as evidence corroborating its structural case, explaining that the post-merger pricing evidence "confirms the predictive assessments made by the structural market analysis"98 Accordingly, he dismissed Count II of the complaint as moot. But he went on to explain that, had the question required an answer, he would have held that "[t]he antitrust plaintiff must show at least the rough contours of a relevant market" and that the merging parties would have a "substantial share" of that market.99 Only then, he explained, can direct evidence of anticompetitive effects prove market power "'in lieu of the usual showing of a precisely defined relevant market and a monopoly market share.'"100 This seems problematic. Admittedly, relevant markets cannot and need not be delineated with scientific precision, but a "substantial share" of a "rough[ly] contour[ed]," relevant market (whatever that means), as opposed to market share in an accurately defined relevant market, would seem to provide little helpful information. 101 Moreover, other than to put icing on the cake, what purpose is served by requiring market-definition and market-share analysis if the transaction's actual anticompetitive effects are clear?102 Logically, and as a matter of policy, it is difficult to understand why, if direct or performance evidence is sufficient to prove market power or anticompetitive effects in Section 1 and 2 cases, it is not sufficient to prove the same in a consummated Section 7 case, where there is an actual track record of post-merger market performance. Market definition and shares, after all, are simply surrogates or predicators — assessment tools — of the transaction's actual or probable effect on market performance, which is (or should be) the important ultimate question. 103 This is true regardless of whether the applicable statute is Section 1, 2, or 7. Without question, courts should be extremely careful when evaluating and relying on performance evidence, and particularly expert studies, because it can be very ambiguous, and its validity depends so heavily on the particular models the experts use, the variables the model includes, and the reliability of the data the model uses. The old adage that statistics can show almost anything the analyst wishes them to show comes into play here. Accordingly, pragmatic and policy concerns might argue that a structural analysis should confirm the conclusion of the performance analysis before a violation can be found (although the same would be true in a Section 1 or 2 case). The arguments against the sufficiency of performance evidence in Section 7 cases are the related points that (1) almost every merger decision begins with the emphatic statement that market definition is an essential element, or "necessary predicate," 104 in every Section 7 case105 ; and (2) Section 7, by its own terms, prohibits only mergers with probable anticompetitive effects "in any line of commerce" (i.e. , a product market) "in any section of the country" (i.e., a geographic market), 106 while Sections 1 and 2 of the Sherman Act do not mention a restraint on competition, monopolization, or attempted monopolization of a "line of commerce" or "section of the country." The question is whether proof of a line of commerce and a section of the country are essential elements of a Section 7 claim simply because of this language. The better answer would seem to be no. First, there is no requirement in the language of Section 7 that the specific line of commerce or area of the country affected be identified.107 And if accurate, credible evidence shows that a merger results in actual adverse competitive effects, this subsumes that those effects occurred in some line of commerce in some area of the country. In fact, the performance analysis itself identifies the relevant market. 108 Second, another party (the Antitrust Division or an injured payor) could have challenged the ENH merger under Section 1 of the Sherman Act and, presumably, proved it by the same performance evidence that complaint counsel introduced. Basing the type of necessary proof on the identity of the plaintiff has no logic.109 Third, because the important ultimate question is the transaction's actual or likely effect, the requirement of structural evidence would seem to raise form over substance. Logically, structural evidence should corroborate performance evidence, not vice versa.110 VI. ENH's Defenses Unrelated to the issue of whether its price increases resulted from market power, ENH also argued (1) that prior to the merger, it and Highland Park were a single entity under the NHN umbrella since NHN was the sole member of both, and thus that the merger resulted in no new joining of economic power; (2) that the merging hospitals' nonprofit status should be considered in the competitive-effects analysis and showed that ENH had no incentive to behave anticompetitively; and (3) that Highland Park was a flailing, weakened-competitor firm, which mitigated the probability of anticompetitive effects from the transaction. The ALJ rejected all these arguments. A. "Learning About Demand" As a theoretical matter, the defense seems plausible. One requirement for a perfectly competitive market is that all participants — buyers and sellers — have perfect knowledge of prices and other relevant competitive variables. That a seller increases its prices upon learning that its competitors are obtaining higher prices and that the demand for its services is greater than it realized is not indicative of market power. The main problem with the defense, according to the ALJ, was that the facts did not fit the theory. Highland Park, according to the ALJ, was actually obtaining lower reimbursement than Evanston from a number of payers and in the aggregate.111 In addition, Evanston raised prices above those charged by Highland Park, arguing that as an academic teaching hospital, it "was entitled to even higher rates than Highland Park." 112 The ALJ rejected that argument on the ground that Evanston was not perceived as an academic medical center, and so it was inappropriate to compare Evanston's post-merger prices with those of academic medical centers. And finally, even if the defense were otherwise supported, the theory would not explain ENH's increasing prices at Highland Park as well as at Evanston. B. Quality of Care Quality has been a big bugaboo in antitrust analysis for a long time, particularly in health-care antitrust cases where it is a particularly important competitive variable in the eyes of patients (and thus in Stage 2 competition among hospitals). Because hospital services are differentiated, nonprice variables, such as quality, play a large role in consumer choice. As the ALJ noted, "The precise role of quality of care in the antitrust context has yet to be determined."115 Quality is difficult to factor into the antitrust analysis for several reasons: First, what is quality? Related to that, how broadly is quality defined? Does in include any nonprice competitive variable (e.g. , a convenient parking deck) or just clinical factors? Second, assuming quality can be defined, what methodology should be used to assess quality or the merger's effect on quality — i.e., how should the level of quality be measured? Third, how does the fact-finder compare quality across providers or to some "competitive" level of quality? Fourth, how does the fact-finder balance the merger's effect on quality against the effect on other competitive variables, particularly price? Fifth, where does the analysis of quality fit into the framework used to assess the transaction's effect on competition? And sixth, how are the burdens of going forward and of persuasion allocated between the parties on the issue of quality improvements (which largely depend on the point at which quality is folded into the antitrust analysis)? 116 Here, the methodologies used by the two sides' experts seemed to pass like two ships in the night. ENH's expert focused on improvements and additions in services and processes, while complaint counsel's expert focused on clinical outcomes and patient satisfaction. So there was no effective way to compare the expert's conclusions; doing so would have required comparing apples and oranges. The ALJ did not choose between the methodologies, although he seemed to focus on ENH's. He simply found that the improvements were not merger specific, that those at ENH were no greater than those at other area hospitals, and that they did not improve competition. C. Single-Entity Status Perhaps ENH's argument was that under the Copperweld doctrine,119 the fact that NHN was the sole member of both ENH and Highland Park was determinative by itself, that the actual degree of control by NHN over its hospitals was irrelevant, and thus that the hospitals were a single entity as a matter of law. But in the case of for-profit stock companies, such as those involved in Copperweld, the parent, by operation of law from the very nature of its ownership of its subsidiaries' stock, has complete ultimate control over their activities, regardless of the actual degree of integration. When push comes to shove between a parent and subsidiary, the subsidiary cannot implement economic interests separate or divergent from those of the parent. But here, the ALJ seems correct that, under the NHN umbrella, the member hospitals did not exhibit "economic unity" or "function as an economic unit, but rather retained autonomy with respect to hospital administration, staff, delivery of health services, budget and expansion plans."120 As a result of the degree of autonomy permitted each hospital under the arrangement, NHN could not control their activities or, as in the Copperweld situation, "'assert full control at any moment if [its hospitals] fail[ed] to act in [NHN's] best interests.'"121 This analysis and the future Commission and court of appeals decisions may have implications in analyzing the circumstances under which so-called hospital "virtual merger" transactions result in a single entity for antitrust purposes. This is a crucial issue if the hospitals then contract together with payors or allocate services between themselves — activities that may result in per se violations of Section 1 of the Sherman Act if the hospitals are not a single entity. Very generally, a virtual merger results where two hospital systems combine all, or almost all, of their hospital operations into a joint-operating-type arrangement, but each parent entity retains ownership of its hospital's assets.122 In most situations, each system also retains at least some reserved powers over its hospital. Thus, there is no single ultimate parent entity over the hospitals and the systems may not cede absolutely all authority and control over the hospitals to the operating entity, but the hospitals integrate their operations to such an extent that they actually function as a single entity. 123 The initial decision's analysis indicates that the presence (and by analogy, lack) of a single ultimate parent entity is not necessarily determinative of single-entity status, but that the court may need to conduct a relatively detailed factual inquiry to determine whether, as a practical matter, the hospitals function as a single economic unit or have divergent economic interests.124 In the case of hospital virtual mergers, the result may hinge in large part on the degree to which any reserved powers dilute the operating entity's ability to control the hospitals' short-term and long-term operations. D. The Merging Hospital's Nonprofit Status E. "Flailing Firm" Conclusion The appellate bodies, however, may disagree with some of the ALJ's other conclusions and analyses. Based on the ALJ's findings and conclusions about ENH's post-merger pricing, there appears to be a good chance that that the appellate bodies would agree with complaint counsel both that the relevant geographic market should have been limited to the three ENH hospitals and that complaint counsel's performance evidence was sufficient to prove a violation, even absent the structural evidence.132 It is simply difficult, under the Merger Guidelines' methodology for defining relevant geographic markets, to square the ALJ's findings about ENH's price increases with his decision to include hospitals in addition to the ENH hospitals in the relevant market. And there seems to be no logical reason to require a structural analysis if the actual effects of the transaction are (and this is an important qualification) very clear. Because the vast majority of merger challenges occur prior to consummation and thus before the existence of any post-merger performance evidence, another interesting and important question is how the ultimate decision in ENH might affect the analysis of those challenges. For example, if the Commission and court of appeals agree with complaint counsel that post-merger performance evidence is sufficient to prove a violation in a post-merger challenge, will econometric estimates of post-merger unilateral price increases based on merger simulation or other econometric models be sufficient to prove a violation of Section 7 in a premerger challenge? 133 And if patient-flow data is irrelevant, will courts weigh payer testimony about their "needs" and "must have" hospitals more heavily than in prior hospital-merger decisions? In sum, the ENH case raises a plethora of interesting issues to which the enforcers, the courts, antitrust economists, the antitrust bar, and hospitals will need to devote much attention and thought. Make no mistake about it: The ultimate analysis and result in ENH will be extremely important. The decision may very well determine whether hospital-merger enforcement returns to its previous state of hibernation or whether the enforcement agencies become even more active in challenging hospital mergers than they were in the 1990s. The answer to this question could have dramatic ramifications for both health-care and antitrust policy. Notes 1 FTC Dkt. No. 9315 (Oct. 20, 2005), available at http://www.ftc.gov/os/adjpro/ 2 15 U.S.C. § 18. 3 See Mark Taylor, SPLITTING HEADACHES: FTC JUDGE'S RULING COULD HAVE FAR-REACHING IMPACT, Modern Healthcare, Nov. 14, 2005; Margaret E. Guerin-Calvert, et al., THE FTC'S ROUND ONE VICTORY IN ITS CHALLENGE TO THE EVANSTON NORTHWESTERN HOSPITAL MERGER: WHAT DID THE ALJ FIND, AND WHAT ARE ITS IMPLICATIONS?, AHLA Hospitals and Health Systems Practice Group Member Briefing (Nov. 2005); Theresa E. Weir, AN OVERVIEW AND ROADMAP TO THE KEY ISSUES IN "THE EVANSTON CASE," ABA Antitrust Health Care Chronicle, Oct. 2005, at 11. 4 See, e.g. , Announcement, "The Evanston Case: Views From Both Trenches," AHLA Teleseminar, Jan. 5, 2006; Announcement, "Implications for Merger Analysis in Health Care After the Evanston Decision," ABA Antitrust Law Section and American Hospital Association Teleseminar, Dec. 1, 2005. 5 "FTC Announces Formation of Merger Litigation Task Force," FTC Press Release (Aug. 28, 2002), available at http://www.ftc.gov/opa/2002/08/mergerlitigation.htm. 6 In general, antitrust merger jurisprudence posits that "mergers should not be permitted to create or enhance market power or to facilitate its exercise." U.S. Dep't of Justice & Federal Trade Comm'n, Merger Guidelines § 0.1 (1992 & Supp. 1997) ( "Merger Guidelines" ). "Market power," as defined in the Merger Guidelines and as used in this article, means "the ability profitably to maintain prices above competitive levels" or "lessen competition on dimensions other than price, such as product quality, services, or innovation" "for a significant period of time." Id. & id. n.6.
7 Most antitrust analysis is based on the stated or unstated assumption that the relevant market is homogeneous — that is, that each of the firms produces perfect substitutes for the products and services of all other firms in the market. In this case, the only competitive variable — the only thing customers care about — is price. In a differentiated services (or heterogeneous) market, the services of the different firms are not so different that they constitute separate relevant markets but they do differ sufficiently so that different customers prefer the products or services of different firms for reasons other than price. In hospital markets, these differences may be geographical (i.e., patients may prefer one hospital over another because it is closer to their home) or may be based on a myriad of other factors (e.g. , breadth of services, quality of services, medical staff, reputation, amenities, courtesy of the nursing staff, etc.). That customers have non-price-based preferences means that each hospital may have some degree of market power — i.e. , a downward-sloping demand curve and thus an ability to charge a price higher than its marginal cost — because those consumers are willing to pay more for its services than for a competitor's services. "Economists recognize that a firm that sells a differentiated product is somewhat insulated from competitors." ABA Section of Antitrust Law, Market Power Handbook 3 (2005); see also Gregory J. Werden, DEMAND ELASTICITIES IN ANTITRUST ANALYSIS, 66 Antitrust L.J. 363, 371 (1998) ("every seller of a product that is differentiated with respect to any relevant dimension almost certainly has some market power").
8 See United States v. Oracle Corp. , 331 F. Supp.2d 1098, 1113 (N.D. Cal. 2004) ( "Oracle" ) ("There is little case law on unilateral effects merger analysis."). This article cites Oracle frequently because that decision is the most detailed to date discussing a differentiated products unilateral-effects merger theory and because it is both interesting and instructive to compare the analyses there with the ALJ's analyses in ENH. Oracle was an Antitrust Division challenge to the proposed merger of Oracle and PeopleSoft, which both sell a particular type of business data-processing software application. The court found for the defendant in a relatively long and complicated opinion. 9 FTC v. Columbia Hosp. Corp. , 1993-1 Trade Cas. (CCH) 70,209 (M.D. Fla. 1993) (preliminary injunction order); FTC v. University Health, Inc. , 938 F.2d 1206 (11th Cir. 1991). 10 FTC v. Tenet Health Care Corp. , 186 F.3d 1045 (8th Cir. 1999); FTC v. Butterworth Health Corp. , 121 F.3d 708 (6th Cir. 1997) (per curiam unpublished opinion), reprinted at 1997-2 Trade Cas. (CCH) 71,863; FTC v. Freeman Hosp. , 69 F.3d 260 (8th Cir. 1995); FTC v. Hosp. Bd. , 38 F.3d 1184 (11th Cir. 1994). 11 Hosp. Corp. of Am., 106 F.T.C. 361 (1985), aff'd, 807 F.2d 1381 (7th Cir. 1986); Am. Med. Int'l, Inc., 104 F.T.C. 617 (1984). 12 Adventist Health Sys./West, 117 F.T.C. 224 (1994). The Antitrust Division has not done much better. Its record is 2 wins (United States v. Rockford Mem'l Corp., 898 F.2d 1278 (7th Cir. 1990); United States v. Hosp. Affiliates, Int'l, Inc. , 1980-81 Trade Cas. (CCH) 63,721 (E.D. La. 1980)) and 3 losses (United States v. Mercy Health Servs. , 107 F.3d 632 (8th Cir. 1997); United States v. Carilion Health Servs., 892 F.2d 1042 (4th Cir. 1989) (per curiam unpublished opinion), reprinted at 1989-2 Trade Cas. (CCH) 68,859; United States v. Long Island Jewish Med. Ctr. , 983 F. Supp. 121 (E.D.N.Y. 1997)), and it is on a two-game losing streak. State attorneys general, when not filing jointly with a federal agency, are winless — 0 and 1. That loss is the most recently decided hospital-merger antitrust challenge (prior to ENH), decided in 2001. Cal. v. Sutter Health Sys. , 130 F. Supp.2d 1109 (N.D. Cal. 2001). 13 This FTC won-loss record is somewhat misleading because it fails to consider several hospital-merger challenges the FTC "won" through consent orders without litigation. 14 Testimony of J. Thomas Rosch before the Senate Committee on Commerce, Science & Transportation (Nov. 14, 2005) (FTC's "likely priorities over the next few years" are "energy, health care and high-tech"); Testimony of William E. Kovacic before the Senate Committee on Commerce, Science & Transportation (Nov. 14, 2005) ("The most pressing matters include energy . . .[,] health care, and information security and privacy."). 15 The complaint ( "Compl." ) is available at http://www.ftc.gov/os/caselist/ 16 Compl. 30, 32. 17 Initial Decision ("I.D.") at 3. 18 Under the traditional or structural approach for proving a violation of Section 7, the plaintiff introduces evidence of the relevant market and shows that the merger results in either a highly concentrated market or a merged entity with an unduly high market share. The plaintiff can rest at that point if it wants, having proved a prima facie case. The burden of going forward then shifts to the defendant to introduce evidence to rebut the plaintiff's quantitative evidence, or evidence of a qualitative nature to show that those statistics present an inaccurate picture of the merger's likely effect on competition. At that point, the burden of going forward shifts back to the plaintiff and merges with its ultimate burden of persuasion to show that the reasonably probable (or actual) effect of the acquisition is a substantial lessening of competition. See generally United States v. Citizens & S. Nat'l Bank, 422 U.S. 86 (1975). 19 Compl. 16. 20 See I.D. at 131. 21 As the ALJ explained, "Simply because the merging firms provide [two services] does not compel a finding that [both] are included in the product market." Id. at 134. Even if the services are purchased together as a package rather than separately (based, for example, on a uniform per diem amount), market power in one of the cluster's products could drive up the price of the cluster.
22 Complaint counsel's expert found "no decrease in the price of outpatient services . . . at the time of the increases in the prices of inpatient services" and that "payors did not accept lower outpatient prices in return for higher inpatient prices." I.D. at 92. 23 "If hospitals raise the price of inpatient services, managed care organizations cannot substitute outpatient services," id. at 29, and thus, "prices for inpatient services are not restrained by prices for outpatient services," id. at 133. 24 Merger Guidelines § 1.0 ("Market definition focuses solely on demand substitution factors — i.e., possible consumer responses [to a price increase]."). 25 See, e.g., FTC v. Tenet Health Care Corp., 186 F.3d 1045, 1052-54 (8th Cir. 1999) (including in the market one hospital 145 miles from the merging hospitals). 26 The seminal articles on critical-loss analysis are Barry Harris & Joseph Simon, THE OFT-FORGOTTEN ROLE OF PRICE-COST MARGINS IN ANTITRUST MERGER ANALYSIS, Int'l Merger Law, Feb. 1991, at 1, and Barry Harris & Joseph Simon, FOCUSING MARKET DEFINITION: HOW MUCH SUBSTITUTION IS NECESSARY?, 12 Res. in Law & Econ. 207 (1989). For more recent discussions, see, e.g. , Daniel O'Brien & Abraham Wickelgren, A CRITICAL ANALYSIS OF CRITICAL LOSS ANALYSIS, 71 Antitrust L.J. 161 (2003); David Scheffman & Joseph J. Simons, THE STATE OF CRITICAL LOSS ANALYSIS: LET'S MAKE SURE WE UNDERSTAND THE WHOLE STORY, Antitrust Source, Nov. 2003; James Langenfeld & Wenqing Li, CRITICAL LOSS IN EVALUATING MERGERS, 46 Antitrust Bull. 299 (2001). 27 See generally Federal Trade Comm'n & U.S. Dep't of Justice, IMPROVING HEALTH CARE: A DOSE OF COMPETITION Ch. 4 at 4-9 (2004 ) ( "FTC-DOJ Report" ). 28 For a helpful discussion of the two-stage hospital-competition model, see Gregory Vistnes, HOSPITALS, MERGERS, AND TWO-STAGE COMPETITION, 67 Antitrust L.J. 671 (2000). 29 I.D. at 23. 30 This was not a novel thought. See Gregory Vistnes, DEFINING GEOGRAPHIC MARKETS FOR HOSPITAL MERGERS, Antitrust, Spring 1999, at 28. 31 See, e.g. , I.D. at 30 ("Patient-flow data and the Elzinga-Hogarty test are inapplicable to geographic market definition for a differentiated product such as hospital services."), id. at 139 ("evidence regarding patient flow data, service areas and the Elzinga-Hogarty test are not probative in determining the relevant geographic market"). 32 Id. at 31 ("Patient flow data should not be used to determine the geographic market for hospital services, even apart from the Elzinga-Hogarty test . . . ."). 33 Professor Elzinga was also a government expert witness in Oracle, another differentiated products merger case, and testified there that his own test was not applicable in defining the relevant geographic market. The court rejected his testimony and held that the test, which defendant's expert had applied to define the market, was "an appropriate method for determining the ‘area of effective competition.'" Oracle, 331 F. Supp.2d at 1161-65. 34 I.D. at 30 ("Because patients do not set the price of hospital services, their willingness to travel tells us nothing about their sensitivity to price changes by merging hospitals."). 35 Contestable zip codes are zip codes within the tentative market that include a significant number of residents who use a hospital located outside of the tentative market — e.g. , a zip code where 20 or 30 percent of the residents of the zip code admitted to a hospital were admitted to a hospital outside the tentative market. The zip code is a "contestable" zip code in the sense (so the theory goes) that the fact that a significant percentage of zip-code residents already use a more distant hospital shows that many more residents of that zip code could and would switch to a hospital outside the tentative market in response to a price increase. See generally Gregory S. Vistnes, DEFINING GEOGRAPHIC MARKETS FOR HOSPITAL MERGERS, Antitrust, Spring 1999, at 28, 31. 36 See also FTC-DOJ REPORT Ch. 4 at 9, which briefly discusses the "silent majority" problem, noting that those patients traveling to more distant hospitals may be traveling for reasons unrelated to price, in which case their willingness to travel would not constrain price increases by the merging hospitals. Clearly, the reason that patients are willing to travel is a crucially important consideration.
37 See Merger Guidelines § 1.22 (discussing geographic market definition in the presence of price discrimination). 38 I.D. at 18; see also id. at 35, 142 ("managed care organizations themselves take into account patient preferences concerning hospital geography when building their hospital networks"). 39 Id. at 142 ("to the extent that employees value convenience, there is a derived demand by managed care organizations for hospitals that are convenient to their enrollees"). 40 See I.D. at 24 (To steer, "[p]atients are given financial incentives, through lower out-of-pocket expenditures, to use the preferred providers. . . . Use of other providers is discouraged by forcing patients to pay larger amounts themselves," but "managed care organizations in Chicago have not successfully engaged in steering their enrollees from one hospital to another in exchange for better rates."), 138 (noting that health plans were unable to steer patients to more distant hospitals to escape ENH's price increases, and explaining that "when ENH raised prices more than 5% after the merger, managed care organizations did not utilize alternative hospital network configurations to avoid the price increase"). 41 Id. at 137, 139. 42 Merger Guidelines § 1.2. 43 For a helpful discussion of the "hypothetical monopolist" analysis and background, see Gregory J. Werden, THE 1982 MERGER GUIDELINES AND THE ASCENT OF THE HYPOTHETICAL MONOPOLIST PARADIGM, 71 Antitrust L.J. 253 (2003). The FTC and Antitrust Division support application of this framework to relevant geographic market definition in hospital-merger cases. FTC-DOJ REPORT Ch. 4 at 20 (2004) ("The hypothetical monopolist test of the Merger Guidelines should be used to define geographic markets in hospital merger cases."). 44 Merger Guidelines § 1.2 (a relevant geographic market exists where "a hypothetical monopolist over that group of locations would profitably impose at least a ‘small but significant and nontransitory' [price] increase, including the price charged at the location of one of the merging firms"). 45 Complaint Counsel's Post-Trial Br. at 55 ("In the market . . . , ENH has the only hospitals, giving it a monopoly in the provision of inpatient services sold to health plans."), available at http://www.ftc.gov/os/adjpro/d9315/ 46 I.D. at 150 ("The geographic market, as proposed by Complaint Counsel . . . included only the ENH hospitals . . ., giving ENH a monopoly in the provision of inpatient services sold to managed care organizations."). 47 See, e.g., id. at 140, 142 ("although patients may use hospitals outside of the geographic market, the evidence demonstrates that those hospitals do not constrain Respondent's pricing . . . and are not hospitals to which managed care organizations can realistically turn to construct their local hospital networks"). 48 Id. at 144. The ALJ went on to conclude that "[s]hould ENH hospitals be excluded from a payor's network, a patient living within the ENH triangle would only have to drive past one hospital to reach a hospital within the geographic market," id., also suggesting that payors could exclude ENH from their networks and still market viable networks; see also id. at 149 ("Based on the evidentiary record, it seems reasonable that in the face of probable, future anticompetitive pricing, managed care organizations could create a network excluding the ENH hospitals and including the next proximal set of geographically close hospitals where consumers could go to seek ‘practical alternative' acute care inpatient hospital services."). He said a plan could develop a viable network if one of the 7 hospitals he listed (three of which were the ENH hospitals) were in the network. Id. at 142. If true, it is difficult to understand why ENH was a "must have" system or how it could have market power. 49 See, e.g., FTC v. Tenet Health Care Corp. , 186 F.3d 1045, 1057 (8th Cir. 1999). 50 I.D. at 54, 58, 59 (emphases added). 51 In Oracle, the court explained: [T]he court found the testimony of the customer witnesses largely unhelpful to plaintiffs' effort to define [the relevant market]. . . . Each of these witnesses had an impressive background . . . . 331 F. Supp.2d at 1130-31 (first emphasis added; remaining emphases in original). Perhaps the other side of the coin is that there is no obvious incentive for health plans to testify against the merger (assuming the plan is not part of a health system whose hospital competes with the merging hospitals and thus fears the efficiency effects of the merger) unless there is some reason it honestly believes that the merger, on balance, will be anticompetitive. If the merger is likely to have pro-consumer effects (e.g., lower prices, higher relative quality for the price, etc.), then presumably, health plans would support it. 52 Id. at 56. 53 Cf. Oracle, 331 F. Supp. 2d at 1131 ("But the issue is not what solutions the customers would like or prefer . . .; the issue is what they could do in the event of an anticompetitive price increase by a post-merger Oracle.") (emphasis in original). 54 I.D. at 142. 55 Indeed, a representative of a health plan testified that customers wanted a hospital within 30 miles of where they lived or worked. I.D. at 35. 56 And there was testimony from health-plan representatives that Condell did compete with ENH. Id. at 39. 57 Indeed, this seemed to be the conclusion of the FTC and Antitrust Division in their 2004 report on competition in health care. See FTC-DOJ REPORT Ex. Summ. at 26 ("Evidence regarding the willingness of consumers to travel and physicians to steer consumers to less expensive alternatives should also be considered by [c]ourts."), Ch. 4 at 19, 20 (2004). 58 Not surprisingly, the ALJ also concluded that entry barriers were high thanks to Illinois certificate-of-need regulation (and the fact that the state would deny an application for new beds in ENH's area) and that, even absent CON regulation, it would take almost three years to build a new hospital. I.D. at 123. 59 Unfortunately, this information, as well as the shares of the other hospitals in the relevant market, were redacted. Id. at 42 60 I.D. at 43. 61 See, e.g. , ABA Section of Antitrust Law, MARKET POWER HANDBOOK 26 (2005) ("single firm market shares help courts in assessing a firm's unilateral market power, while concentration statistics are used to analyze likely collusive effects of a proposed merger"). 62 I.D. at 200. 63 Merger Guidelines § 1.51 (providing that a merger increasing the HHI more than 100 points to over 1800 presumptively results in market power). 64 United States v. Philadelphia Nat'l Bank, 374 U.S. 321, 364 (1962) (30 percent market share "threaten[s]" "undue concentration"). In Oracle, the court explained that "[u]nder PHILADELPHIA NAT[IONAL] BANK, a post-merger market share of 30 percent or higher unquestionably gives rise to the presumption of illegality" but then held that "[a] presumption of anticompetitive effects from a combined share of 35% in a differentiated products market is unwarranted. Indeed, the opposite is likely true." 331 F. Supp.2d at 1110, 1123. 65 The exception is Hospital Corporation of America v. FTC, 807 F.2d 1381 (7th Cir. 1986), in which two mergers increased respondent's market share from 13.6 percent to 26.7 percent. That case, however, was a coordinated-interaction case rather than a unilateral-effects case. 66 See, e.g., Midwestern Mach., Inc. v. Northwest Airlines, 167 F.3d 439, 442 (8th Cir. 1999) ("No restraints, monopolies, or substantial lessening of competition need actually occur to violate section 7."). Some courts, however, hold that the standards under Section 1 and Section 7 have dovetailed such that there is little, if any, difference today. See, e.g., Nelson v. Monroe Reg'l Med. Ctr. , 925 F.2d 1555, 1561 n.3 (7th Cir. 1991). 67 See, e.g., Hosp. Corp. of Am. v. FTC, 807 F.2d 1381, 1389 (7th Cir. 1986) ("Section 7 does not require proof that a merger . . . has caused higher prices . . . . All that is necessary is that the merger create an appreciable danger of such consequences in the future. A predictive judgment, necessarily probabilistic and judgmental rather than demonstrable . . . is called for."). 68 See, e.g. , I.D. at 87. 69 Id. at 166 ("Complaint Counsel acknowledges that ‘large price increases alone do not mean that the merger gave ENH market power.'"); cf. N. Tex. Specialty Physicians, FTC Dkt. No. 9312 (Nov. 29, 2005) Slip Op. at 36 ("We agree that higher physician rates, by themselves, are of no antitrust significance. They may indeed be associated with higher quality of care or with different competitive conditions in various localities."), available at http://www.ftc.gov/os/adjpro/d9312/051201opinion.pdf. 70 See, e.g. , Herbert Hovenkamp, FEDERAL ANTITRUST POLICY § 12.5 at 527 (3d ed. 2005) ("Firms in product differentiated markets face demand curves that slope downward, although perhaps only slightly. At any given output level there will be different customers willing to pay different prices for the output of different sellers. Likewise, substantial product differentiation often entails that different firms have different costs . . . ."); Complaint Counsel's Post-Trial Brief at 45 (complaint counsel's "expert had to determine whether ENH's price increases were attributable to changes in the marketplace that would affect all hospitals equally."). 71 See I.D. at 62: Because prices can vary in a market for a differentiated service for many different reasons, one may not conclude anything about market power by merely using a cross-sectional analysis of hospital prices at a single point in time . . . . See also FTC-DOJ REPORT Ch. 4 at 5 n.22: Here, the competitive effect of the transaction is identified by comparing the change in price at the merging hospitals to the change in price (measured over the same time period) at a set of "control" hospitals. The control hospitals are hospitals in other geographic areas that are otherwise similar to the merging hospitals. Note, however, that a price increase by itself may not be sufficient to prove anticompetitive effect. 72 Complaint-counsel's expert compared ENH's price increases to those of three control groups of hospitals: (1) all general acute-care hospitals located in the Chicago Primary Metropolitan Statistical Area ("CPMSA"); (2) all hospitals in the CPMSA not involved in a merger between 1996 and 2002; and (3) because Evanston was a teaching hospital, all general acute-care hospitals in the CPMSA at which there was some teaching activity. 73 He also conceded, as he had to, that ENH's post-merger pricing patterns to certain payers were "consistent with ENH obtaining market power through the merger." I.D. at 87. 74 Complaint counsel's expert found that ENH's patient mix did change post-merger but not sufficiently to explain the difference between its price increases and those of the control-group hospitals. See id. at 94. 75 Id. at 155. 76 Id. at 154. 77 The ALJ's conclusion that ENH's prices were probably supracompetitive was based on a finding that its prices were higher than those of the other four hospitals in the relevant market. But because the hospitals' services were differentiated, one would expect their prices to be different, and there are many possible benign explanations for why ENH's prices might be the highest. 78 Id. at 155 (emphasis added). Indeed, ENH argued that the evidence showed that it had increased output. Respondent's Post-Trial Brief at 37, available at
http://www.ftc.gov/os/adjpro/d9315/ 79 See, e.g., Richard A. Posner, ANTITRUST LAW 9 (2d ed. 2001) ("A monopolist is a seller . . . who can change the price at which his product will sell in the market by changing the quantity that he sells."); ABA Section of Antitrust Law, MARKET POWER HANDBOOK 1 n.3, 5 (2005) (explaining that "[g]enerally a monopolist will reduce sales to increase price," although noting that there are exceptions — for example, where it can engage in perfect price discrimination); Oracle, 331 F. Supp.2d at 1114 ("the monopolist is able to maximize profit . . . by reducing output"). Indeed, this restriction of output is what leads to the welfare loss from market power. See Dennis W. Carlton & Jeffrey M. Perloff, MODERN INDUSTRIAL ORGANIZATION 143 (2d ed. 1994) ("If a monopoly restricts output and raises its price above marginal cost, society suffers a deadweight loss."). 80 Although the ALJ appeared to find ENH's forced bundling anticompetitive itself, it would seem that even if ENH had not required payors to contract with all three hospitals, it could manipulate the prices at the individual hospitals in individual transactions with payors to exploit whatever market power the system had. The problem would seem to be not that ENH engaged in forced bundling but that it could control the pricing decisions of all three hospitals. 81 See generally Merger Guidelines § 2.2; see also Oracle, 331 F. Supp.2d at 1113-14. In its Chicago Bridge merger decision last year, the FTC explained: Such unilateral effects are most likely to result in either of two circumstances. First, a firm might be able to increase prices in markets where competitors are distinguished primarily by differentiated products and the merging firms produce products that a substantial number of customers regard as their first and second choices (or, more precisely, where a substantial volume of sales are to customers who regard the products of the merging firms as their first and second choices). . . . Second, although no case seems to have dealt directly with such facts, economic learning holds that a firm might be able to increase prices above competitive levels in some [homogeneous product] markets where capacity is constrained and competitors may not be able to increase output in response to an output restriction by the merged firm. Chicago Bridge & Iron Co., FTC Dkt. No. 9300 (FTC Dec. 21, 2004) Slip Op. at 6 n.34, available at http://www.ftc.gov/os/adjpro/d9300/ 82 See generally William M. Landes & Richard A. Posner, MARKET POWER IN ANTITRUST CASES, 94 Harv. L.Rev. 937 (1981). 83 For a more in-depth discussion of this theory, see Carl Shapiro, MERGERS WITH DIFFERENTIATED PRODUCTS, Antitrust, Spring 1996, at 23; Herbert Hovenkamp, FEDERAL ANTITRUST POLICY § 12.3d at 512-16 (3d ed. 2005). 84 See, e.g., Oracle, 331 F. Supp.2d at 1117: [I]t appears that four factors make up a differentiated products unilateral effects claim. First, the products controlled by the merging firms must be differentiated. Products are differentiated if no ‘perfect' substitutes exist for the products controlled by the merging parties. . . . Second, the products controlled by the merging firms must be close substitutes. Products are close substitutes if a substantial number of the customers of one firm would turn to the other in response to a price increase. Third, other products must be sufficiently different from the products controlled by the merging firms that a merger would make a small but significant non-transitory price increase profitable for the merging firms. Finally, repositioning by the non-merging firms [to replicate the services offered by the merging firms] must be unlikely. 85 See id. at 1139 ("The court must demarcate such a[n] . . . area of localized competition between Oracle and PeopleSoft [the merging parties] as a prerequisite to finding any likelihood of unilateral anticompetitive effects.") (emphasis in original). 86 See, e.g. , Herbert Hovenkamp, FEDERAL ANTITRUST POLICY § 12.3d at 514 (3d ed. 2005) ("[T]he closer the products made by the two merging firms, the more likely that the merger will produce a substantial price increase. However, they need not be the ‘closest' rivals before the merger can have sufficiently anticompetitive consequences."); Gregory J. Werden, DEMAND ELASTICITIES IN ANTITRUST ANALYSIS, 66 Antitrust L.J. 363, 408 (1998) ("It is not true that differentiated products mergers necessarily have insignificant effects on prices unless the merging products are next-best substitutes. . . . Mergers of products that are not next-best substitutes clearly can cause significant price effects."). 87 Merger Guidelines §1.51. 88 Id. § 2.21. 89 Oracle, 331 F. Supp.2d at 1118, 1123: In a unilateral effects case, a plaintiff is attempting to prove that the merging parties could unilaterally increase prices. Accordingly, a plaintiff must demonstrate that the merging parties would enjoy a post-merger monopoly or dominant position, at least in a ‘localized competition' space. 90 Respondent's Post-Trial Brief at 54-55 (arguing that complaint counsel's expert failed to testify that Evanston and Highland Park were close substitutes). 91 See I.D. at 31-32. 92 Id. at 33. 93 For a discussion of the diversion ratio and its importance in differentiated products unilateral-effects analysis, see, e.g. , Carl Shapiro, MERGERS WITH DIFFERENTIATED PRODUCTS, Antitrust, Spring 1996, at 23 ("This Diversion Ratio is the answer to the following question: If the price of Brand A were to rise, what fraction of the customers leaving Brand A would switch to Brand B?"). 94 In Oracle, the court was highly critical of the failure of plaintiffs' experts to present any econometric evidence, such as diversion ratios, relating to whether the merging parties were customer's first and second choices. 331 F. Supp.2d at 1172. 95 See generally ABA Section of Antitrust Law, ANTITRUST LAW DEVELOPMENTS 65-73 (5th ed. 2002). 96 See, e.g., United States v. Microsoft Corp. , 253 F.3d 34, 51 (D.C. Cir. 2001) (per curiam). 97 The court in Oracle may have come close. After finding that the government failed to prove its case under a structural analysis, the court explained, "Without the benefit of presumptions [based on concentration or market share], the burden remains upon plaintiffs to come forward with evidence of actual anticompetitive effects," 331 F. Supp.2d at 1165, suggesting that had plaintiffs done so, that would have been sufficient to prove a violation of Section 7. 98 I.D. at 153; see also id. at 200 ("This prediction is confirmed by direct evidence that ENH exercised its enhanced post-merger market power through elimination of a competitor and obtained post-merger price increases significantly above its premerger prices and substantially larger price increases obtained by other comparison hospitals."). Given the direct-effects evidence before him, it is difficult to understand why the ALJ would need to make any "predictions" about effects at all. 99 Id. at 201 (emphasis in original). 100 Id. at 201 (quoting Republic Tobacco Co. v. N. Atl. Trading Co. , 381 F.3d 717, 737 (7th Cir. 2004)). 101 Cf. Oracle, 331 F. Supp.2d at 1154 (noting that "if the market is not precisely defined, then the market participants and their relative shares will be ‘economically inaccurate'"). 102 As some courts have explained, "Market share is just another way of estimating market power, which is the ultimate consideration. When there are better ways to estimate market power, the court should use them." Allen-Myland, Inc. v. Int'l Bus. Machs. Corp. , 33 F.3d 194, 209 (3d Cir. 1994); cf. N. Tex. Specialty Physicians, 2005-2 Trade Cas. (CCH) 75,032 at 103,476 (FTC 2005) ("Simply put, it makes no sense to undertake the exercise of market definition if it will not affect the outcome in any way."). 103 As one commentator has noted in suggesting that performance evidence should be sufficient to prove a violation, "Market structure evidence is the surrogate for bad performance, not the other way around." Herbert Hovenkamp, FEDERAL ANTITRUST POLICY § 12.8 at 550 (3d ed. 2005). 104 United States v. E.I. du Pont & Co. , 353 U.S. 586, 593 (1957). 105 E.g., United States v. Marine Bancorp. , 418 U.S. 602 (1974); United States v. Engelhard Corp. , 123 F.3d 1302, 1305 (11th Cir. 1997) ("Establishing the relevant product market is an essential element in the Government's case."). None of these cases, however, were decided in the face of strong direct evidence of actual anticompetitive post-merger effects and an argument that this evidence was sufficient itself to prove a violation. 106 15 U.S.C. § 18. 107Cf. Herbert Hovenkamp, FEDERAL ANTITRUST POLICY § 12.8 at 549-550 (3d ed. 2005) ("Clayton § 7's ‘may substantially lessen competition' language does not require a given market structure or a given set of proofs about market concentration, firm market share, entry barriers or anything else"; also noting that if a merger resulted in anticompetitive behavior, "the merger has lessened competition"). 108 For example, if hospitals A and B merge, and the court finds that, by themselves, they could not raise price but rather would need to coordinate their pricing with hospital C — the next-closest hospital — before prices could be raised, the analyst knows that the relevant market consists of hospitals A, B, and C. 109 The FTC has no authority to enforce Section 1 of the Sherman Act. Analysis of the merger under Section 5 of the FTC Act, 15 U.S.C. § 5, would be identical to that under Section 1. Section 5, however, does apply to nonprofit entities, see 15 U.S.C. § 44, and thus complaint counsel could not have challenged the merger under Section 5. 110 The ALJ also relied on a number of ENH pre- and post-merger documents from which he concluded that the primary purpose for the transaction was to terminate competition between the two hospitals and "attain enhanced market power . . . [for] negotiations with managed care organizations," I.D. at 157, and that the parties recognized that this is what actually happened. Most of the documents are quoted, paraphrased, or discussed at id. 44-48. Although some of the documents are troublesome, others do not seem as "hot" as the ALJ suggests. But some of the documents exemplify the need to prepare documents and explain intent and effect carefully. For example, if a document explains merely that a purpose for the transaction is to "increase prices," it is important to know whether the intent and effect is that this result from aggregating market power or differentiating and offering a better product or service. The decision also teaches that the documents of consultants can be as harmful as the documents of the parties themselves. See, e.g., id. at 48, 53, 54 (merging hospitals' consultant telling hospitals they have leverage and egging them on to raise prices).
111 Id. at 173. 112 Id. at 171. 113 Cf. N. Tex. Specialty Physicians, 2005-2 Trade Cas. (CCH) 75,03 at 103,477 (FTC 2005) (noting that higher quality of care may provide a benign explanation for higher provider prices). ENH's argument assumes, as some payer representatives indicated in their testimony, that health plans are willing to pay higher prices for higher quality, a proposition with which many providers would disagree. Their perception is that most health plans view providers' services as relatively homogenous and thus that price is the salient competitive variable in selecting network providers. 114 Merger Guidelines § 4. For a discussion, see Richard D. Raskin & Bruce M. Zessar, TELLING THE EFFICIENCIES STORY: PRACTICAL LESSONS FROM THE HOSPITAL-MERGER FIELD, Antitrust, Spring 1999, at 21. 115 I.D. at 175. 116 For example, since the important question is whether the merger will result (or has resulted) in an anticompetitive "quality-adjusted" price increase and plaintiff has the burden of persuasion, it could be argued that plaintiff has the burden to show that the price increase did not result from an increase in quality. Or, quality could be considered as a procompetitive justification of the type considered in a Section 1 rule-of-reason analysis after the plaintiff proves the defendants have market power. In that situation, the defendants would have the burden of going forward, but not the burden of persuasion. Or, quality could be considered as one of the qualitative factors that might be used to combat plaintiff's prima facie case based on post-merger concentration or market-share statistics. Again, the defendant would have the burden of going forward. Or, finally, the quality improvements could be treated, as the ALJ seemed to do in ENH, as efficiencies, and the defendants would have the burden of persuasion to prove them (although the plaintiff would continue to bear the ultimate burden of persuasion to prove that the merger would likely lessen competition substantially | ||||||
