|This document is available in two formats: this web page (for browsing content) and PDF (comparable to original document formatting). To view the PDF you will need Acrobat Reader, which may be downloaded from the Adobe site. For an official signed copy, please contact the Antitrust Documents Group.|
PUBLIC REDACTED VERSION
ORAL ARGUMENT IS REQUESTED
IN THE UNITED STATES COURT OF APPEALS
FOR THE TENTH CIRCUIT
UNITED STATES OF AMERICA,
Plaintiff-Appellant, v. AMR CORPORATION, AMERICAN AIRLINES, INC., and
AMERICAN EAGLE HOLDING CORPORATION,
ON APPEAL FROM THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF KANSAS
(Honorable J. Thomas Marten)
REPLY BRIEF FOR APPELLANT UNITED STATES OF AMERICA
TABLE OF CONTENTS
TABLE OF AUTHORITIES
A.A. Poultry Farms, Inc. v. Rose Acre Farms, Inc., 881 F.2d 1396 (7th Cir. 1989)
Advo, Inc. v. Philadelphia Newspapers, 854 F. Supp. 367 (E.D. Pa. 1994)
Advo, Inc. v. Philadelphia Newspapers, 51 F.3d 1191 (3d Cir. 1995)
Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227 (1st Cir. 1983)
Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209 (1993)
Bushnell Corp. v. ITT Corp., 175 F.R.D. 584 (D. Kan. 1997)
Coastal Fuels of Puerto Rico, Inc. v. Caribbean Petroleum Corp., 79 F.3d 182 (1st Cir. 1996)
D&S Redi-Mix v. Sierra Redi-Mix and Contracting Co., 692 F.2d 1245 (9th Cir. 1982)
ILC Peripherals Leasing Corp. v. IBM Corp., 458 F. Supp. 423 (N.D. Cal. 1978), aff'd, 636 F.2d 1188 (9th Cir. 1980)
Janich Bros. v. American Distilling, 570 F.2d 848 (9th Cir. 1977)
Los Angeles Land Co. v. Brunswick Corp., 6 F.3d 1422 (9th Cir. 1993)
Morgan v. Ponder, 892 F.2d 1355 (8th Cir. 1989)
Multistate Legal Studies, Inc. v. Harcourt Brace Jovanovich, 63 F.3d 1540 (10th Cir. 1995)
Neuman v. Reinforced Earth Co., 786 F.2d 424 (D.C. Cir. 1986)
Pacific Engineering & Production Co. v. Kerr-McGee Corp., 551 F.2d 790 (10th Cir. 1977)
Photovest Corp. v. Fotomat Corp., 606 F.2d 704 (7th Cir. 1979)
Re/Max International, Inc. v. Realty One, Inc., 173 F.3d 995 (6th Cir. 1999)
Reazin v. Blue Cross and Blue Shield of Kansas, Inc., 899 F.2d 951 (10th Cir. 1990)
Rothery Storage & Van Co. v. Atlas Van Lines, 792 F.2d 210 (D.C. Cir. 1986)
Spaulding v. United Transportation Union, 279 F.3d 901 (10th Cir. 2002)
Stearns Airport Equipment Co. v. FMC Corp., 170 F.3d 518 (5th Cir. 1999)
Todd v. Exxon Corp., 275 F.3d 191 (2d Cir. 2001)
United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir.), cert. denied, 122 S. Ct. 350 (2001)
15 U.S.C. § 2
Fed. R. Civ. P. 56(c)
3 Philip Areeda & Herbert Hovenkamp, Antitrust Law (2d ed. 2002)
Janusz Ordover, Predatory Pricing, in 3 The New Palgrave Dictionary of Economics and the Law 77 (Peter Newman, ed. 1998)
Lawrence Sullivan & Warren Grimes, The Law of Antitrust (2000)
Richard Posner, Antitrust Law (2d ed. 2001)
1. This case is about predation, which is defined under section 2 of the Sherman Act, 15 U.S.C. § 2, as "abnormal" business behavior that makes no business sense except for its ability to harm competition. Multistate Legal Studies, Inc. v. Harcourt Brace Jovanovich, 63 F.3d 1540, 1550 (10th Cir. 1995). We demonstrated in our opening brief (U.S.Br. 33-46)(1) that the most objective way to determine whether American's conduct was rational, except for eliminating rivals, is to compare the costs of the extra capacity with the revenues it generated, because there is "no reason consistent with an interest in efficiency" for an airline to add capacity on a route if the additional revenues generated will fall short of the added costs. See Pacific Engineering & Production Co. v. Kerr-McGee Corp., 551 F.2d 790, 797 (10th Cir. 1977). We showed that American engaged in a calculated course of predation to protect its monopoly in numerous DFW city-pair markets whenever it perceived a threat of low-cost carrier ("LCC") hub formation. American inundated the LCCs' routes with so much money-losing capacity that it drove the LCCs below their break-even load factors ("BELFs"). U.S.Br. 9 n.9. American knew that this strategy would be "very expensive." U.S.Br. 9-10, 32. CEO Crandall cautioned in advance that there was "no point to diminish profit" by implementing it, unless American was confident it would "get them [the LCCs] out [of DFW]" and allow American to recoup its losses through maintenance of monopoly prices. GA753; U.S.Br. 10, 32.
2. On issues of fact, American does not deny that its strategists expected the incremental capacity to lose money. Rather, American argues (AABr. 27-31) that methodological flaws in the government's incremental cost studies doom its case. The rules of appellate review of summary judgments (U.S.Br. 28) lead to a quick rejection of American's argument, which is a reargument of conflicting expert opinions. See infra pp. 11-18. American's claim that it cannot charge supracompetitive prices at DFW (AABr. 47-50) is similarly untenable -- particularly in light of the district court's finding (ignored by American) indicating that it does. Op. 1149-50.
3. On issues of law, American fares no better. It argues that as a matter of law and economics adding money-losing extra capacity is not predatory unless it causes market-wide average variable cost ("AVC") to rise above price. In so arguing, American misrepresents marketwide AVC as the exclusive judicially endorsed test for evaluating predatory conduct. And it misstates the reasons for what judicial use of AVC there has been. Without judicial support, American offers an economic rationale for using AVC (alleged protection of "equally efficient competitors"), which experts ranging from Nobelist Professor Stiglitz to Judge Posner demonstrate is flawed. See infra pp. 3-11. Equally unattractive is American's claim that consideration of out-of-market recoupment is legally impermissible even when economic analysis most indicates the need for antitrust concern. AABr. 54-56. Scholarly authority and common sense compel rejection of this claim. See infra pp. 18-26. Finally, American's effort to justify a Sherman Act meeting competition defense that would apply when the predator's prices are below an appropriate measure of cost, and recoupment is assured (AABr. 34-43), is insubstantial. No appellate court has accepted this defense, which would protect inefficient monopolists, and this Court should not do so. See infra pp. 26-29.
Predation is defined as business behavior that makes no business sense except for its ability to harm competition. It logically follows that tests for predation should, to the extent possible, evaluate the business sense of conduct using standard business decision-making criteria. That means comparing incremental revenue and cost when incremental conduct, such as adding capacity, can be clearly identified. Such a comparison reveals that American's capacity additions made no sense except for the elimination of LCCs. American, however, argues that the exclusive test for predation should be a comparison of price with AVC over all units sold. This test does not comport with the definition of predation, is not simple to apply in the airline industry, and was not in fact the basis on which American made its capacity decisions.
1. American's various responses to our showing of predation are unconvincing. Its assertion (AABr. 17) that the government "seeks to avoid the need to meet Brooke Group's stringent standards" is incomprehensible because the government relies on Brooke Group's two-part test. Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209 (1993); see U.S.Br. 31-32, 37-40. American's emphasis on this point makes sense only in light of its consistent effort to pretend that, applied to this case, Brooke Group somehow mandates the use of routewide AVC as the exclusive rule of decision. The government agrees that AVC can be a useful test, particularly as a safe harbor for conduct when incremental or marginal cost cannot confidently be measured -- unlike the manipulation of capacity at issue here. But this and other appeals courts have concluded that marginal cost, not AVC, "is the best indicator of the competitive health of the industry." Pacific Engineering, 551 F.2d at 797. AVC is merely used as a surrogate for marginal cost when appropriate.(2) U.S.Br. 40-46; GA648. Even Professor Areeda, the original proponent of the AVC test, concluded that defendants should not be allowed to rely on AVC when marginal cost "significantly exceed[s] average variable cost." 3 Philip Areeda & Herbert Hovenkamp, Antitrust Law ¶740c at 431  (2d ed. 2002) ("A&H").(3) That is the case here. GA343-45, 394-95, 552-53.
American is correct that "capacity alone . . . tells the Court nothing as to illegality"(4) (AABr. 18), but its argument that empty seats had no effect on LCCs (AABr. 19) is an attempt at misdirection. Contrary to American's suggestion, unsold seats were both an inevitable result of its predation, and tangible evidence that its capacity additions were irrational except for their ability to exclude LCCs.(5) In the end, American added three seats for every additional local passenger it carried. U.S.Br. 12 & n.13. American has neither refuted nor explained these figures.(6) That explanation is found in American's business plans: American increased capacity to attract passengers away from the LCCs and force them below their BELFs by " [Redacted] ."(7) GA855. American's strategists and CEO knew both that LCC passengers would respond to these capacity additions, thus "[Redacted]" the LCCs, and that the resulting empty seats would "[Redacted] ." GA855; U.S.Br. 9-10, 35. Professor Berry's careful measures of costs and revenues quantitatively confirm American's prediction.
American also ignores the fact that the costs of the added flights were "avoidable" or "incremental" costs -- costs it could have saved by not adding the flights, and thus the costs most appropriately measured in testing American's conduct. Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227, 232 (1st Cir. 1983). Contrary to American's suggestion, the issue is not simply "whether American sold too many seats at too low prices." AABr. 19. Rather, the incremental cost of all the added seats matters. GA395. As Barry Wright explains, if a firm "fail[s] to cover these 'avoidable' or 'incremental' costs . . . it would do better to discontinue production." 724 F.2d at 232; see also Brooke Group, 509 U.S. at 223 (suggesting an "incremental" measure of cost).
American wrongly claims (AABr. 20 & n.16) that case law mandates a market-wide price-cost analysis. The cases, however, mandate only that conduct claimed to be predatory "must be capable . . . of producing the intended effects on the firm's rivals." Brooke Group, 509 U.S. at 225. In other words, it must be a "threat to competition." Bushnell Corp. v. ITT Corp., 175 F.R.D. 584, 588 (D. Kan. 1997). Most of the cases American cites are discussed in Morgan v. Ponder, 892 F.2d 1355, 1362 (8th Cir. 1989), which explains that they "all focus on the basic question of whether the alleged predatory act poses a genuine threat to the overall competition." There, the challenged conduct was found incapable of eliminating competition.(8) Here, LCC competition was eliminated. A&H fully agrees with the government's position.(9) In fact, Brooke Group's focus on below-cost pricing in only the "generic" segment of the market (U.S.Br. 37-39) dispels American's "just a fraction of the market" argument.(10) AABr. 20.
3. Citing no authority, American argues that market-wide AVC must be used because it indicates whether an equally or more efficient rival can be excluded from the market. AABr. 21-22, 24-27. This, however, is merely American's incorrect theory, which Professor Stiglitz explained (GA628-36; 648-49) is deficient because it "would allow the incumbent to drive out . . . a firm which could supply the marginal increment of output at lower cost than the incumbent."(11) GA631. Judge Posner recently reached the same conclusion. Richard Posner, Antitrust Law at 219 (2d ed. 2001) ("Posner").
The A&H gasoline hypothetical (U.S.Br. 43-45) provides the basis for a clear example. The refiner produces 1,000 gallons a day at an average cost of $1.00, and its cost of additional gasoline is $1.06. Suppose a new entrant can produce 1,000 gallons a day at an average cost of $1.04. Although the entrant can supply all incremental output more efficiently than the incumbent, American's AVC test would allow the incumbent to exclude the more efficient entrant by selling up to 2,000 gallons at a price of $1.03, which is below the incumbent's incremental cost. Thus American is flat wrong that "competition is threatened only when service in the market as a whole is priced below cost."(12) AABr. 25 (emphasis added).
American fares no better with its contention that for methodological reasons the government's tests do not actually measure incremental costs.(13) AABr 27-31. The probative weight of the government's cost evidence plainly presents a genuine issue of material fact on which the experts disagreed. Fed. R. Civ. P. 56(c). Thus in this Court's de novo review, summary judgment must be reversed. Spaulding v. United Transportation Union, 279 F.3d 901, 904 (10th Cir. 2002).
Professor Berry examined the costs of American's conduct using four tests:
Professor Berry found that American failed each test everywhere it was applicable.(15) GA351, 354-84.
These tests use the same cost and revenue accounts that American used to make decisions about route and flight profitability, including whether to serve a
route and with how many flights. GSA3-4, 33-36, 197 (American "[Redacted] "), 233-35 ("
"); GA341-43, 345. Professor Berry found American's cost accounts reliable (U.S.Br. 53), noting that they are "designed for decision-making." GA342; accord GSA15-16. Professor Maher, the government's accounting expert, described the accounting system as "modern, sophisticated" and "comparable to the best." GSA15. Modern managerial accounting systems, such as American's, "report costs that reflect economic reality" (GSA3, 7), and American's system "was designed to capture only those costs that are incremental to flight activity." GSA18.
Professor Berry's analysis of American's costs and revenues was cautious and thorough. He explained why his tests appropriately measured costs. GA351-54, 386-93. He identified incremental costs by analyzing and rejecting "possible non-predatory explanations" for cost increases, and testing for "false positive" findings. GA354, 384-85; see also GA355-59, 365, 371-72, 374-75, 379, 384-407, 423-34, 449-53. He also controlled for various factors affecting costs. GA352, 362-63 & n.126, 449-53. He used the several tests together as cross-checks to avoid misleading indications of predation. GA392. He explained why his cost tests were appropriate to identify predation. GA325, 352, 384-85, 433. And he conducted a regression analysis that identified a striking pattern: while American's capacity changes were systematically profitable in most markets, including markets where American competes with Southwest, they were systematically unprofitable in LCC markets. GA352, 449-53.
American argues that evidence based on its data is inadequate because it "did not actually measure [incremental] costs and revenues." AABr. 2, 12, 27, 28, 29 (emphasis added); cf. Op. 1202. American claims that its accounts cannot be used to calculate the incremental cost of added capacity because they do not "isolate the costs actually attributable to the capacity additions."(16) AABr. 27-29. American's only discernible rationale is that its accounts employ cost allocations.(17) AABr. 27-30. But cost allocation is a key part of managerial accounting that permits relevant and sensible recording and assigning of costs.(18) GSA5, 8-10, 12-13; GA329-31. American ignores Professor Maher's conclusion that the particular cost allocations in American's system were "reasonable" and that "the treatment of costs" in American's accounts provides "reasonable measures of profitability for evaluating business decisions relating to whether to add capacity to a route." GSA19-21, 27; accord GA329-31, 350. He further explained that Professor Berry's cost measure conservatively evaluates the incremental costs of adding capacity to a route because it excludes some costs likely to vary with the addition of capacity. GSA29-34; GA353. American's expert claims otherwise, so the issue is disputed. Moreover, Professor Berry re-did his test omitting the particular costs American cites in its brief (AABr. 30 citing AA0128), and American still failed.(19) GA385-86.
American's claim (AABr. 31-33) that Professor Berry's Tests 1 and 4 are legally improper "profit maximizing" tests is incorrect. American fails to contest our argument (U.S.Br. 48-50) on this point. American's additional citations to Professor Stiglitz's theories (AA0053-54, AA0094-95), and to proposed DOT tests (AA0220), do not concern Professor Berry's tests.
Three important points are clear. First, the district court faced genuine issues of material fact about the government's cost evidence, on which the experts disagreed. Second, Professor Berry's tests showed abrupt, significant, and persistent losses on the added capacity. GA358, 372, 378, 382-84. Third, American expected losses when it planned its predation, and never claimed that the losses were due to a good-faith miscalculation.
Finally, American simply is wrong that incremental cost tests are too "vague" or "complex" for business decisionmaking.(20) AABr. 64, 29 n.22. As A&H explains, a firm's analysis of whether to engage in predation is "no different" from the endless business decisions it routinely faces, such as whether to add a new product or enter a new market. A&H ¶727c at 326-27 ; see U.S.Br. 21 n.15. In each instance, a "rational firm makes an investment only after calculating that the anticipated payoff seems to be worth the costs and risks." A&H ¶727c at 326 . Thus American, like all rational firms, uses an incremental cost test routinely when it asks: will this conduct make money or not? Here, the answer plainly was "not" -- except for the elimination of competition that Section 2 condemns.(21)
1. American predated in four markets to protect its supracompetitive profits in thirty monopoly markets from the threat posed by LCC expansion and hub formation at DFW. GA317-24, 484-87. American accurately predicted that predation would be "very expensive" (Op. 1152; GA1192): it eventually cost American $41 million. Op. 1213. But forestalling LCC hub formation at DFW for just two years was worth at least $[Red]million.(22) GA419. American clearly had a realistic expectation of recovering its losses from the challenged conduct, Brooke Group, 509 U.S. at 222, when it "commit[ted] its resources to predation." A&H ¶727c, at 327 . While the government need not show that American actually recouped its predatory investment, the continuing absence of any LCC hub at DFW (U.S.Br. 20, 64-65) provides strong evidence of actual recoupment as well.
Business strategy documents paint a clear picture of the benefits and costs that American's management, including CEO Crandall, reasonably expected from predation.(23) See U.S.Br. 7-10, 57. American now contends that these documents indicate "subjective beliefs" and fail to provide the requisite "objective evidence." AABr. 46 & n.32. But Brooke Group's reference to "objective evidence," 509 U.S. at 241, cannot sensibly be interpreted to exclude carefully formulated strategic business plans anticipating that the exclusion of LCCs would make it profitable to undertake money-losing capacity additions.(24) Whether American anticipated profiting from exclusion is what the recoupment issue is all about.
2. American contends that the absence of "relevant structural barriers at DFW" made supracompetitive fares impossible (AABr. 47-50), and that recoupment outside the four markets in which predation occurred is irrelevant as a matter of law. Id. 54-60. Both contentions are wrong.
a. American ignores the fact that no inquiry into "structural barriers" is necessary when there is direct evidence of supracompetitive pricing. See, e.g., Re/Max International, Inc. v. Realty One, Inc., 173 F.3d 995, 1016-19 (6th Cir. 1999) (reversing summary judgment in light of direct evidence of monopoly power); Coastal Fuels of Puerto Rico, Inc. v. Caribbean Petroleum Corp., 79 F.3d 182, 196-97 (1st Cir. 1996) (dictum). The district court found that fares in markets served by Southwest and LCCs were "competitive prices" and were "significantly" lower than fares where there was no competition from Southwest or an LCC. Op. 1149-50. It follows that the "significantly" higher fares in the monopoly markets are "prices above a competitive level," as required by Brooke Group, 509 U.S. at 225, quoted at AABr. 46. Despite this supracompetitive pricing, American has maintained dominant shares in the thirty markets where the government alleged it had monopoly power. GA317-24, 484-85. In trumpeting the rise of LCC competition at DFW (AABr. 48-49), American ignores the facts that it has faced no non-stop LCC competition in these monopoly markets and that LCC entry has been far greater at other major hubs than at DFW.(25) GA278-79, 530.
American's assessment of market structure also is based on a highly selective review of the record. American points out that Delta operated a modest hub at DFW, but ignores strong evidence that Delta's presence did not negate American's monopoly power. GA312-17, 866; see also Op. 1147-48. American observes that Southwest serves many destinations out of Dallas Love Field, but omits the crucial fact that none of the markets alleged to be monopolies could have been served by Southwest as a matter of federal aviation law.(26) GA113, 125-26, 319-20.
b. American anticipated recoupment from the elimination of actual competition in the four predation markets, and from the elimination of potential competition in the remaining DFW monopoly markets. The out-of-market recoupment would derive both from destroying or disciplining the target LCCs, thus ending their hub building threat, and from signaling non-target LCCs, (U.S.Br. 61-62). Signaling deters entry by non-target LCCs through the establishment of a "reputation for predation." There is no support for American's argument that only in-market recoupment is legally relevant.(27)
The support American offers (AABr. 55) is mainly a few words plucked from a sentence in Brooke Group that directs courts to conduct "a close analysis of . . . the scheme alleged." 509 U.S. at 226. Recoupment in all the DFW monopoly markets is both the "scheme alleged" and the scheme that was meticulously planned by American at the highest levels of the company. See GA108-09, 115-17; U.S.Br. 7-10. Moreover, Brooke Group had no difficulty with the plaintiff's theory that recoupment of the predatory sacrifice in generic cigarettes would occur through higher prices for branded cigarettes. See U.S.Br. 39, 55.
American also offers Advo (AABr. 57), which indirectly supports the government's recoupment theory by noting that often "[p]redation makes economic sense" only because of the potential for a payoff outside the predation market. 51 F.3d at 1196 n.4; see U.S.Br. 56, 60.(28) American ignores this point and advocates a rule that precludes an attack on predation in the very scenario in which predation can be most profitably employed.
Without questioning the economic soundness of the government's recoupment theory, American argues that the concept of a "reputation for predation" should be rejected as a matter of law. Significantly, this argument provides no basis for rejecting consideration of American's substantial recoupment through preventing the target LCCs from expanding service at DFW. Moreover, American's argument runs counter to the great weight of scholarly authority. The leading antitrust treatise explains:
A&H ¶726d5, at 322 . Judge Bork has explained that "chasten[ing] rivals" is a form of predation. Neuman v. Reinforced Earth Co., 786 F.2d 424, 427 (D.C. Cir. 1986). Judge Posner finds the concept of reputation for predation sound. Posner at 211-13. And one of American's experts specifically says that "courts should consider" it. Janusz Ordover, Predatory Pricing, in 3 The New Palgrave Dictionary of Economics and the Law 77, 80-81 (Peter Newman, ed. 1998).
American's reliance on Advo (AABr. 57) is misplaced. Advo did not reject all reputation arguments.(29) Rather, it distinguished that plaintiff's reputation theory from other reputation theories, and rejected the plaintiff's because it implied "barriers to entry . . . in all markets." 51 F.3d at 1202. American recognizes that the government's theory is different and applies more narrowly, but urges this Court to reject it because it "likely would prevent summary judgment in every case" involving "multiple markets." AABr. 57-58. This is false; summary judgment can be based on issues other than recoupment, and it could be based on recoupment in the vast majority of cases, if the infrequency of the multiple-market scenario in reported cases is a useful guide.
American argues that it is difficult to "distinguish between a reputation for predation and one for aggressive, lawful competition," so that the government's recoupment theory "could lead to 'mistaken inferences' of predatory conduct." AABr. 59-60 (emphasis added). This misapprehends the government's theory. It is not the character of American's reputation that determines liability, but rather the character of its conduct. The only distinction we call on the court to make is between a reputation from predation and a reputation from aggressive, lawful competition. Hence, the government's theory of recoupment requires no distinction the court did not already have to make about the legality of the conduct establishing the reputation.
Finally, American claims that the government failed to show an actual "reputation for predation." AABr. 60-63. But this is just American's take on a disputed issue of fact, and there is substantial evidence on the other side. See U.S.Br. 61-62. The target LCCs withdrew from several markets and were not replaced (U.S.Br. 65), and there remains no LCC hub at DFW. Moreover, the issue is not whether American actually has recouped, but whether it had a realistic expectation of recoupment when it undertook its predation.(30)
American does not significantly challenge the government's showing (U.S.Br. 67-70) that a meeting competition defense is both unavailable under the Sherman Act and factually inapplicable here. It does not dispute that the text of the Sherman Act is devoid of any meeting competition defense. It does claim that this Court, as well as the Sixth and Seventh Circuits, have recognized the defense (AABr.35-37), but it gives this game away by conceding (id. 37-38) that no court of appeals has ever recognized the defense when the predator's prices are below an "appropriate measure" of cost.(31) And, as American says (AABr. 38), "[i]f a defendant's prices are above" the appropriate measure of cost, "then it does not need to raise a meeting competition defense."
A proposed defense that merely meeting a rival's price excuses what otherwise would be illegal predation is understandably without support in precedent, for it would undermine the goals of the antitrust laws. It would make Brooke Group's below-cost pricing prerequisite superfluous when it is most important -- when an entrenched, high-cost monopolist faces new, more efficient competition. And it would reward the predation of an inefficient firm with the maintenance of its monopoly.
Finally, as a factual matter, American cannot avail itself of any meeting competition defense, because it undercut its rivals' prices by offering greater value. Op. 1188. Its frequent flier program, AAdvantage, offers significant value to passengers.(32) American does not deny the value of that program. Rather, it argues that it should be ignored because its value is "intangible" and because the government "cites no evidence in the record (there is none) purporting to quantify the value of American's program." AABr. 43. These arguments are disingenuous. American knows full well the value of its frequent flier program and keeps it on its corporate books. And such data is in the record. GSA208-10. Moreover, if American were correct that "any analysis of quality . . . is inherently speculative and uncertain" (AABr. 44), the defense should not be allowed because it cannot be administered consistent with the purposes of the Sherman Act.(33)CONCLUSION
For the reasons stated here and in our principal brief, the grant of summary judgment should be reversed.
I, Robert B. Nicholson, a member of the bar of this Court, hereby certify that this brief contains 6,998 words, as counted by Word Perfect 7.0 used to prepare this brief, and thus is within the 7,000-word limitation set by the Court's order of September 24, 2001.
I, Robert B. Nicholson, a member of the bar of this Court, hereby certify that today, March 19, 2002, I caused copies of the accompanying REPLY BRIEF FOR APPELLANT UNITED STATES OF AMERICA to be served by Federal Express on the following persons:
1. We use these abbreviations: "Op." (district court's opinion); "U.S.Br." (government's brief); "AABr." (American's brief); "GA" (government's appendix); "AA" (American's appendix); "GSA" (government's supplemental appendix).
2. No court of appeals has adopted AVC as a "dispositive" measure. AABr. 19 & n.15. The cases American cites either do not adopt any standard, e.g., Advo, Inc. v. Philadelphia Newspapers, 51 F.3d 1191, 1198 (3d Cir. 1995), or, like this Court, utilize AVC as a proxy for marginal cost. E.g. Stearns Airport Equip. Co. v. FMC Corp., 170 F.3d 518, 532 (5th Cir. 1999) ("Ideally, [we] would look at the true marginal cost . . . . But because the true marginal costs of production are difficult to generate, this Court attempts to estimate them by using average variable cost.").
3. Our opening brief cites the 1996 edition of this volume. A 2002 edition is now available. Paragraph numbers and text of cited material are the same in both, but page numbering differs. In this brief we cite the 2002 edition's pages with the 1996 edition's pages in brackets. By Rule 28(j) letter, we will provide 2002 page numbers for A&H citations in our opening brief.
4. The government's experts evaluated "the entire package of [pricing and yield management] decisions that accompan[ied] the capacity change[s]." GA351 n.118.
5. American misleadingly suggests that its capacity remained largely unchanged after LCC exit. AABr. 4. The district court found that American reduced flights on DFW-MCI, DFW-ICT, and DFW-COS following LCC exit. Op. 1156, 1160, 1164. Moreover, capacity, i.e., the number of seats, decreased dramatically ([Red.]%) on these three routes. GSA251-56.
6. American suggests that some of the challenged capacity additions may have been warranted in light of high load factors (AABr. 4-10), but American normally has not added substantial capacity in markets with similar or even higher load factors (GA430-32, 548-51), no doubt because it knew doing so would have been unprofitable. Its capacity additions were systematically unprofitable when, and only when, confronting LCCs. GA424. American's further suggestion (AABr. 4), that Delta's exit in Kansas City required it nearly to double its number of flights, is wrong because in such circumstances American typically reduced capacity or added little. See GA355-56, 544.
7. American unconvincingly claims (AABr. 16-17) that its documents are insignificant "sound bites reflecting American's competitive zeal." We agree that courts should ignore documents that merely reflect a "greed-driven desire to succeed and glee at a rival's predicament." A.A. Poultry Farms, Inc. v. Rose Acre Farms, Inc., 881 F.2d 1396, 1402 (7th Cir. 1989). American's documents are meaningful because they are not heated bluster, but coldly calculated analysis. They articulate in detail American management's strategy to " [Redacted] " to "
8. American wrongly claims that "those cases did not indicate that only 'a handful' of sales were at issue." AABr. 25. For example, Morgan involved sales to only one customer. 892 F.2d at 1361-62. And Janich Bros. v. American Distilling, 570 F.2d 848 (9th Cir. 1977), emphasized that buyers bought allegedly predatory half gallons "in comparatively small amounts" (less than 5% of total sales), and "[b]ecause of this" a below cost price "would not be likely" to eliminate competition. Id. at 856 (emphasis added).
9. As A&H explains (¶742c1 at 464 ):
Because adding capacity was American's "mechanism," the added capacity is the relevant "line" or "arena" in which to make the comparisons.
10. After its extensive reliance on Brooke Group's formulation of the "appropriate measure of cost" test (AABr. 2, 10-11, 14-19, 32, 63-66), American's criticism of the government's reliance on Brooke Group's application of that test is curious. AABr. 22-24. Neither aspect of the opinion was a clear holding in the case, which presented only an issue of recoupment. 509 U.S. at 218-19. But both aspects provide useful guidance to the courts.
11. LCCs "are not burdened with inefficient work rules" and can have operating costs per mile 30-50% below American's. Op. 1151.
12. American incorrectly argues that the A&H hypothetical is inapposite because evidence that marginal cost exceeds AVC "is not present here," and that whenever it adds flights to a route, its "marginal or incremental cost of adding service [is] lower than the AVC of all flights on [the] route." AABr. 26. If, however, American were to add a flight attracting only one additional passenger, the marginal and incremental cost of carrying that passenger would be the entire cost of the added flight -- far greater than the average cost for all passengers on the route. Professor Berry explained that American's marginal and incremental costs exceed AVC here because American added so much capacity that its route-wide load factors dropped. GA343-45, 394-95, 552-53.
13. The district court's analysis of this issue was brief and cryptic. Op. 1202. Apart from two criticisms that are plainly wrong (see infra notes 16-17), the court appears not to have addressed American's contentions.
14. Professor Berry applied this "range of tests" to avoid misleading indications in arriving at his conclusion that American engaged in predation. GA385, 392. While we focus on the district court's errors with regard to Tests 1 and 4, American's criticisms of Tests 2 and 3 (AABr. 33-34) are incorrect. GA328, 332-33, 335, 343-45, 395-407.
15. American criticizes Tests 1 and 4 because they cannot both be applied to all episodes. AABr. 30. Contrary to American's claims, Test 1 was applied to Long Beach, GA382-84, and Test 4 was applied to the second Kansas City episode, GA369-70, as well as to Colorado Springs. GA379-80. Moreover, the law requires an "appropriate" measure of costs, and Professor Berry's caution in applying the tests only where appropriate is hardly grounds for criticism. As he explained, "no test can be applied in a vacuum," and he used "a range of tests" to be confident of his results. GA352.
16. Similarly, American says that Test 4 arbitrarily measures revenue by ignoring "seasonal variations in travel demand and passenger mix that make no two months alike in the airline industry." AABr. 29; see also Op. 1202. But Professor Berry used three different time periods to "examine the robustness of this result," where possible, "equalizing seasonality factors." GA354, 362, 379. Moreover, although American criticizes Berry's identification of the average price paid by incremental passengers, American's conduct almost uniformly failed Test 4 using an alternative calculation based on the (higher) average price paid by all passengers. GA347-48 & n.115.
17. American suggests (AABr. 28) that the cost increases identified by a before-and-after analysis might not be "attributable to the capacity additions," but American does not suggest another cause. The district court wrongly speculated that the increase in costs might be due to labor or fuel. Compare Op. 1202 with U.S.Br. 52. American has abandoned this argument. Moreover, Professor Berry carefully analyzed whether other causes existed, and found none. GA333-34, 351-52.
18. The superiority of cost allocation over simple cash accounting is familiar to anyone who has figured on a cents-per-mile basis the cost of using an automobile (including expenses that are frequent (e.g., gasoline), periodic (e.g., scheduled maintenance), and occasional (e.g., new tires)). Allocation of such costs on a per-mile basis matches real economic costs to the activities with which they are associated (GSA8-10), and avoids the errors of over-inclusion (e.g., counting the entire tire replacement cost against one trip) or under-inclusion (e.g., counting all other trips as if they did not produce any wear and tear on the tires). American implements these principles in its allocation decisions. GSA213 ("
19. Citing its expert, Dr. Dorman, American claims that its accounting system inadequately shows incremental cost because it "allocates certain costs down to individual flights and routes that are not driven by activity at the flight or route level." AABr. 30. But Professor Maher explained that "Dr. Dorman applies the wrong concept of avoidable cost in his analysis," while American's accounting approach, used by Professor Berry, "is sensible and is consistent with common practice in managerial accounting." GSA44, 47; see GA329-30. As yet another check, Professor Berry re-calculated Test 4 using an additional "incremental flight cost measure that American uses": American still failed. GA385-86.
20. American is wrong that "straightforward" simplicity makes AVC the test of choice. AABr. 29 n.22, 64. While AVC once "was thought to be a judicially manageable measure of cost, determining average variable cost has proven difficult and costly." Lawrence Sullivan & Warren Grimes, The Law of Antitrust 150 (2000). Airlines charge many fares on a single flight, so no single fare provides a useful comparison with AVC, and when airlines set fares, they do not know what per-passenger AVC will be, since that depends on how many passengers fly. Moreover, largely the same analysis of costs is required for AVC and Tests 1 and 4. GA334-35, 393. Which costs are variable depends on the magnitude and duration of the changes in capacity, output, and other factors that may cause costs to vary. And which such changes are relevant must be determined by reference to the challenged conduct, which also defines the increment for purposes of determining incremental cost.
21. It is not surprising that American's documents reveal its conclusion that "diminish[ing] profits" by adding capacity would be worthwhile only if it got the LCCs "out" of DFW, and American points to no evidence that it performed an AVC analysis prior to adding the flights.
22. American does not address this calculation. The district court labeled it "speculative" (Op. 1215 n.24), but the two-year period was not offered as a prediction of how long LCC hubs actually would be forestalled. Rather, it was used to show that the predatory sacrifice "easily" could be recouped. GA419. Had the two-year period been a prediction, hindsight demonstrates that it would have been conservative.
23. American's claim (AABr. 46 n.32) that these documents do not show it expected supracompetitive fares after predating hinges on its mistaken contention, discussed presently, that it was not charging monopoly fares before predating.
24. In Brooke Group the Court was not presented with strategic business plans detailing the precise predatory scheme alleged by the plaintiff. Rather, the plaintiff relied on documents merely "speak[ing] of a desire to slow the growth" of generic cigarettes, when the defendant's conduct did not "threaten in a serious way" to achieve that effect. 509 U.S. at 241-42. See also note 7, supra.
25. American's DFW passenger share has remained over 70%, while the collective share of LCCs has been around 2%. GA708, 1368. American claims that this 2% figure "is misleading," AABr. 48 n.34, but its use was prompted by the district court's remark that the LCCs' passenger share at DFW had "increased significantly." Op. 1209; see U.S.Br. 66. The 2% figure is from the same page of testimony (GA708) American cited in urging the court to make that remark. See GSA193-94.
26. As detailed by the district court (Op. 1148-49), service at Love Field has been restricted by federal law, which permits commercial service from Love Field in aircraft with more than 56 seats, which all Southwest's planes have (Op. 1149), only to points in Alabama, Arkansas, Kansas, Louisiana, Mississippi, New Mexico, Oklahoma, and Texas. American's predation in Wichita occurred before federal law allowed such service to Kansas. Wichita has not subsequently been served from Love Field.
27. American is correct that no case cited in our opening brief squarely holds that out-of-market recoupment is legally relevant (AABr. 57), but it is also true that no case squarely rejects it. American claims we "seriously misconstrue" Reazin v. Blue Cross and Blue Shield of Kansas, Inc., 899 F.2d 951 (10th Cir. 1990). AABr. 58. But we cited the case only for its discussion of an analogous and successful signaling effect by an entrenched monopolist that "constituted willful maintenance of its monopoly power." 899 F.2d at 954-55, 973.
28. American wrongly asserts that the district court in Advo, 854 F. Supp. 367, 376 n.31 (E.D. Pa. 1994), "expressly rejected the same type of recoupment theory the government makes here." AABr. 56 n.42. The last sentence of the cited footnote, which American omits, makes clear that the plaintiff's theory was one of cross-subsidization rather than recoupment.
29. As argued in our opening brief (U.S.Br. 63 n.44), other cases involving "reputation" issues are "inapposite." American disagrees (AABr. 58 n.43) as to Los Angeles Land Co. v. Brunswick Corp., 6 F.3d 1422, 1427 (9th Cir. 1993), but the court's point in that case was that "proof of monopoly power" cannot rest on conduct alone "particularly where a court has already determined that the acts in question were not exclusionary." In this case, "proof of monopoly power" does not rest on conduct alone, and the conduct in question was exclusionary.
30. American's statement of facts may be read to suggest that the target LCCs exited for reasons other than predation. AABr. 3-10. But American did not argue this in its summary judgment motion, and its management did not share this view when it planned and implemented its LCC strategy. E.g. GA919-20 (SunJet is "major concern"); GA891 (prevent Western Pacific from "
31. American places substantial weight on language in a district court opinion within the Ninth Circuit (AABr. 38), but the Ninth Circuit subsequently disavowed it. Compare ILC Peripherals Leasing Corp. v. IBM Corp., 458 F. Supp. 423, 433 (N.D. Cal. 1978), aff'd, 636 F.2d 1188 (9th Cir. 1980), with D&S Redi-Mix v. Sierra Redi-Mix and Contracting Co., 692 F.2d 1245, 1248 (9th Cir. 1982). American's reliance on the 1996 edition of A&H (AABr. 37) is hard to fathom, because the treatise never endorsed a meeting competition defense of the kind American advances, and, as American concedes (id. 40-41) the 2002 edition expressly would not allow the defense when the incumbent, as here, floods the market with extra capacity. American's attack on Professor Hovenkamp for expressing these views in the 2002 edition (AABr. 41-43 nn. 27-28) manages to be at once abusive and frivolous. In 1998 he consulted with the government on this case and was paid $2,250 for nine hours of work.
32. No LCC target except Vanguard had a frequent flier program, and Vanguard's program was inferior to American's. Op. 1188. Thus it is irrelevant to determine "by what margin the value of American's program exceeds the value of an LCC's program." AABr. 43.
33. American's claim (AABr. 63-67) that its anticompetitive conduct should be excused because it will be too hard to fashion appropriate relief was not examined in the district court, was not mentioned in the district court's opinion, and therefore cannot be usefully considered by this Court. Economists, for both the government and the defendants, agreed that a procompetitive remedy can be fashioned. GA656.