Plaintiff-appellant, the United States of America, appeals from an order entered on May 20, 1994 in the United States District Court for the Western District of New York (Telesca, Chief Judge ) granting a motion made by defendant-appellee Eastman Kodak Co. ("Kodak") to terminate two antitrust consent decrees. The first decree was entered in 1921 ("the 1921 Decree") after a finding by the district court that Kodak had monopolized the sale of cameras and photographic supplies. The 1921 Decree continues to impose a variety of restrictions on Kodak's business practices. Most notably, Section X of the 1921 Decree prevents Kodak from selling "private label" film, which is film marketed under a brand name other than Kodak's, typically that of a retail outlet. The second decree was entered in 1954 ("the 1954 Decree") without any adjudication of law or facts. Section V of the 1954 Decree continues to prevent Kodak from selling its film in a "bundle" with its photofinishing services.

After a nine-day evidentiary hearing, the district court granted Kodak's motion to terminate both decrees. See United States v. Eastman Kodak Co., 853 F. Supp. 1454, 1487-88 (W.D.N.Y.1994). In deciding to terminate the 1921 Decree, the district court relied on its findings that Kodak no longer possesses market power over the sale of film and that the elimination of the decree's remaining provisions would benefit consumers. With respect to the 1954 Decree, the district court found that the purpose of the decree--the creation of a competitive photofinishing market--had been accomplished, that neither Kodak nor its affiliates had market power in the photofinishing or film markets, and that allowing Kodak to sell film and photofinishing as a bundle would enhance competition. For the reasons that follow, we affirm the order of the district court terminating both the 1921 and the 1954 Decrees.

a. The 1921 Decree

The 1921 Decree arose out of antitrust enforcement proceedings initiated by the government more than 80 years ago. These proceedings led to a finding by the district court in 1915 that Kodak had monopolized the amateur camera, film, and photofinishing industries through acquisitions and a variety of exclusionary practices. See United States v. Eastman Kodak Co., 226 F. 62, 79-80 (W.D.N.Y.1915), appeal dismissed, 255 U.S. 578, 41 S. Ct. 321, 65 L. Ed. 795 (1921). Six years later, while its appeal of the district court's decision still was pending in the Supreme Court, Kodak entered into the 1921 Decree with the government and withdrew its appeal.

The 1921 Decree imposed both short- and long-term obligations on Kodak. In the short term, the decree required Kodak to divest itself of many of its acquisitions so as to "effectually dissolve the combination found to exist in violation of the [antitrust] statute." Id. at 81. These provisions were satisfied long ago. In the longer term, the consent decree proscribed certain sales and distribution practices that the government believed to be anti-competitive. A number of these restrictions remain in force today. Section X of the decree prevents Kodak from selling private-label film by requiring its products to "be labeled in such manner as to show clearly that the [product] is manufactured by [Kodak]." Sections VI and VII of the decree enjoin Kodak from entering into exclusive dealing contracts and from imposing other restraints on its dealers.

b. The 1954 Decree

The 1954 Decree concerned Kodak's practice of tying its photographic film and photofinishing services. Until the 1954 Decree, the sales price of Kodak color film included Kodak photofinishing. Since Kodak sold approximately 90 percent of the color film in the United States during that time, this practice allowed Kodak to maintain a 90-percent share of the photofinishing market as well. The government filed suit to stop this practice, and, prior to trial, Kodak and the government entered into the 1954 Decree. The decree principally required Kodak to make its color photofinishing technology available to competitors. Kodak fully complied with this requirement. The only provision of the 1954 Decree that remains operative today is Section V, which prevents Kodak from "[t]ying or otherwise connecting in any manner the sale of its color film to the processing thereof." The broad language of Section V precludes Kodak from engaging in the otherwise lawful practice of bundling its film and photofinishing services.

2. Today's Markets for Film and Photofinishing

a. Film

The marketplace for film has changed considerably in the last 80 years. Today, five companies manufacture all of the amateur color negative film (hereinafter "film") sold in the United States: Kodak, Fuji, Konica, Agfa, and 3M. All five are "well-financed, billion-dollar, multinational corporations selling film all over the world." 853 F. Supp. at 1471. On a world-wide basis, Kodak is the leading seller of film, with a 36-percent market share, followed closely by Fuji, which has 34 percent of the market. Id. at 1471 & n. 10. Konica has a 16-percent share of the world-wide market, followed by Agfa, with ten percent, and 3M, which has a four-percent market share.

In the United States, Kodak is far and away the leading seller of film with 67 percent of the market. Id. at 1472. It is followed by Fuji, which has between eleven and twelve percent of the United States market. When market share is measured in dollar terms, rather than in unit sales, Kodak's market share in the United States is even more impressive. Kodak accounts for 75 percent of United States film sales in dollar terms.1 Id. Part of Kodak's high market share in the United States is attributable to the large number of retail outlets that carry its products. Almost all retail outlets that sell film carry Kodak film. Only thirty percent of retail outlets carry Fuji film. However, Fuji film tends to be carried by mass-merchandisers, such as K-Mart and Wal-Mart, which account for between 45 and 50 percent of all film sales. Id. at 1474.

Notwithstanding Kodak's success in the United States, photographic experts agree that there is very little difference in quality among the various brands of film. Kodak and Fuji films are both considered excellent, and "[i]t would be very hard for any one to tell a good Fuji print from a good Kodak print." Konica and Agfa film, while slightly coarser-grained, also produce very good photographs. While 3M film is of slightly lower quality than the four other brands, it still is of acceptable quality to most consumers.

b. Photofinishing

Today, there are three principal types of photofinishers: mail-order macrolabs, wholesale and captive macrolabs, and minilabs. See 853 F. Supp. at 1482-84. Mail-order photofinishers operate large film processing labs ("macrolabs") and often have the lowest prices, but with the disadvantage of a longer turn-around time. Construction of a macrolab requires a several million dollar investment and takes six to nine months. Wholesale or "captive" macrolabs process film dropped off at retailers, such as drug stores.2 Wholesale labs tend to process film in two or three days, but also can provide overnight service at an additional price. Minilabs, the third and newest type of photofinisher, are installed at retail locations, such as supermarkets, to provide on-site photofinishing. A new minilab can be purchased for as little as $40,000 and can provide one-hour film developing. However, minilabs have a somewhat higher per-print cost and cannot handle the volume of work required by large retail customers. In 1992, minilabs and wholesale labs each processed 36 percent of the film in the United States, captive labs processed 21 percent, and mail-order labs processed only 7 percent.

In the last ten years, Kodak has reclaimed a large share of the photofinishing market through acquisitions. The most important was a joint venture to establish Qualex, Inc., a company that accounts for approximately 30 percent of the overall United States photofinishing business. Id. at 1486. In the United States, Qualex is the largest wholesale photofinisher, with more than 70 percent of the wholesale market, and the second largest operator of minilabs.

After a nine-day evidentiary hearing, the district court granted Kodak's motion to terminate both the 1921 and 1954 Consent Decrees in their entireties. See 853 F. Supp. at 1487-88. The first issue addressed in the district court's comprehensive opinion was the legal standard under which to evaluate Kodak's request to terminate or modify the consent decrees. The district court decided that the "flexible" standard that the Supreme Court recently set forth in Rufo v. Inmates of Suffolk County Jail, 502 U.S. 367, 112 S. Ct. 748, 116 L. Ed. 2d 867 (1992), should be applied. See 853 F. Supp. at 1465. Applying Rufo, the district court reasoned that the consent decrees could be modified or terminated on the basis of significant changes in market conditions. See id. at 1487-88.

With respect to the 1921 Decree, the district court first determined that Kodak no longer possesses market power over the sale of film. Id. at 1471-73. In so doing, the court defined the relevant geographic market to include not only the United States, but also Japan and Western Europe (collectively "the world-wide market"). Id. at 1471. Finding that Kodak only has a 36-percent share of the highly competitive world-wide market, the district court concluded that Kodak does not possess market power over film. Id.

Alternatively, the district court determined that, even if the relevant geographic market were defined as the United States, its determination that Kodak lacks market power would not be altered, despite Kodak's 67-percent market share in the United States. The district court reasoned that "[p]rice elasticities are better measures of market power" than market share data, id. at 1472-73, and found that Kodak would lose ten percent of its sales to its foreign competitors if it raised its price for film by five percent, a finding that the court deemed incompatible with the presence of market power. The district court also determined that, although Kodak film sells for a small price premium in certain types of retail outlets, the premium simply is evidence "of the perceived quality difference that exists in the minds of consumers who are satisfied with Kodak products." Id. at 1477. Having determined that Kodak no longer possesses market power under either geographic market definition, the court determined that termination of the 1921 Decree would enhance competition in the sale of film.

With respect to the 1954 Decree, the court found that the decree's purpose of creating a competitive photofinishing market had been accomplished. The court determined that the relevant product market includes all three types of photofinishers, and that the relevant geographic market is the United States. Id. at 1483-86. In this market, Kodak has only a thirty-percent share, which the court found to be too small to give rise to an inference of market power, particularly in light of the low barriers to entry. Id. at 1486. Finally, the court determined that allowing Kodak to bundle film and photofinishing would enhance competition in both markets. Id. at 1486-87.

1. Standard for Terminating Antitrust Consent Decrees

The government contends that the district court misunderstood and misapplied the legal standard governing termination of an antitrust decree. In opposing Kodak's motion, the government argued that the modification of an antitrust consent decree was governed by United States v. Swift & Co., 286 U.S. 106, 52 S. Ct. 460, 76 L. Ed. 999 (1932), where the Supreme Court stated that "[n]othing less than a clear showing of grievous wrong evoked by new and unforeseen conditions" permits modification or termination of an antitrust decree. Id. at 119, 52 S.Ct. at 464. Kodak argued for the "less stringent and more flexible" standard set forth in Rufo v. Inmates of Suffolk County Jail, 502 U.S. 367, 112 S. Ct. 748, 116 L. Ed. 2d 867 (1992). The trial court agreed with Kodak and indicated that it was applying the Rufo standard. 853 F. Supp. at 1463-66.

On appeal, the government does not argue for application of Swift's "grievous wrong" standard, but it rejects the Rufo test as too lenient in antitrust cases. The government contends that the controlling standard for antitrust cases is set forth in United States v. United Shoe Machinery Corp., 391 U.S. 244, 88 S. Ct. 1496, 20 L. Ed. 2d 562 (1968), in which the Court explained that the "grievous wrong" standard of Swift was, in large part, the product of the peculiar circumstances of the case: a defendant's attempt to escape the impact of a consent decree a short time after its entry, where the same dangers of monopoly and unfair restraint of trade still were present. Id. at 248, 88 S.Ct. at 1499; see also Rufo, 502 U.S. at 379-80, 112 S.Ct. at 757-58. Looking beyond the "grievous wrong" language, the United Shoe Court explained Swift as teaching that an antitrust decree

may be changed upon an appropriate showing, and ... that it may not be changed in the interests of the defendants if the purposes of the litigation as incorporated in the decree (the elimination of monopoly and restrictive practices) have not been fully achieved.

391 U.S. at 248, 88 S. Ct. at 1499. The government therefore argues that Kodak should not be relieved from the 1921 and 1954 Decrees absent a showing that the purposes of the decrees have been fully achieved.

In response, Kodak argues that the Court's recent decision in Rufo is controlling. The Rufo case involved attempts by state and county officials to modify a consent decree governing conditions in a correctional institution. In rejecting the district court's reliance on the "grievous wrong" standard of Swift, the Court recognized that Federal Rule of Civil Procedure 60(b)(5), "in providing that ... a party may be relieved from a final judgement or decree where it is no longer equitable that the judgment have prospective application, permits a less stringent, more flexible standard" for modifying or terminating a consent decree. 502 U.S. at 380, 112 S.Ct. at 758. Specifically addressing the case "when a party seeks modification of a ... consent decree that arguably relates to vindication of a constitutional right," id. at 383 n. 7, 112 S. Ct. at 760 n. 7, the Court held that a party seeking to modify or terminate a consent decree "bears the burden of establishing that a significant change in circumstances warrants revision of the decree," id. at 383, 112 S.Ct. at 760.3 In formulating this standard, the Court drew upon a number of cases from this court. See, e.g., New York State Ass'n for Retarded Children, Inc. v. Carey, 706 F.2d 956, 969 (2d Cir.), cert. denied, 464 U.S. 915, 104 S. Ct. 277, 78 L. Ed. 2d 257 (1983); King-Seeley Thermos Co. v. Aladdin Indus., Inc., 418 F.2d 31, 34 (2d Cir.1969).

Although the Rufo Court seemed to limit the "significant change in circumstances" test to the type of decree presented in that case--one relating to the vindication of constitutional rights--this court has, on two occasions, applied Rufo in other contexts. First, in Still's Pharmacy, Inc. v. Cuomo, 981 F.2d 632, 636-37 (2d Cir.1992), we applied the Rufo standard to a consent decree relating to New York State's compliance with federal Medicaid regulations. Our most recent case to address the applicability of the Rufo standard is Patterson v. Newspaper & Mail Deliverers' Union, 13 F.3d 33 (2d Cir.1993), cert. denied, --- U.S. ----, 115 S. Ct. 58, 130 L. Ed. 2d 16 (1994), where an action had been brought by a union and its employees to set aside an affirmative-action consent decree. We held

that the flexible standard outlined in [Board of Educ. of Oklahoma City v. Dowell, 498 U.S. 237, 247, 111 S. Ct. 630, 636-37, 112 L. Ed. 2d 715 (1991) ] and Rufo is not limited to cases in which institutional reform is achieved in litigation brought directly against a governmental entity.... If a decree seeks pervasive change in long-established practices affecting a large number of people, and the changes are sought to vindicate significant rights of a public nature, it is appropriate to apply a flexible standard....

Id. at 38. Applying this standard, we upheld an order of a district court vacating a consent decree where, although not every provision of the decree had been satisfied, we were satisfied that the decree had served its primary purpose. Id. at 39.

In this case, the parties devote a considerable portion of their briefs to the issue of whether Rufo or United Shoe is controlling in the context of an antitrust consent decree. We think that both cases bear on the issue at hand. Rufo teaches that the power of a court to modify or terminate a consent decree is, at bottom, guided by equitable considerations. Indeed, Federal Rule of Civil Procedure 60(b)(5) makes no exception for antitrust decrees. However, we also believe that United Shoe provides a useful starting point for evaluating an antitrust defendant's request to modify or terminate a consent decree. In most cases, the antitrust defendant should be prepared to demonstrate that the basic purposes of the consent decrees--the elimination of monopoly and unduly restrictive practices--have been achieved. See United Shoe, 391 U.S. at 248, 88 S. Ct. at 1499.

Contrary to Kodak's arguments, we see nothing in Rufo that undermines the vitality of this approach. The decision in United Shoe was cited favorably by the Court in Rufo, 502 U.S. at 379-80, 112 S.Ct. at 757-58, and it was both cited and applied one year earlier in Board of Educ. of Oklahoma City v. Dowell, 498 U.S. 237, 247, 111 S. Ct. 630, 636-37, 112 L. Ed. 2d 715 (1991) (concluding that the termination of a decree was warranted where "the purposes of the desegregation litigation had been fully achieved"). The United Shoe standard promotes adherence to settlement agreements voluntarily entered into by parties to a litigation and ensures that consent decrees are not so easily modifiable as to discourage parties from reaching constructive settlements. See Patterson, 13 F.3d at 38.

Of course, cases may arise in which modification or termination of a consent decree is appropriate even though the purpose of the decree has not been achieved. For example, there may be significant changes in the factual or legal climate. See Rufo, 502 U.S. at 383, 112 S.Ct. at 759-60. Cf. United States v. Western Elec. Co., 46 F.3d 1198, 1207 (D.C.Cir.1995) (holding that unanticipated developments in telecommunications industry warranted waiver of consent decree provision). However, as a general matter, we believe that an antitrust defendant should not be relieved of the restrictions that it voluntarily accepted until the purpose of the decree has been substantially effectuated, or when time and experience demonstrate that the decree "is not properly adapted to accomplishing its purposes." King-Seeley, 418 F.2d at 35.

In view of the foregoing, we reject the government's contention that the district court misapplied the legal standards governing termination of an antitrust decree. The district court properly rejected the government's reliance on Swift and recognized that parties should be released from consent decrees when it is no longer equitable to enforce the decrees. In exercising its equitable discretion to terminate the 1921 and 1954 Decrees, the district court required Kodak to prove that: (1) it no longer possesses market power over film and photofinishing, and therefore that the primary purposes of the decrees--the elimination of monopoly and unduly restrictive practices--have been achieved; and (2) termination of the consent decrees would benefit consumers. We conclude that, under the legal standards established by Rufo and United Shoe, the district court's resolution of these issues in Kodak's favor provided a proper basis for the court's decision to terminate the consent decrees.

The district court found that the relevant geographic market for film is world-wide. The government contends that this determination is clearly erroneous and that the relevant geographic market in this case should be limited to the United States. In contending that the district court's world-wide market definition is overbroad, the government argues that the district court's own factual findings and certain items of undisputed evidence provide direct empirical proof that Kodak exercises market power in the United States. This, the government argues, is inconsistent with the world-wide geographic market found by the district court and weighs heavily in favor of a more narrowly defined geographic market, the United States.

a. The District Court Decision

The district court defined the relevant geographic market as the " 'area of effective competition in which the seller operates, and to which the purchaser can practicably turn for supplies.' " 853 F. Supp. at 1468 (quoting Tampa Electric Co. v. Nashville Coal Co., 365 U.S. 320, 327, 81 S. Ct. 623, 628, 5 L. Ed. 2d 580 (1961)). Applying this definition, the court found that the market for film is world-wide. In so doing, the court relied on findings that "each [film] manufacturer produces film for world-wide distribution using generally the same emulsions, and that the films are of comparable quality," id. at 1469, that foreign film manufacturers have excess production capacity, and that the supply of film is "elastic," meaning that foreign film manufacturers quickly could increase the supply of film for consumption in the United States if Kodak attempted to increase prices and reduce output. See id. at 1468.

In defining the geographic market, the district court also relied on two journal articles that address the process of defining the relevant geographic market in antitrust cases: William M. Landes & Richard A. Posner, Market Power in Antitrust Cases, 94 Harv.L.Rev. 937 (1981) (hereinafter "Landes & Posner"), and Kenneth G. Elzinga and Thomas F. Hogarty, The Problem of Geographic Market Delineation in Antimerger Suits, 18 Antitrust Bulletin 45 (1973) (hereinafter "Elzinga & Hogarty"). Landes & Posner advocate a general rule that, "if a distant seller has some sales in a local market, all its sales, wherever made, should be considered a part of that local market for purposes of computing the market share of a local seller." Landes & Posner, supra, at 963. They reason that a foreign seller with a United States presence has proved its ability to sell in this market--overcoming such factors as higher transportation costs as well as political and economic barriers--and could increase its sales here should local prices rise. See id. at 963-66. Elzinga and Hogarty advocate a more empirical approach to defining markets, contending that markets properly are defined by determining the areas from which, and to which, a significant amount of the relevant product is shipped. See Elzinga & Hogarty, supra, at 73-75. This approach is sometimes referred to as the "shipments" or the "imports/exports" approach.

The district court found that application of these two approaches also supported the definition of the relevant market as the world, rather than the United States. In particular, the court relied on the testimony of Kodak's economics expert, Professor Hausman, who opined that the Landes & Posner test was satisfied because: (1) each of the foreign film manufacturers had proven its ability to establish market share in the United States; (2) the foreign film manufacturers have excess production capacity; and (3) the supply of film is elastic. See 853 F. Supp. at 1468. With respect to the Elzinga & Hogarty approach, Professor Hausman testified that he applied the imports/exports analysis and determined that the world-wide flow of imports and exports of film were "significant," also leading to the conclusion that the relevant geographic market was world-wide. Id.

The district court rejected the government's contention that Professor Hausman's economic analysis was tainted by the so-called Cellophane4 fallacy. The issue in Cellophane was whether du Pont had monopoly power over the sale of cellophane, and, in resolving this issue, the Court defined the relevant product market as comprising "flexible wrapping materials." In defining the product market, the Court relied heavily on the high cross-elasticity of demand between cellophane and other wrapping materials, such as wax paper. This decision was criticized by many commentators because it failed to recognize that a monopolist--i.e., a seller that has reduced output and raised prices--always faces a highly elastic demand; its products are so overpriced that even inferior substitutes begin to look good to consumers. See, e.g., Landes and Posner, supra, at 960-61. In this case, the government argued that foreign film is an adequate substitute for Kodak film only because Kodak's film prices in this country are excessive. The district court found the government's Cellophane argument to be misplaced, because foreign film is an excellent substitute for Kodak film, whereas the other wrapping products discussed in Cellophane were inferior for many applications. See 853 F. Supp. at 1469-70.

While conceding that foreign competition should be taken into account in defining markets, the government argued that Kodak's ability to engage in price discrimination against this country's consumers directly shows that Kodak is exercising market power in the United States, and, therefore, that the world-wide geographic market proposed by Kodak is too broad. In support of this contention, the government noted that Kodak's average wholesale prices were lower in Europe (where Kodak has a 43-percent market share) than in the United States (where Kodak has a 67-percent market share), and its prices are even lower in Japan (where Kodak has less than a ten-percent market share) than in Europe. The district court rejected the government's price discrimination argument on two grounds. First, the court noted that the data upon which the government relied represented Kodak's prices at only a single point in time. Id. at 1475. This, even the government's expert acknowledged, is not necessarily evidence of systematic price discrimination. Id. The court also found that differing currency exchange rates and costs of distribution in different parts of the world made the government's price comparisons of "limited evidentiary value." Id.

The district court also rejected the government's argument that the "premium" price that Kodak film commands in the United States was evidence of market power in this country. The court found that Kodak film was priced 4.5 percent higher than Fuji film in the smaller retail outlets for film, and that this price premium has declined over time. Id. at 1474. This small difference in pricing, the court found, "results not from an intention on Kodak's part to extract a premium price for its film, but from a purposeful marketing strategy on Fuji's part to undercut whatever price Kodak is charging for its film, which Fuji considers to be the industry standard." Id. Also contributing to the price premium is "the perceived quality difference that exists in the minds of consumers who are satisfied with Kodak products." Id. at 1477.

Having rejected the government's arguments that Kodak engages in price discrimination and that Kodak film is priced at monopolistic levels, the district court stood by its definition of a world-wide market.

b. Definition of the Relevant Market

In determining whether a firm possesses market power, "the first step in a court's analysis must be a definition of the relevant markets." Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263, 268 (2d Cir.1979), cert. denied, 444 U.S. 1093, 100 S. Ct. 1061, 62 L. Ed. 2d 783 (1980). Without a definition of the relevant market, there is no way to measure a company's ability to act as a monopolist. Walker Process Equip., Inc. v. Food Machinery & Chem. Corp., 382 U.S. 172, 177, 86 S. Ct. 347, 350, 15 L. Ed. 2d 247 (1965). The relevant product market here--amateur color negative photographic film--is not in dispute. Kodak and the government disagree, however, as to the relevant geographic market. As the district court properly noted, the relevant geographic market is defined as the "area of effective competition ... in which the seller operates, and to which the purchaser can practicably turn for supplies." Tampa Electric, 365 U.S. at 327, 81 S.Ct. at 628. The district court applied this definition and found that an examination of the elasticities of supply and demand supported a world-wide geographic market for film. We see no error in the district court's analysis.

Looking first to the area where film sellers operate, we find strong support for the district court's determination that film sellers operate on a world-wide scale. Kodak is the only domestic film manufacturer. Of its four competitors, all of which sell film in the United States, "Fuji manufactures film in Japan and in the Netherlands, Konica manufactures film in Japan, Agfa manufactures film in Germany, and 3M manufactures film in Italy." 853 F. Supp. at 1468. Since Kodak has two-thirds of the United States film market, a full one-third of all film used in the United States comes from abroad. Additionally, the district court accepted Professor Hausman's opinion that "foreign manufacturers could quickly increase the supply of film for U.S. consumption if Kodak attempted to restrict output and raise prices." Id. Moreover, the flow of imported film has been continuous and systematic through the entire United States, and there is no evidence in the record of transportation costs or tariffs that put imported film at a significant cost disadvantage. The foregoing provides an ample basis for the district court's finding that foreign film producers act as a check on Kodak's ability to raise domestic prices.

Looking to the area in which film purchasers turn for supplies, the district court found that film purchasers are, for the most part, price sensitive and will shift among Fuji, Kodak, and private label film on the basis of changes in price. Id. at 1473. Thus, while many consumers state a preference for the familiar Kodak brand name, the empirical evidence of what consumers actually do indicates that consumers find non-Kodak film to be an acceptable substitute. The increasing price sensitivity of consumers is evidenced by the fact that private label film has become the fastest growing segment of the film market, with at least a 12.8-percent share of the market. Id. at 1473. Also, as noted above, one-third of all the film purchased in the United States is manufactured abroad. These findings amply support the district court's factual determination that a significant segment of United States film purchasers turn to foreign sources of film.

The government, however, contends that, by relying on the significant cross-elasticity of demand between Kodak film and other brands, the district court fell victim to the Cellophane fallacy. It contends that, because Kodak already is pricing its products at monopolistic levels in the United States, consumers' willingness to switch to other brands of film when the price of Kodak film rises actually demonstrates that Kodak possesses market power in the United States. We disagree.

The economic error allegedly committed by the Court in Cellophane was in failing to recognize that a high cross-elasticity of demand may, in some cases, be the product of monopoly power rather than a belief on the part of consumers that the products are good substitutes for one another. As the district court succinctly stated: "At a high enough price, even poor substitutes look good to the consumer." 853 F. Supp. at 1469. That is, in the Cellophane case, the high cross-elasticity between cellophane and wax paper simply may have been a function of the high price that du Pont demanded for cellophane. This case, however, does not involve a comparison of two highly-differentiated products like cellophane and wax paper. The district court found, and we agree, that the film produced by Kodak's competitors is of comparable quality to Kodak's film and is an excellent substitute. Moreover, the government's contention assumes that Kodak film is priced well above competitive levels, and we do not believe that the small but declining price premium that Kodak obtains for its film bears out this assumption.5 Under these circumstances, we do not believe that the district court fell victim to the Cellophane fallacy.

Based on the foregoing evidence that foreign film sellers operate as a check on Kodak's ability to raise prices and that a substantial segment of United States film purchasers turn to foreign sources of film, we agree with the district court that the "area of effective competition" for film encompasses the entire world.

c. Empirical Challenges to a World-Wide Market

While conceding that "foreign competition should be taken into account in defining markets, and that the prospect of foreign firms increasing their sales into the United States may sometimes prevent American firms from maintaining prices above competitive levels," the government argues that the world-wide geographic market found by the district court is overly broad and that the relevant geographic market should be limited to the United States. In making this argument, the government relies on certain empirical evidence before the district court that allegedly establishes that Kodak exercises market power over film in the United States. In particular, the government relies on the following "direct proof" of Kodak's market power in the United States: (1) Kodak's ability to engage in price discrimination against United States customers; (2) Kodak's ability to sell film no better than the film sold by its rivals at a substantial price premium; and (3) Kodak's excessive profits, as evidenced by an "own elasticity" of two. Kodak's ability to exercise market power in this country, the government argues, conflicts with the world-wide market definition that resulted from examination of the elasticities of supply and demand for film. For this reason, the government asserts, the relevant geographic market should be limited to the United States, an area in which Kodak can and does exercise market power. We address each of the three types of direct evidence of market power offered by the government in turn.

i. Price Discrimination

The government first argues that Kodak's ability to engage in geographical price discrimination against its United States customers establishes that the world-wide market definition adopted by the district court is overly broad. It is the position of the government that Kodak charges a higher price for its film in the United States than it charges for the same film in other parts of the world. This, the government argues, is proof that Kodak exercises market power in the United States. The district court disagreed, finding that the factual record did not support the government's price-discrimination theory. 853 F. Supp. at 1474-75. In response, the government further argues that the district court compounded its error by "suggest[ing] that the government had not carried [its] burden of persuasion" on this point.

Support for the government's theory that a market can be defined on the basis of price discrimination can be found in the 1992 Department of Justice and Federal Trade Commission Horizontal Merger Guidelines ("the Merger Guidelines"). The Merger Guidelines provide that a geographic market generally will be defined as a region in which a "hypothetical monopolist" could exact a " 'small but significant and nontransitory' increase in price, holding constant the terms of sale for all products produced elsewhere." Merger Guidelines Sec. 1.12. While this standard fully encompasses foreign competition and can result in the definition of a world-wide market, it applies only in the absence of "geographic price discrimination." Id. Sec. 1.22. If the company is capable of geographic price discrimination, then smaller geographic markets, defined by the regions in which the company is able to price discriminate, will be recognized. See id. Thus, the Merger Guidelines recognize that a geographic market is limited to the area in which prices are relatively uniform.

The Supreme Court's recent decision in Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 451, 112 S. Ct. 2072, 119 L. Ed. 2d 265 (1992), also recognizes that a company's ability to price discriminate against certain classes of consumers may be evidence of power in a particular market. In that case, the Court rejected Kodak's theory that its lack of market power in the photocopier and micrographics market precluded, as a matter of law, the possibility that it could exercise market power in derivative aftermarkets, i.e., parts and service for equipment that it sold. Id. at 455, 475-78, 112 S. Ct. at 2076, 2086-88. In rejecting Kodak's theory, the court relied, in part, on the possibility that Kodak could price-discriminate against various classes of consumers. Id. at 475-78, 112 S.Ct. at 2086-88; see also U.S. Anchor Mfg., Inc. v. Rule Indus., Inc., 7 F.3d 986, 997 (11th Cir.1993) (relying on "unusual circumstance of severe price discrimination against a distinct group of consumers based solely on brand preference"), cert. denied, --- U.S. ----, 114 S. Ct. 2710, 129 L. Ed. 2d 837 (1994).6

Even if we accept the government's theory that the relevant geographic market may be defined on the basis of price discrimination, there simply is no probative evidence in the record to support the assertion that Kodak engages in geographic price discrimination. To support its argument, the government relies upon pricing data supplied by Kodak in response to an interrogatory. This data indicates that the price of Kodak film is subject to regional variations. However, evidence that Kodak film sells for different prices in different parts of the world is insufficient to establish price discrimination without proof that Kodak's costs are uniform throughout the world. See Areeda et al., supra, Sec. 522, at 125. While there was no evidence in the record that Kodak's costs are the same in each region, there is evidence that Kodak's marketing and distribution systems differ markedly from country to country. Even the government's expert conceded that such variations would have to be "factor[ed] ... in" to determine whether there was price discrimination. Further, as the district court noted, fluctuations in currency exchange rates also are relevant to this inquiry. See 853 F. Supp. at 1475. Absent proof that Kodak's costs are uniform throughout the world, the government's price-discrimination argument simply is conjecture.

We also reject the government's argument that the district court improperly placed upon it the burden of persuasion on this point. In this proceeding, Kodak had the burden of proving that it no longer possessed market power over film in the relevant geographic market. To discharge its burden of establishing that the market for film is a world-wide one, Kodak put forth evidence that foreign film producers act as a check on its ability to raise prices and that consumers in the United States look to foreign sources of film. As discussed above, this evidence provides a sufficient basis for the definition of a world-wide market for film. In rebuttal, the government argued that the market definition proposed by Kodak was overbroad in view of, inter alia, Kodak's ability to price discriminate against its United States customers. Having chosen to rebut Kodak's proposed market definition by relying on its theory of price discrimination, it was incumbent upon the government to produce probative evidence of systematic price discrimination, which it failed to do. In the absence of such proof, the district court properly rejected the government's theory.

i. Premium Pricing and Consumer Perception

In a similar vein, the government argues that the district court erred in failing to appreciate the significance of United States customers' strong preference for Kodak film and the resulting premium price that Kodak is able to obtain for its film in this country. This, the government contends, demonstrates that Kodak possesses market power over film in the United States and, therefore, that the relevant geographic market should be limited to the United States.

In addressing Kodak's argument regarding consumer loyalty, the district court recognized that "50 percent of [surveyed] consumers will only buy Kodak film, while another 40 percent of consumers prefer Kodak film, but are willing to purchase another brand of film." 853 F. Supp. at 1475 (footnote omitted). The court attributed this loyalty "to the fact that Kodak has earned consumer trust by producing a product the consumer perceives to be superior." Id. at 1477. With respect to the government's contention that Kodak sells for a monopolistic premium, the district court found that, in mass merchandisers, which account for approximately one-half of all film sales, Kodak and Fuji film are priced within one percent of each other; on occasions, Fuji film is the higher priced, while at other times, Kodak film is more expensive. Id. at 1474. In smaller food and drug stores, which account for the vast majority of other sales of film, the court found evidence of an average price difference of 4.5 percent. The court determined that these small price differences resulted from Fuji's policy of undercutting Kodak's price in order to make its product more attractive to retailers as well as from consumers' preference for Kodak film.

The government argues here, as it did below, that Kodak's ability to sell film of similar quality as its competitors for a price premium while maintaining a very high market share is strong empirical evidence of Kodak's market power within the United States, and that the district court erred in failing to recognize this. In support of this argument, the government relies on the 4.5-percent retail price premium charged in the smaller outlets, and the ten-percent premium that Kodak film allegedly commands at the wholesale level. Since Fuji's strategy is to undercut Kodak's wholesale prices, the government argues, "it is all the more significant that Kodak nevertheless maintains a dominance of 67% to 10% over Fuji in U.S. sales." With regard to consumers' preference to Kodak film, the government argues that the issue is not how Kodak obtained its market power--whether as a result of its illegal monopoly ninety years ago or through a perceived difference in quality resulting from Kodak's leadership in the photography field--but only whether Kodak has market power.

The government's arguments here are not without some force. While Fuji may be Kodak's equal on a world-wide basis, Kodak's dominant position in the United States cannot seriously be disputed. However, the evidence before the district court demonstrates that the difference in price between Kodak film and that of its competitors is small and declining. The district court found, and the government does not dispute, that one-half of Kodak's film is sold at virtually no premium and the rest is sold at an average premium of less than five percent. These price differences are not of the magnitude to raise significant market power concerns. Moreover, the declining nature of this premium is evident from the fact that, only five years ago, Kodak film fetched a 6.3-percent price premium at mass merchandisers, and this premium has since fallen to less than one percent. See 853 F. Supp. at 1474. This supports the district court's finding that the premium paid for Kodak film is dissipating rapidly as its competitors' favorable reputations continue to grow. Under these circumstances, we believe that the small and declining price premium paid for Kodak film in the United States was not sufficient to call in question the district court's definition of a world-wide market.

For similar reasons, we must reject the government's contention that the strong preference of United States customers for Kodak film demonstrates Kodak's market power in the United States. Although consumers do state a strong preference for Kodak film when surveyed, empirical evidence introduced at trial by Kodak demonstrates, and the district court found, that "film purchasers are ... price sensitive, and will shift between Kodak, Fuji and private label film on the basis of changes in price." 853 F. Supp. at 1473. In this regard, the district court accepted Professor Hausman's conclusion that if Kodak were to raise prices by five percent, it would lose ten percent of its film sales. Accepting these factual findings--as we must, since they are well supported by the record--in conjunction with the small price premium for Kodak film, we cannot say that the district court erred in rejecting the government's argument that the preference of American consumers gives Kodak market power over the sale of film in the United States. Accordingly, this evidence is not inconsistent with a world-wide market definition.

iii. "Own Elasticity of Demand"

The last item of direct evidence relied upon by the government to support a finding that Kodak has market power within the United States is the fact that Kodak's "own elasticity of demand" is two. According to the government, an "own elasticity" of two indicates that Kodak is earning excessive profits from its film. This, the government argues, is strong evidence that Kodak exercises market power in the United States and weighs against a world-wide market definition.

The economist's term "own elasticity of demand" expresses the change in quantity of goods a firm will sell in response to a change in the price it charges for the goods. For a firm operating with an own elasticity of two, a price increase by the firm would result in its sales dropping by twice the percentage of the price change. Kodak's expert, Professor Hausman, opined that Kodak film is subject to an own elasticity of two, and, therefore, that Kodak would lose ten percent of film sales if it were to raise the price of its film by five percent.

The significance of an own elasticity of two, in the government's view, is that it indicates that the sales price of Kodak film is twice the short-run marginal cost. This conclusion follows from the "Lerner index," a mathematical formula expressing the relationship between a firm's own elasticity and the marginal cost of the goods sold. See Landes & Posner, supra, at 939-40. From this, the government argues that Kodak is earning monopolistic profits from the sale of its film. Thus, from Kodak's own elasticity of two, the government concludes that Kodak is exercising significant market power in the United States.

We are unpersuaded by this argument for a number of reasons. First, the government contends that the Lerner index applies to Kodak as the seller of a differentiated product that sets prices independently of its rivals. But, the contention that Kodak film is differentiated from the film sold by its rivals is directly contradicted by the district court's findings with respect to the elasticities of supply and demand for film. Accordingly, we believe that the use of the Lerner index in this manner rests on unwarranted factual assumptions.

Second, even if we were to accept the government's contention that Kodak's short-run marginal costs equal one-half of the product's sales price, we do not think that it necessarily follows that Kodak is earning monopolistic profits. Certain deviations between marginal cost and price, such as those resulting from high fixed costs, are not evidence of market power. See Landes & Posner, supra, at 939. In this case, there was overwhelming evidence that Kodak's film business is subject to enormous expenses that are not reflected in its short-run marginal costs. In particular, Professor Hausman testified that (1) Kodak uses between eight and nine percent of its revenues for research and development, and (2) Kodak's film is produced at plants costing hundreds of millions of dollars. See 853 F. Supp. at 1473. The government's own witnesses and submissions confirm that the fixed costs in the film industry are huge. In light of these considerations, we do not believe that the district court erred in rejecting the government's argument that Kodak's own elasticity of two proves that Kodak is exercising market power in the United States. Even if the government's evidence in this regard supported an inference of market power, and we are not persuaded that it does, it certainly did not compel such an inference. Accordingly, we do not believe that the government's "own elasticity" argument calls in question the district court's determination that the market for film is world-wide.

d. Conclusions as to the Relevant Geographic Market

To prevail on this appeal, the government was obliged to show that the district court abused its discretion in defining the geographic market for film as a world-wide one. We conclude that the government has failed to make such a showing. The district court applied the correct legal principles, and its factual findings relied upon herein are well supported by the evidence in the record. While another fact-finder might have weighed the evidence differently, our review of the district court's decision is limited to ensuring that the district court exercised its broad discretion in a proper manner. We are convinced that it did so, and we therefore affirm its definition of a world-wide geographic market.

Having upheld the district court's world-wide market definition, we also must uphold its determination that Kodak lacks market power over film. On a world-wide basis, Kodak has only a 36-percent share of the market for film. This low market share, in connection with the other factors addressed in the district court's opinion, see 853 F. Supp. at 1471-72, strongly supports the finding that Kodak no longer possesses market power. Since the government concedes that the restrictions set forth in the 1921 Decree should not be maintained if we affirm the district court's finding that Kodak lacks market power, we therefore affirm the district court's decision to terminate the 1921 Decree. With respect to the 1954 Decree, which prohibits Kodak from bundling its film with photofinishing, the government does not challenge the district court's finding that Kodak lacks market power over photofinishing. Given that Kodak lacks market power over both film and photofinishing, we see no abuse of discretion in the district court's decision to terminate the 1954 Decree.CONCLUSION

For the foregoing reasons, the order of the district court is affirmed.