Full opinion text
MEMORANDUM OPINION AND ORDER JOYCE HENS GREEN, District Judge. This matter now comes before the Court on the motion of plaintiff Federal Trade Commission (“FTC”) for a preliminary injunction to prevent the merger of defendants Owens-Illinois, Inc. (“Owens-Illinois”) and Brockway, Inc. (“Brockway”), two of the leading glass container manufacturing companies, until the conclusion of administrative proceedings by the FTC to determine whether the proposed acquisition would substantially lessen competition in violation of Section 7 of the Clayton Act, 15 U.S.C. § 18, and Section 5 of the Federal Trade Commission Act, 15 U.S.C. § 45. In consideration of the motion, defendants' opposition thereto, the three-day hearing before the Court, and for the reasons set forth below, plaintiff's motion is denied. I. FACTUAL BACKGROUND A. The Proposed Acquisition Owens-Illinois, a privately held corporation based in Ohio, produces a variety of packaging products, including glass containers (for soft drinks, beer, wine, liquor, food, pharmaceuticals, and cosmetics), plastic containers and closures, forest products, and specialty glassware (including scientific and laboratory glassware). Owens-Illinois attributed 36% of its profit in 1986 to its sale of glass containers and 23% to plastics and closures. The company has fourteen glass plants and approximately twenty plastic container plants throughout the United States. In 1986, Owens-Illinois’s glass container sales were over $1 billion, making it the second largest domestic seller of glass containers. The principal business of Brockway, a publicly held corporation and Owens-Illinois’ main competitor in the glass container industry, is the manufacture of glass, plastic, and metal containers, as well as caps, lids, and closures for the packaging of consumer and industrial products, including food, beer, liquor, wine, soft drinks, toiletries, cosmetics, and pharmaceutical and proprietary products. Brockway is the third largest producer of glass containers in the United States, with annual sales in 1986 of $681 million. In 1986, Brockway derived approximately 64% of its revenue from the manufacture of glass containers, 22% from plastic, and 10% from metal. Brockway’s glass manufacturing is conducted at eleven plants nationwide, and it produces plastics at seventeen locations. On September 23, 1987, Owens-Illinois, through BI Acquisition Corporation, commenced a cash tender offer for all outstanding common shares of Brockway at $60 per share, for a total of about $750 million, plus an additional $110 million for expenses and debt retirement. The offer was scheduled to expire January 7, 1988, was then extended through February 19, 1988, and most recently through February 29, 1988. After learning of the proposed merger, the Federal Trade Commission began investigating the acquisition for potential violations of the Clayton Act. After receiving voluminous information from Owens-Illinois, the FTC voted 3-2 on November 18, 1987 to seek a temporary restraining order and a preliminary injunction to prevent consummation of the merger. On January 11, 1988, after this action had commenced, the FTC voted 4-1 to issue an administrative complaint against the defendants. B. The Glass Container Industry To place this case in its proper economic context, it is important to examine briefly here the dramatic technological changes during the past fifty years in the development of packaging materials, including glass, plastic, metal, and paper. Prior to World War II, most foods and beverages were sold in glass containers and metal cans, but today a wide variety of packaging is offered to container purchasers and to consumers. The complete conversion of three products in particular exemplifies the dynamic changes that the packaging industry is capable of initiating and responding to. Milk, packaged in glass bottles for many years, is now sold almost entirely in paperboard containers. In the larger sizes, milk is also being packaged in plastic. The next development may be aseptic packaging that would give milk shelf-stability sufficient to alleviate the need for store refrigeration. Another dramatic historical shift in packaging was the conversion of baby food in the 1950s and continuing through the 1970s from cans to glass. And, motivated by consumer preference for nonbreakable containers, bleach containers rapidly converted in the early 1960s from glass to plastic containers. As discussed further below, two of the most recent developments in the packaging industry have been the introduction and growing use of plastic containers with high-barrier properties, particularly the “PET” container introduced in 1977-78, which offer many of the same advantages traditionally unique to glass — such as impermeability to moisture and oxygen, clarity, and heat resistance — in addition to providing package purchasers and consumers the primary attractions of plastic — non-breakability and lighter weight. Customer preference for plastic packaging of some products traditionally packaged in glass continues to spur conversions in packaging. In addition to these trends away from glass, there is also some indication of growing self-manufacture by packagers, particularly with plastic and paper aseptic containers. Based on these trends in packaging, the plastics division of Owens-Illinois has targeted glass containers as the primary opportunity for growth in plastics, while the glass division of Owens-Illinois seeks to erode the predominance of the metal can. Because the production of glass has markedly declined over the past several years, Owens-Illinois characterizes the dynamics of this market as a “life and death struggle” between glass and other packaging materials. In addition to this brief sketch of particular developments in the packaging industry, it is important to examine for a moment the present position of defendants Owens-Illinois and Brockway in the glass container business, an industry which had sales in 1986 of almost $5 billion. Since 1980, the glass container industry has been undergoing significant consolidation. The number of firms in the industry has decreased from 26 to its current level of 18 and, during this same time, 29 glass plants have closed. The domestic glass container industry currently comprises six principal producers with four or more production facilities and a “fringe” of twelve other firms. Owens-Illinois, the second largest domestic producer of glass containers, had 22.2% of the 1986 furnace capacity and 22.7% of the dollar sales. Brockway is the third largest producer, with 14.9% of the furnace capacity and 15.0% of the dollar sales. If the proposed merger were to occur, Owens-Illinois would control approximately 37% of the glass container capacity, with the next largest producer, Anchor/Diamond-Bathurst, holding approximately 24%. Assuming the relevant market is “all glass containers,” as proposed by the FTC, the Herfindahl-Hirschman Index, an economic measure of market concentration, would increase by approximately 662 points to 2,129 based on capacity and by 681 points to 2,219 based on dollar sales, putting the glass container industry into what the Department of Justice considers a “highly concentrated” category where anti-competitive effects are presumed. On the other hand, if the market is viewed as including plastic, metal, and paper containers, as well as glass, then the FTC agrees that the anticompetitive effect of the merger is insignificant. II. STANDARDS FOR PRELIMINARY INJUNCTIVE RELIEF Under Section 13(b) of the Federal Trade Commission Act, 15 U.S.C. § 53(b), which effectuates the congressional intent expressed in section 7, 15 U.S.C. § 18, to “arrest the anticipated anticompeti-tive effects of acquisitions ... in their in-cipiency,” FTC v. Warner Communications, Inc., 742 F.2d 1156, 1160 (9th Cir.1984), the FTC may sue to enjoin a proposed merger [wjhenever the Commission has reason to believe— (1) that any person, partnership, or corporation is violating, or is about to violate, any provision of law enforced by the Federal Trade Commission, and (2) that the enjoining thereof pending the issuance of a complaint by the Commission and until such complaint is dismissed by the Commission or set aside by the court on review, or until the order of the Commission made thereon has become final, would be in the interest of the public.... Rather than the traditional four-part test for preliminary injunctive relief requiring proof of “irreparable harm,” Virginia Petroleum Jobbers Ass’n v. Federal Power Comm’n, 259 F.2d 921, 925 (D.C.Cir.1958), the applicable standard for relief in this action is the statutory “public interest” standard. FTC v. Exxon Corp., 636 F.2d 1336, 1343 (D.C.Cir.1980). The Court is required to determine whether “[ujpon a proper showing that, weighing the equities and considering the Commission’s likelihood of ultimate success, such action would be in the public interest.” 15 U.S.C. § 53(b). Because the request is only for interim relief until administrative proceedings are complete, the Court is “not to make a final determination on whether the proposed merger violates Section 7, but rather to make only a preliminary assessment of the merger’s impact on competition.” Warner Communications, Inc., 742 F.2d at 1162. While the courts possess broad authority to enjoin mergers when such action would serve the public interest in the effective enforcement of the antitrust laws, “it is well recognized that the issuance of a preliminary injunction prior to a full trial on the merits is an ‘extraordinary and drastic remedy’ ” because “as a result of the short life-span of most tender offers, the issuance of a preliminary injunction blocking an acquisition or merger may prevent the transaction from ever being consummated.” FTC v. Exxon Corp., 636 F.2d at 1343 (citing Medical Society v. Toia, 560 F.2d 535, 538 (2d Cir.1977)). Indeed, in this case, defendants have represented that the granting of preliminary relief to the FTC would “spell the doom” of the proposed merger. III. ANALYSIS A. Likelihood of Success on the Merits Section 7 of the Clayton Act, 15 U.S.C. § 18, prohibits mergers and acquisitions whose effect “in any line of commerce in any section of the country ... may be substantially to lessen competition, or to tend to create a monopoly.” To demonstrate a reasonable likelihood of success under Section 7, the Court must be persuaded that the Commission has “raised questions going to the merits so serious, substantial, difficult and doubtful as to make them fair ground for thorough investigation, study, deliberation and determination by the FTC in the first instance and ultimately by the Court of Appeals.” Warner Communications, Inc., 742 F.2d at 1162; FTC v. Beatrice Foods Co., 587 F.2d 1225, 1229 (D.C.Cir.1978). As discussed below, the Court does not so find. The FTC has the burden of proof in presenting this motion for a preliminary injunction to show a likelihood of success on the merits on two issues: whether the effect of the merger may be “substantially to lessen competition” and whether the issuance of an injunction would be in the public interest. Determining anticompeti-tive effects requires a case-specific analysis of the facts. Fruehauf Corp. v. F.T.C., 603 F.2d 345, 353 (2d Cir.1979). The first and most critical task is to define the “relevant product market.” Brown Shoe Co. v. United States, 370 U.S. 294, 320, 82 S.Ct. 1502, 1521, 8 L.Ed.2d 510 (1962); see also United States v. E.I. duPont de Nemours & Co., 353 U.S. 586, 593, 77 S.Ct. 872, 877, 1 L.Ed.2d 1057 (1957). Despite their differences of opinion on many other issues in this case, the parties agree that this is the determinative issue in this litigation. In large part, the instant case in a 1980s revival of the 1960s classic United States v. Continental Can, 378 U.S. 441, 84 S.Ct. 1738, 12 L.Ed.2d 953 (1964), except in the modern setting plastic jars and bottles have replaced metal cans as the leading competitor of glass containers. In Continental Can, the Supreme Court held that a merger between the second largest can producer and the third largest glass producer violated Section 7 of the Clayton Act, finding that interindustry competition between glass and metal containers warranted treating them as a relevant product market for all end uses for which they competed. The Court rejected the trial court’s “unduly narrow construction of the ‘competition’ protected by § 7 and of ‘reasonable interchangeability of use or the cross-elasticity of demand,’ ” reasoning that merely because glass and metal had distinctive characteristics did not “automatically remove them from the reach of section § 7.” 378 U.S. at 453, 84 S.Ct. at 1745. The Court further observed that “there is and has been a rather general confrontation between metal and glass containers and competition between them for the same end uses which is insistent, continuous, effective and quantitywise very substantial.” Id. In analyzing the competition between metal and glass, Continental Can looked specifically at the substitution of containers in certain end uses, the attempts by producers to expand their market shares, and that producers took the pricing of the competing container into account in marketing their own product. The structure, history, and probable future of the industries involved were also considered. Id. at 458, 84 S.Ct. at 1747. Determining that the relevant market was glass and metal, the Court concluded that “though the interchangeability of use may not be so complete and the cross-elasticity of demand not so immediate as in the case of most intrain-dustry mergers, there is over the long run the kind of customer response to innovation and other competitive stimuli that brings the competition between these two industries within § 7’s competitive-preserving proscriptions.” Id. at 455, 84 S.Ct. at 1746. Though Continental Can spoke directly to the industries at issue in this case, the Court’s conclusion that glass and metal compete in the relevant product market, is, of course, not the important lesson to be drawn from the case. The Court made clear that the pivotal question of relevant market can be determined only after close scrutiny of the facts in each case. Brown Shoe dictates as much: both cross-elasticities of demand and supply determine the “outer boundaries” of the production market, and the factors that may be considered in defining “submarkets” are “such practical indicia as industry or public recognition of the submarket as a separate economic entity, the product’s peculiar characteristics and uses, unique production facilities, distinct customers, distinct prices, sensitivity to price changes, and specialized vendors.” Id. 370 U.S. at 325, 82 S.Ct. at 1524. Warner Communications, 742 F.2d at 1163. There are many cases where courts have applied these basic principles and ultimately defined markets both broadly and narrowly, confirming the necessity for individualized examination of the facts presented. As the Court commented in Continental Can, “the [legal] guidelines offer no precise formula for judgment and they necessitate, rather than avoid, careful consideration based on the entire record.” 378 U.S. at 449, 84 S.Ct. at 1743. See also Columbia Metal Culvert Co., Inc. v. Kaiser Aluminum & Chemical Corp., 579 F.2d 20, 26-27 (3d Cir.), cert. denied, 439 U.S. 876, 99 S.Ct. 214, 58 L.Ed.2d 190 (1978) (observing that the “well-trodden trail illuminated by Chief Justice Warren in Brown Shoe ... has led to a variety of destinations”). Despite the necessity for a fresh examination of the state of competition in the packaging industry, however, Continental Can does provide useful guidance on the critical issue in this case — how extensive must actual competition be in various end use segments in order for competing packaging materials to be considered within the relevant product market. Particularly important to this case, the Court reasoned that glass and metal did not have to compete in all end uses, declaring that “complete interindustry competitive overlap need not be shown.” Id. 378 U.S. at 457, 84 S.Ct. at 1747. See also United States v. E.I. duPont de Nemours & Co., 351 U.S. 377, 395, 76 S.Ct. 994, 1007, 100 L.Ed. 1264 (1956) {“Cellophane”) (competitive effects within smaller parts of the market are not relevant). At the same time it defined the relevant product market broadly as glass and metal packaging, Continental Can declined to include all other packaging (plastic, paper, and foil) in the market, finding that the existence of the broader market “does not necessarily negative the existence of sub-markets of cans, glass, plastic or cans and glass together, for ‘within this broader market, well-defined submarkets may exist which, in themselves, constitute product markets for antitrust purposes.’ ” Id. 378 U.S. at 457-58, 84 S.Ct. at 1747 (citing Brown Shoe, 370 U.S. at 325, 82 S.Ct. at 1524). The reasoning of Continental Can remains valid today. The existence of a wide array of containers for products with equally varying characteristics does not necessarily mean that demand for certain containers is inelastic or that purchasers and consumers would not accept substitute packaging. If product differentiation were a controlling variable in the relevant product market analysis, then “one can theorize that we have monopolistic competition in every nonstandardized commodity with each manufacturer having power over the price and production of his own product.” Cellophane, 351 U.S. at 393, 76 S.Ct. at 1006. However, “this power ... is not the power that makes an illegal monopoly. Illegal power must be appraised in terms of the competitive market for the product.” Id. (emphasis added). This assessment, in turn, is made by examining elasticity of demand and supply. Id. The law is clear that products need not be fungible to be considered substitutes; they need only be “reasonably interchangeable by consumers for the same purposes.” Cellophane, 351 U.S. at 394-95, 76 S.Ct. at 1007; Brown Shoe, 370 U.S. at 325, 82 S.Ct. at 1524; see also Neumann v. Reinforced Earth Co., 786 F.2d 424, 429-30 (D.C.Cir.1986). Keeping these principles in mind, and the necessity for focusing on actual competition within an industry, the relevant product market in this case must be scrutinized by analyzing (1) the reasonable interchangeability of use between the glass containers and substitutes (“cross-elasticity of demand”), and (2) the reasonable interchangeability of production facilities (“cross-elasticity of supply”). 1. Cross-Elasticity of Demand In analyzing the scope of the relevant product market then, it must be determined whether glass containers compete with other types of containers in the packaging industry. The FTC bases its case on the premise that, for a significant number of users within eleven specific “end use segments” of the glass container market, there are not sufficiently acceptable substitutes to glass containers and that, therefore, a small but significant (defined as 5% to 10%) nontransitory increase in the price of glass would not cause purchasers to switch to alternative forms of packaging. Defendants paint an entirely different picture of the elasticity of demand for glass containers. They contend that there is substantial and vigorous competition among glass, plastic, metal cans, and other types of packaging containers, that purchasers faced with a supracompetitive increase in the price of glass would shift to alternative packaging, and that the allegedly inelastic end uses represent a clear minority of the glass container business. The following lengthy, but not exhaustive, discussion of the eleven end use segments should be viewed as the trees, not the forest, of this Opinion. The economic experts for both the FTC and defendants agree that the large majority of glass container purchasers do have substitutes for glass containers and that there is significant intermaterial competition. The eleven end use segments presented by the FTC constitute approximately $450 million in glass sales or only about 25.8% of the total glass container tonnage. Thus, in the vast majority (at least 75% even by the FTC’s own estimate) of end uses for glass containers, other packaging materials, including plastic, metal, and paper, compete directly and vigorously with glass. Furthermore, the FTC concedes, as it must, that not all end users within the eleven end use segments are committed to glass containers, and the FTC has not suggested that a market narrower than all glass containers would be appropriate. The inquiry could end here, since it is possible to conclude for these reasons that, even aggregated, the end use segments at issue, assuming arguendo they are indeed as inelastic as the FTC suggests, do not constitute a sufficient part of the glass market to allow a finding of substantial anticompetitive effect under Section 7. Continental Can, 378 U.S. at 457, 84 S.Ct. at 1747; Cellophane, 351 U.S. at 404, 76 S.Ct. at 1012. Nonetheless, a brief examination of each of those segments proposed by the FTC is undertaken both because of the importance of this case to the public and the litigants and because such an examination reveals the extensive present and future intermaterial competition in the glass and other packaging industries. This analysis serves to confirm the conclusion that those few end use segments proven to be inelastic are not significant enough, in and of themselves, to constitute a relevant product market and are not representative of the glass container market as a whole. a. Alleged Inelastic End Uses The eleven end use segments that the FTC contends contain significant purchasers whose demand for glass containers is inelastic are: baby food and baby juice; spaghetti sauce; mayonnaise; pickles; shelf-stable juices; jams and jellies; wine and wine coolers; distilled spirits; single-serve soft drinks; soluble beverage products; and scientific and lab glassware. In reviewing the record on these end use segments and the potential for substituta-bility of containers, certain characteristics of glass and its modem competitor plastic are recurrent themes. The FTC repeatedly commented that glass has a distinctive combination of characteristics not found together in the same combination in other types of containers: impermeability, retort-ability and ability to be “hot filled,” clarity, reclosability, inertness, rigidity, a “quality image,” microwaveability, and recycla-bility. On the other hand, defendants showed that modem plastic containers have many, if not all, of these same desirable characteristics as glass, as well as two major advantages over glass — nonbreaka-bility and lighter weight. As will be seen below, these differing characteristics of glass and plastic are of varying importance for individual end users and for different types of consumers within those end uses. The analysis below can only highlight certain portions of the extensive record in this case, consisting of approximately 15 volumes of exhibits and almost four days of testimony and argument. 1. Baby Food and Baby Juice The most inelastic end use for glass containers alleged by the FTC is packaging of baby food and baby juice. The vast bulk of baby food and juice is packaged in glass, constituting approximately $107 million in glass container sales in 1986. The essential testimony presented by the FTC on this end use segment was from the Beech-Nut Nutrition Corp., one of the three major producers of baby food and baby juice products. Beech-Nut uses only glass containers for its products and believes “there exists no clear, retortable plastic package that is cost-competitive with glass.” The company indicated that it would not switch to alternative packaging even in the event of a 10% increase in the price of glass containers. The leading baby food producer, Gerber, also uses only glass to package its baby food, but is currently test marketing a flavor of baby juice in a 4-oz. plastic container and marketing baby juice in a 750-ml. plastic container. Despite Gerber’s belief that consumers have a preference for plastic packaging of baby food and are willing to pay a premium for it (Gerber’s test market product in plastic costs 25% more than its glass counterpart reflecting that the containers themselves cost almost twice that of glass), Gerber nonetheless seems generally pessimistic about the potential for developing in the near future a container for baby food that is costrcompetitive and able to withstand thermal processing. Gerber also expressed its concerns about the competitive effects of the proposed merger. Although defendants acknowledge that plastic packaging has not yet penetrated the baby food and juice market to any significant degree, they contend that the advent of packaging alternatives is but a matter of time. For example, retortable plastic containers are already being used, in other end uses to varying degrees. Defendants’ industry expert Michael F.X. Gi-gliotti suggested that plastic and aseptic packaging of baby foods is attractive because they are “easy to use and unbreakable.” In addition to conducting their own research on alternative packaging, the baby food companies, like other glass container purchasers, are frequently courted by the major plastic companies attempting to convert them to new containers. Despite defendants’ optimism for the future of plastics, however, the record reflects that conversion to an acceptable and cost-competitive alternative container for baby food and baby juice is not likely within the next few years. In sum, while plastic containers are promising to make inroads in the baby food and juice container business, this end use appears firmly committed to glass for the near future. 2. Spaghetti Sauce Spaghetti sauce is marketed predominantly in glass containers and constituted about $115 million in glass container sales in 1986. The leading producer of spaghetti sauce, Ragú, and another major producer Prego (Campbell Soup Co.), each believe that there are currently no reasonably viable substitutes for glass containers and that alternative containers will not be available for several years. The retortable, high-barrier containers that exist and could be used for spaghetti sauce are not considered reasonable substitutes by the major producers because they are not clear and because the containers are “considerably” more costly than the currently used glass containers. Cans are an even less acceptable alternative to purchasers than is plastic. Some spaghetti sauce is packaged in metal cans, but neither Ragú nor Prego consider metal an acceptable substitute because cans are neither clear nor resealable. Pre-go would not switch to cans in the event of a 10% increase in the price of glass containers, and Ragu would resist such a conversion even if the price of glass increased by 20 to 25%. Owens-Illinois and Brockway are Ragu’s principal glass suppliers. Overall, the spaghetti sauce end use segment appears largely inelastic. But, the commitment to glass appears motivated primarily by factors — cost and perceived consumer preferences — that could well change in the near future. 3. Mayonnaise Glass is also the traditional packaging container for mayonnaise. Mayonnaise and other “spoonable dressings” accounted for $129 million in glass sales in 1986. Plastic containers appear to be a viable, but not yet accepted, substitute for glass in the consumer sizes of mayonnaise. The Best Foods Division of CPC International, Inc., producers of Heilman’s and Best Foods mayonnaise, prefers glass to plastic containers because glass is clear, “relatively economical,” preferred by consumers, less permeable to oxygen, and has historically been the accepted package for mayonnaise. Best Foods conducted a preliminary study of the possibility of using plastic packaging for its mayonnaise, but concluded that a clear and impermeable container would cost about 20% more than glass. The company states that it would not, therefore, switch to plastic containers in the event of a 10% increase in the price of glass. Best Foods buys its glass containers primarily from Owens-Illinois and Brockway. Despite this conclusion, Best Foods is continuing to study applications for plastic packaging. The record showed that the technology to produce alternative packaging for mayonnaise, including wide-mouth PET jars, is becoming available. Some producers of mayonnaise already use alternative packaging, including squeezable opaque plastic containers, plastic tubes, and wide-mouth clear plastic containers (marketed by Saffola), but these currently represent only a small part of the market. Although the trend toward plastic in the mayonnaise packaging business is clearly not a stampede, alternative packaging appears to be for more than simply “niche” products as the FTC contends. For instance, Brockway lost part of its mayonnaise container business when Kraft switched from glass to the 16-oz. squeezable plastic container. Although plastic may become a more prevalent container for mayonnaise in the future, the solid majority of the mayonnaise purchasers currently appear firmly committed to glass. 4. Pickles Pickles have also traditionally been packaged in glass and this end use segment accounted for $111 million in glass sales in 1986. Most pickles are hot-packed and would require high-barrier plastic containers that are not yet cost competitive with glass. The Campbell Soup Co., producer of Vlasic pickles, had indicated that it would not switch to plastic containers in the event of a 10% increase in the price of glass. As with mayonnaise, pickle producers see glass as more desirable because it is clear and not oxygen permeable. Once again, however, the record indicates that the real barrier to switching to plastics is not technological infeasibility, but price and purchasers’ perception of consumer preferences. In fact, about 10% of Campbell Soup’s Vlasic pickles are already packaged in plastic containers, other companies have plastic pickle jars in test-market, and growing use of plastic containers for pickles is predicted by industry observers. Demand for glass containers by pickle purchasers thus appears to be presently firm, but is beginning to soften toward at least partial acceptance of plastic packaging. 5. Shelf-Stable Juices Shelf-stable juices are currently packaged in a variety of containers. Glass may comprise less than half of the containers used for shelf-stable juices. Glass container sales in this category were approximately $260 million in 1986. Two producers of shelf-stable juices testified that plastic is not a cost-competitive substitute for glass containers, particularly in the smaller sizes. In the 16-oz. “single serving” size, a high-barrier plastic container for shelf stable juice would be about 40-50% more expensive than glass (not including the capital investment required for conversion). The cost differential between glass and plastic narrows with increasing size, however, particularly in the large (40-to 64-oz.) container sizes, and some major producers have started using plastic even though it is more expensive. For the past three years, Oceanspray has been packaging its cranberry juice in a 64-ounce plastic container and in a concentrated form in an aseptic package that produces 42 fluid ounces. On the other hand, Welch, one of Oceanspray’s competitors, itself had investigated the possibility of switching some of its juice products in the 64-ounce size to plastic, but found it too costly. The other major shelf-stable juice currently packaged in plastic in the 64-oz. size is Gatorade; for Quaker Oats, plastic is considered an “acceptable” alternative to glass because of its lighter weight. Some shelf-stable juices are currently packaged in metal cans, in sizes directed to particular types of consumers. For example, Welch packages in cans its juice in the 5.5-oz. size for children and lunch wagons, the 12-oz. size for vending machines, and the 46-qz. size for grocery stores and food service. Welch testified, however, that its can sales have been constantly declining over the past few years. Shelf-stable juices are also sold in aseptic packaging, but this packaging appears largely to be only in areas of children or “lunch box” consumption. Welch packages some of its juice in aseptics, even though it considers aseptic packaging to be “very slow and costly” and is not interested in self-manufacture. Thus, while both Welch and Quaker Oats indicated that they would not switch away from their current use of glass containers in the event of a 10% increase in the price, this opinion appears based on perceived preferences and sizing concerns rather than on concrete obstacles to conversion. Even these purchasers acknowledged that some of their own shelf-stable juices are packaged in metal, aseptic, and ■ plastic packing, and that their glass container products compete with products in alternative packaging. Both companies are currently investigating alternative packaging and observed that an increase in the cost of glass would prompt them to further their research. In contrast to the testimony of Welch and Quaker Oats, the largest producer of single-serve shelf-stable juice, New England Apple Co., stated that a 5 to 10% increase in the price of glass would prompt the company to shift to aluminum cans and would cause it immediately to consider substituting plastic for glass. Thus, to a significant extent, glass, plastic, and metal effectively compete in the shelf-stable juice end use segment. 6. Jams and Jellies Glass is the traditional container for jams and jellies, amounting to $66 million in glass container sales in 1986. The major jelly producers, including Smuckers, Kraft, and Welch Foods, Inc., currently use some “squeezable” plastic containers, but glass is still predominant. Similar to the concerns of mayonnaise and pickle producers, purchasers of glass containers for jams and jellies prefer glass because the product is hot-packed and plastic containers that can withstand this thermal processing cost more than glass. Producers also perceive a consumer preference for clear containers. Welch testified that it would not switch the rest of its jam and jelly products to plastic in the event of a 10% price increase in glass containers. Welch purchases the majority of its glass from Brockway and Owens-Illinois, but always keeps a third supplier qualified. Even among the producers who appear most committed to glass, however, some elasticity is evident. Consider that until three years ago, Welch Foods (which has about 13% of the jam and jelly market) used only glass containers, but now packages about 18% of its jellies in a plastic, squeezable, multi-barrier, hot-fillable, opaque container — even though the cost of the container is higher than a comparable size in glass. Welch considers the plastic container to be a “niche” product (i.e., attractive to a limited number of consumers because of its convenience); it nonetheless comprises a significant fraction of its sales. Clear plastic wide-mouth containers for jams and jellies have not yet been marketed, but similar containers have been developed at least on a limited basis for other hot-filled food products, such as tomato sauce and apple sauce. Notably, a significant amount of Welch’s research and development effort is spent investigating new plastic packaging for jams and jelly. Welch acknowledged that the plastic container for jams and jellies is a new technological innovation, and that as producers develop more skill in manufacturing plastic containers, costs will decline. The inelasticity of the end use segment for jellies and jams accordingly also appears to be based largely on perceived consumer preference and on the presently higher cost of plastic packaging. The success of plastic containers and consumer acceptance generally indicates, however, that these barriers may soon be overcome. 7. Wine and Wine Coolers Wine is sold primarily in glass bottles. The wine industry is distinct among the eleven end uses in its low reliance on the domestic glass container producers. About two-thirds of the glass wine bottles are made by the producers, and foreign imports are significant. In fact, Brockway does not even manufacture “green glass,” which represents about 75% of the wine industry’s total glass usage. In addition, while premium wines are bottled almost entirely in glass, an increasing amount of “popular wines” is being sold in cans, plastic, and “bag in the box” containers. Wine coolers, the fastest growing beverage in the United States and now a major end use for glass containers, are primarily sold in 12-oz. single-serve glass containers, but are also sold in metal, plastic, and “bag in the box" containers. As with shelf-stable beverages and soft-drinks, oxygen permeability is more of a problem in the smaller sizes, and high-barrier plastic bottles are more expensive than glass containers. In short, an examination of demand in this end use segment shows that demand for glass bottles for premium wine is inelastic, but self-manufacture and foreign imports make this end use only minimally reliant on the glass container industry. Furthermore, although glass is still the predominant container, a variety of packaging materials compete for a notable portion of the container business of non-premium wines and wine coolers. 8. Distilled Spirits The weakness of the FTC’s case with regard to the end use of distilled spirits is revealed by the fact that the Commission readily acknowledges that liquor producers have inelastic demand for glass containers only in certain package sizes and qualities of liquor. The FTC has shown that only the 750 ml., 1 liter, and the “higher priced brands” are not converting to plastic. One of the more significant and recent inroads for the PET plastic container has been for liquor. Some producers of national brands are beginning to convert entirely to plastic containers in certain sizes, primarily in the 1.75-liter size. Joseph E. Seagram & Sons, for example, whose glass purchases amounted to $51.5 million in 1987, has switched almost completely to plastic in the 50 ml. size and plans to convert 90% of its 1.75-liter liquor volume that is currently in glass to plastic. This conversion for Seagram alone represents a shift in packaging demand roughly the size of the baby food container demand and somewhat larger j than the spaghetti sauce market. In short, the penetration of plastic into the liquor market is significant, and even may soon be complete, in some areas. The producers’ desire to continue using glass in the remaining sizes is primarily motivated by quality and image concerns. Generally, then, this end use segment can be considered fairly elastic. Indeed, plaintiff’s expert, Dr. Steven Nelson, testified that intermaterial competition for liquor packaging is more intense than in any other of plaintiff’s allegedly inelastic end use segments. 9. Single-Serve Soft Drinks The FTC concedes that the demand for the vast majority of single-serve soft drink containers is elastic and producers would readily shift to alternative packaging in the event of a anticompetitive price increase in glass. One major national supermarket chain reported that 100% of the soft drinks they carry (occupying the most shelf space of any product in the store) is packaged in cans or plastic. For Owens-Illinois, the competition between glass, plastic, and metal soft drink bottles is intense. When a significant portion of the large sizes of soft drink containers converted to plastic in the late 1970s, Owens-Illinois experienced a 20% drop in demand for its glass containers. Owens-Illinois’ development of the single-serve “Plastishield” glass container for soft drinks (a lighter-weight glass bottle surrounded by a foam label) was a response to the competitive threat posed to glass containers by cans. But today, Owens-Illinois finds its Plastishield container competing against plastic containers; the company considers the Vz-liter PET “our single number one threat.” Many soft drink bottlers, including Coca-Cola, self-manufacture 16-oz. PET bottles. For soft drinks in sizes of 16-oz. and below (“single serve”), however, the predominant container is still glass, particularly for the “off” and “slow moving” brands, i.e., those other than Coca-Cola and Pepsi. It is on this inelastic minority (representing by the FTC’s own estimate only about 15% of the total demand for glass soft drink containers) that plaintiff grounds its case. Even for this “slower” part of the market, however, demand may not be as inelastic as plaintiff contends. Plastic has virtually replaced glass 16-oz. soft drink containers in some areas of the country, including Michigan, New York City, and the Carolinas. One bottler, for example, switched to plastic in the 16-oz. size because of a glass shortage, even though he preferred using glass bottles. A few of the smaller brands are, in fact, being marketed in plastic containers. And, “warehouse” brands (those brands, often private, that are stocked by grocery stores in their warehouses rather than put directly on the shelves by the producer), are generally packaged in cans or two-liter PET containers. Warehouse brands represent approximately 4% of the glass shipments for beverages. In addition to plastic, aluminum packaging for soft-drinks is popular in the 12-oz. size. While some purchasers, such as L. & A. Juice, maker of Fifth Avenue Seltzer, may strongly prefer glass for its “image,” it is clear that glass, plastic, and aluminum packaging actively compete in the soft-drink sector. Although glass soft-drink containers continue to command a large sector of the market, the record supports a more elastic characterization of the single-serve demand than that proffered by the FTC. Joseph Lemieux, Owens-Illinois’ President and Chief Operating Officer, testified to the heavy pressure on the company to keep its glass containers competitive with cans and plastic in the single-serve size: he predicted that “[i]f we lost that market, we are facing disaster in this industry.” Thus, aside from a small portion of the market that appears committed to glass primarily for reasons related to shelf-life, the single-serve soft drink end use segment appears relatively elastic. 10. Soluble Beverage Products The FTC selected “soluble beverage products” (focusing on instant tea and coffee) as another inelastic end use. Glass containers for instant coffee alone accounted for $82 million in sales in 1986. Examination of the marketing in this end use segment reflects that while this category too may have some purchasers who strongly prefer glass containers, for the large part, the end use is elastic. One major supermarket chain estimates that approximately 73% of the products on its shelf space devoted to soluble beverages are in materials other than glass. The preference for glass in soluble beverage containers, as for other end uses, is motivated by perceived consumer preference and achievement of desirable functional characteristics (impermeability, reseala-bility, and clarity) at a lower cost than presently allowed by plastic and other containers. But, here too, plastic appears to be fast making inroads with at least one major producer. Procter and Gamble, one of the leading producers of soluble beverage products, is presently test-marketing its Folger’s instant coffee in a clear plastic container. The company expects that the test will show a greater consumer acceptance for plastic and that its competitors may follow its lead if it switches to plastic. In short, the soluble beverages end use segment appears fairly elastic even though it currently relies largely on glass. 11. Scientific and Lab Glassware The last, and the least developed factually, end use segment that the FTC offers as an example of inelastic demand for glass is “scientific and lab glassware.” Owens-Illinois and Brockway are the two principal suppliers of chemical, laboratory, and seien-tifie glass containers. Other than the general testimony of a distributor of glass and plastic products for scientific and laboratory applications who stated that glass use in this area is inelastic, however, the case with regard to scientific and laboratory glassware was insufficiently developed. There is not an adequate basis for any firm conclusion regarding elasticity or significance for this end use segment. b. General Observations on Elasticity of Demand One common thread throughout the discussion of demand-side elasticity for glass is the potential for constraints imposed on purchasers and suppliers by state requirements for recycling of packaging materials or outright bans on plastics. Because glass is presently much more capable of being recycled than plastic, recycling requirements and potential opposition to wider use of plastics by environmentalists, make plastic a less attractive alternative to glass. Although adequate information was not provided on the current status of states laws on glass and plastics recycling to allow full assessment of this factor, the record reflects a widespread concern among purchasers about plastics because of the likely expense of recycling and the uncertain scope of future legislation. The extensive record in this case also provides several examples of elastic end use segments that include products very similar to those in the alleged inelastic end use segments. Defendants presented the statements of several glass container customers who stated that they have reasonable alternatives to glass for their products and could readily turn to these alternatives in the event of a significant nontransitory price increase for glass containers. Some testified, in fact, that they viewed the acquisition as “procompetitive” because they expected to received benefits through lower prices and cost and productivity improvements made possible by the acquisition. While the existence of clearly elastic end uses for glass is admitted by the FTC and does not necessarily negate the argument that certain other segments are inelastic, that fact greatly diminishes concerns about the anticompetitive effects of this merger. The elasticity further strongly suggests the great potential for spillovers in technology, improvements in cost, and changes in consumer preferences, the three factors that constitute the remaining obstacles to substitution in the eleven alleged inelastic uses. To illustrate, the largest elastic end use for glass containers is beer; it accounts for over 30% of the United States demand for glass containers, and there is no dispute that glass containers for beer compete directly with cans for packaging. Other products that are quite similar to those in the alleged inelastic end use segments but that have converted, or may soon convert, to plastic include: pourable dressings (accounting for 150,000 tons of glass per year, approximately the size of the pickle market); edible oils (constituting approximately 200,000 tons per year, or the size of the baby food market); ketchup; barbecue sauce and relishes; and some brands of peanut butter (the imminent conversion by Procter and Gamble of Jiff Peanut Butter to plastic represents approximately 135-50 million units of glass per year). The FTC contends that the effect on inelastic purchasers from these conversions will not be significant because these converted products have low sensitivity to oxygen, and for the sensitive products in the alleged inelastic end use segments, the ability of plastic to prevent permeation with the new high-barrier containers can be achieved only at a significantly higher cost. Defendants have shown, however, that the typical cost differential between plastic and glass does not prohibit conversion. The successful substitution of plastic containers for some products traditionally packaged in glass indicates that customers either do not notice the price increase or have been willing to pay more for plastic containers, even in one case where the increase has been as much as 30%. Perhaps as importantly, defendants have also shown that this price disadvantage of plastic decreases with the production increase of types and quantities of plastic containers. And, as noted below, these large conversions “free up” capacity in the glass container industry. In conclusion, both in the aggregate and with regard to the specific end uses at issue, a combination of new and existing technologies, products, marketing strategies, and changing consumer preferences strongly indicate that all but a few of the end use segments have largely elastic demand for glass. The trend in packaging is definitely toward a greater variety of container alternatives rather than fewer. While this Court, unlike the trial court in Cellophane, did not visit an “Annual Packaging Show” to observe the various packaging materials in question, 351 U.S. at 402-03, 76 S.Ct. at 1011, defendants essentially brought such a show into the courtroom by setting up several “grocery store shelves” of common and unusual products packaged in a variety of containers. Through the assembled array of products, admitted individually as exhibits, defendants confirmed the evidence in the record indicating not only that alternative packaging is feasible and is being used for many of the end use segments in question, but also that different types of packaging compete with each other in the eyes of modern consumers. In sum, analysis of the eleven end use segments presented by the FTC in addition to the other evidence of intermaterial competition in the packaging business already points to the conclusion that the relevant market in this case must be broadly defined as rigid-walled containers, which comprise glass, plastic, metal, and paper, rather than as an “all glass” container market. 2. Cross-Elasticity of Supply The second main factor to examine in determining the relevant product market is cross-elasticity of supply, i.e., whether there is production substitution among sellers. The FTC contends that because glass container plants produce only glass containers and since, with minor modifications, glass lines can produce a range of glass containers, the supply side of the equation confirms that the relevant product market is “all glass containers.” Owens-Illinois and Brockway suggest the focus on glass plants and lines is too narrow, that they both view themselves as being in the “packaging” as well as the glass business, that glass lines can be modified to produce plastic containers, that glass purchasers effectively bargain among various suppliers, and there are other potential shifts in production that would occur to expand the packaging alternatives for the end use segments in question should there be an attempt to raise glass prices. These factors, defendants argue, support a market definition including glass, plastic, metal, and paper packaging. For simplicity, most of the analysis of the potential for shifts or increases in glass production capacity is discussed below in the context of assessing, as the FTC has argued, the ability of a hypothetical cartel to impose price increases on only inelastic users. It is worth noting briefly, however, a few aspects of the potential for the alleged inelastic end users to substitute supply sources for glass (as distinct from shifting to alternative packaging). While the self-manufacture of glass containers is negligible (except for wine), inelastic end users could obtain greater flexibility in their own glass supply by turning to “stock” glass containers that are more widely available and for which supplier competition would be greater. For example, there are “stock” wide-mouth jars, “beer” bottles, and standard 16-oz. soft drink bottles. Further, as shall be seen, purchasers may be able to shift contracts to the “fringe” and foreign glass manufacturers. 3. Effect on Competition Once the relevant market has been defined, there must be an examination of whether the merger will result in anticom-petitive effects. The Court has already concluded that the relevant product market in this case is broader than “all glass containers.” Since the FTC admits that the effect of the merger is negligible if the market definition is expanded beyond “all glass containers,” the Court need go no further. Nonetheless, anticompetitive effect cases are difficult to package neatly, and it is important to explore fully whether, even assuming an all-glass container market, the effects of the merger would be anticompetitive. The conclusion reached after reviewing both the post-merger concentration of the glass market and the potential for collusion and price discrimination is that even were the market to be defined narrowly, the FTC has not persuasively shown that its hypothetical glass container cartel could maintain a scheme to discriminate against inelastic end users. a. Market Concentration As earlier discussed, Owens-Illinois and Brockway are the second and third largest manufacturers of glass containers, with 22.2% and 14.9%, respectively, of 1986 furnace capacity. See supra page 32. After the merger, Owens-Illinois would become the largest manufacturer, with 37% of the industry capacity, compared to Anchor/Diamond-Bathurst, now the leader, with 24% of the industry capacity. The Herfindahl-Hirshman Index (“HHI”), an economic measure of market concentration, now at 1467 would increase significantly after the merger to 2129, placing the glass industry in the “highly concentrated,” rather than the “concentrated” category as defined by the Department of Justice Merger Guidelines. Defendants do not directly challenge the FTC’s actual calculations in this regard, they do, however, strongly suggest two “basic errors” in the FTC’s application of the HHI which led to this litigation: (1) the relevant product market is not properly defined, and (2) the type of collusion alleged by the FTC is not one to effect an “across the board” increase in price, but rather one to price discriminate against certain inelastic end users representing no more than 25% of the market. b. Potential for Collusion and Price Discrimination The specific type of “collusion” alleged by the FTC in this case is that the glass container producers will price discriminate against the alleged inelastic end users, ie., raise prices to those purchasers who will not readily switch to alternative packaging, but not raise prices to elastic end users. Collusive agreements to raise prices while maintaining profits can be achieved by identifying inelastic customers and restricting output. Price increases can be defeated by (1) consumer switching to other products, (2) purchasers changing to different suppliers outside the market, (3) expansion of capacity by current suppliers, or (4) expansion of capacity through modification of existing or construction of new facilities. Merger Guidelines, 114492. 1. Identifying Inelastic Customers The record indicates that identification of inelastic customers for glass containers would not be difficult. While predictions of a company’s actual decision in the event of a price increase (i.e., whether it would turn to other producers, turn to alternative packaging, or investigate self-manufacture), and how swiftly that decision would be implemented, could never be certain, ironically, this very litigation has contributed significantly to Owens-Illinois’ knowledge of the various users who have the strongest preferences not to switch to alternatives, or who cannot readily resist a price increase. Counterbalancing the ability of the members of the collusive agreement to identify inelastic users is the power of certain purchasers to defeat price increases. Many of the purchasers in the allegedly inelastic end use segments are major, sophisticated buyers whose orders for glass are infrequent and large. These purchasers, therefore, are in a good bargaining position with respect to producers attempting to impose a price increase. Some of these purchasers testified, in fact, that they have in the past been able to defeat proposed increases in the price of glass containers. Moreover, some producers use competitive bidding, require cost justification for price increases, contract with several competing suppliers, and qualify more suppliers than they need, factors which all exert pressure on purchasers not to raise prices. And, these large customers often buy glass containers for elastic end uses as well as the allegedly inelastic end uses at issue, indicating that they could more easily defeat unjustified price increases and threaten producers with loss of contracts for the elastic end uses were prices to be raised for the inelastic end uses. Concentration of the buying side of the market would tend to inhibit collusion not only directly because of the potential to defeat price increases, but also because a market composed of few, significant buyers increases the incentives of “cheating” within the cartel, “for with a single transaction, [a cheating member] may be able to increase his sales and hence profits dramatically.” Hospital Corporation of America v. F.T.C., 807 F.2d 1381, 1391 (7th Cir.1986), cert. denied, — U.S. -, 107 S.Ct. 1975, 95 L.Ed.2d 815 (1987). 2. Capacity The FTC argues that the glass container industry is already operating at a very high capacity, which enhances the potential for collusion and price discrimination. High capacity utilization levels favor collusive behavior because firms with excess capacity have an incentive to cut prices to sell more of their product or to increase output to capture any supracompetitive profits that might be available from committed or inelastic end users and thus undermine the collusive agreement. Owens-Illinois’s own estimates show consistently high capacity across the firms in the glass industry: Manufacturer Capacity Utilization Owens-Illinois 96% Anchor 92 Anchor-Hocking 90 Arkansas 86 Ball-Incon 86 Brockway 90 Columbine (Coors) Gallo 90 102 Glenshaw 87 Hillsboro 88 Industrial (Tropicana) 95 Kerr 91 Latchford 90 Leone 90 Liberty 92 Manufacturer Capacity Utilization Miller 100% Triangle 93 Wheaton 89 Total Industry: 92 Industry and glass purchasers are well aware that excess capacity exerts downward pressure on prices and profits. The FTC points out that Owens-Illinois has indicated that it intends to offset the productivity gains from consolidation with curtailment of capacity in order to maintain its market share. Precise calculations, of course, are impossible, but Dr. Nelson, the FTC’s economic expert, estimated that even a 5% increase of capacity in the industry would tend to depress prices. In contrast to the picture painted by the FTC, Owens-Illinois argues that substantial excess capacity exists in the glass industry as a whole (they contend that the percentage of practical operating capacity utilized in the industry was 89% in 1985, 80% in 1986, and 88% in 1987), although Owens-Illinois has been operating at its effective maximum capacity, approximately 92%. Furthermore, Owens-Illinois estimates that the additional capacity available on a short-term basis from existing facilities might be as high as 26%. Increases in capacity can come from running through holidays, “electric boosting,” delaying furnace rebuilds, and improvements in forming machines. Another source of increased capacity, suggests defendants, would be to bring back on-stream some of the furnaces that have been idled or shut-down in recent years. The FTC argues, however, that such modifications could consume three months to one year to implement and that increased production can only be achieved through expenditure of significant amounts of capital, labor, and electricity. Existing capacity commi