Substantial Lessening of Competition - The Section 7 Standard SUBSTANTIAL LESSENING OF COMPETITION-THE SECTION 7 STANDARD Thomas 0. Barnett* I. Introduction ............................................................... 293 II. The Consumer Welfare Standard ............................. 295 III. The Definition of a Competitive Effect ..................... 298 A. Coordinated Effects ............................................ 299 B. Unilateral Effects ............................................... 300 IV. Proving Competitive Effects ..................................... 304 A. Custom er Testimony ........................................... 306 B. Company Documents .......................................... 310 C. Econom ic Analyses ............................................. 311 I. INTRODUCTION The topic for this year's Milton Handler Lecture is "anticompetitive effects." As one of the Deputy Assistant Attorneys General in charge of the Antitrust Division's merger enforcement efforts, I will address anticompetitive effects in the context of the "substantial lessening of competition" standard found in Section 7 of the Clayton Act. This year's lecture takes place in the wake of two recent judicial merger decisions-the decision in the Arch Coal case written by my co-panelist Judge Bates,' and the decision in the Oracle case by Judge Vaughn Walker in the Northern District of California.2 Both of those merger challenges were brought by federal enforcers-the FTC in the Arch Coal case and the DOJ's Antitrust Division in the Oracle case. Each was a high-profile battle, and in both cases the defendants won and the government lost. The Arch Coal and Oracle .Deputy Assistant Attorney General, Antitrust Division, U.S. Department of Justice. The views and opinions expressed herein are those of the author and do not necessarily represent the official position or policies of the U.S. Department of Justice. FTC v. Arch Coal, Inc., 329 F. Supp. 2d 109 (D.D.C. 2004). United States v. Oracle Corp., 331 F. Supp. 2d 1098 (N.D. Cal. 2004). COLUMBIA BUSINESS LAW REVIEW decisions have sparked musings by antitrust commentators regarding everything from the continuing viability of the government's theories of anticompetitive effects (e.g., can unilateral effects arguments succeed?) to the government's litigation strategy (e.g., does customer testimony matter, and if so, how?), to the evidence that plaintiffs may have to present in order to prevail (e.g., is econometric analysis essential?). Before turning to those specific questions, I would like to address a larger policy question posed by a few commentators-specifically, whether the Arch Coal and Oracle decisions suggest that something might be wrong with the government's Section 7 enforcement efforts, and, if so, whether a major shift in enforcement policy or practice may be needed. The answer to both of these questions is "No." Although the government did lose two recent merger cases in district court-three if one includes 2001's SunGard case3-those losses should be kept in perspective. When one considers how very rarely merger cases are brought before judges, the government's overall win/loss record including consent decrees and abandoned transactions is actually quite strong. Importantly, defendants tend to litigate only those merger cases that they think they are most likely to win. Given this "adverse selection bias," and the uncertainty inherent in any court proceeding, it should not come as a surprise if on occasion the government loses a case. I respectfully suggest that the Division's case against Oracle's acquisition of PeopleSoft falls squarely within this framework. The Division continues to believe that it presented evidence sufficient to prove that the transaction is anti-competitive and that another district court judge faced with the same record of the case could well have decided to enjoin the transaction. Nevertheless, we recognize the important role that the federal courts play in our merger ' United States v. SunGard Data Systems, Inc., 172 F. Supp. 2d 172 (D.D.C. 2001). [Vol. 2005 No. 2:293] THE SECTION 7 STANDARD enforcement system, and we respect Judge Walker's decision. I will have more to say regarding the Oracle case later. In Part II, I will discuss what we are looking for when we set out to analyze the competitive effects of a transaction. In Part III, I will briefly describe the different types of anti- competitive effects that plaintiffs in merger cases set out to prove. I will conclude in Part IV by offering some observations regarding categories of evidence that may be presented to establish that a transaction is likely to result in a substantial lessening of competition. II. THE CONSUMER WELFARE STANDARD Section 7 of the Clayton Act prohibits transactions likely to create a "substantial lessening of competition.'" Thus, the search for anticompetitive effects is central to the application of the statutory standard. The precise definition of such anticompetitive effects, however, has been the subject of extensive debate over the years. Today, most would agree that proper enforcement of the antitrust laws focuses on consumer welfare. Thus, a merger that results in a combined firm with the power to profitably charge supra-competitive prices or reduce output below competitive levels will be said to result in a substantial lessening of competition. This focus on consumer welfare necessarily draws the analysis to issues like price, quantity, quality, and innovation. Other potential standards exist, of course. For example, a focus on overall social welfare would promote efficiency with less concern for distributional issues. One also could examine a transaction's effect on employment levels or environmental quality. Some have urged (shortsightedly, in my view) the promotion of nationalistic economic goals by discouraging acquisitions of domestic firms by foreign buyers while encouraging the combination of domestic competitors into "national champions." One critical development during the last thirty years of antitrust enforcement has been the consensus that antitrust 15 U.S.C. § 18 (2000). No. 2:293] THE SECTION 7 STANDARD COLUMBIA BUSINESS LAW REVIEW should focus on consumer welfare. A key insight is that many of the other suggested goals (such as increased employment or strong national firms) are best served by the consumer welfare standard. This evolution-or revolution- traces back to former Assistant Attorney General Bill Baxter in the early 1980s, whose contributions are difficult to overstate. This consumer welfare standard means that we do not block a merger in the United States because it might reduce employment levels in the Rust Belt or because it might raise environmental concerns in a particular wetlands area. Those are certainly important public policy concerns, but they should be addressed separately, with a recognition of the economic trade-offs that they present. So in the United States, and increasingly in Europe and other parts of the world, we have consumer welfare in mind when we examine a transaction to determine whether it is likely to result in a substantial lessening of competition. That analysis necessarily focuses on factors that have the most direct impact on consumers: price and quantity. In an effort to inject some precision into the discussion, and at the risk of frightening away those scarred by their undergraduate economics courses, I offer the following graph:5 ' Graph provided by Ken Heyer, Economics Director, Antitrust Division, U.S. Department of Justice (adapted from the work of Oliver Williamson (see Oliver E. Williamson, Economies as an Antitrust Defense: The Welfare Tradeoffs, 58 AM. ECON. REV. 18, 21-23 (1968)). [Vol. 2005 THE SECTION 7 STANDARD Figure 1 Welfare: Producers Gain and Consumers Lose Price, A = Increase in Producer Surplus B = Loss in Consumer Surplus C = Transfer from Consumers to Producers Marginal Cost 2 Cost Savings/ Producer Profits Demand Quantity Q2 -- Q1 Figure 1 is a basic supply and demand graph. P1 represents the competitive price, and Q1 represents the competitive output, both determined by the point at which the "Marginal Cost 1" curve intersects the Demand curve. The graph also illustrates the potential effects of a merger that creates market power, but which also generates efficiencies. The result is that price increases to P2 and quantity decreases to Q2. These changes cause society to experience a dead weight loss (area B), and will likely lead to some transfer effects (area C). This particular graph illustrates how the consumer welfare and total welfare standards can diverge, although I think it is a rare case in practice. Marginal cost has been reduced in this example, producing certain efficiencies. However, the combined firm's market power has also been increased, leading to reduced supply and increased prices. In the end, consumers are unequivocally worse off, regardless of the benefits to the combined firm, because they are forced to pay higher prices because of the merger. Producers are substantially better off (see area A), because they are able to price well above their marginal costs ("Marginal Cost 2") and those costs have come down. In this example, total surplus has increased, but consumer surplus has dropped. No. 2:2931 COLUMBIA BUSINESS LAW REVIEW Figure 1 illustrates the "consumer welfare versus social welfare" debate that continues in the United States as well as in other countries, such as Canada. I am not going to attempt to resolve that debate here, except to repeat that the enforcement authorities in the United States look most frequently at the question of what is best for consumers. In practice, a substantial lessening of competition typically can be established most directly by a post-merger price increase, although our Horizontal Merger Guidelines do not limit the agencies to a rigid or narrow search just for post-merger price increases.6 III. THE DEFINITION OF A COMPETITIVE EFFECT A transaction may substantially lessen competition under Section 7 if it causes significant anticompetitive effects in a relevant market. However, because in all premerger investigations we must speak in terms of probabilities and possibilities, of incipient effects and likely outcomes, some uncertainty is an inevitable part of competitive effects analysis. Much of the work in this area over the past few decades could fairly be characterized as an effort to make that combination of analysis and estimation more of a science than an art. In the taxonomy of the antitrust kingdom, we divide the genus of anticompetitive effects into two basic species: unilateral effects and coordinated effects. Before addressing each of these species, two points warrant mention. First, all of this analysis serves as a means to assess the statutory standard of a substantial lessening of competition. Second, while methods of detecting and explaining such effects are constantly evolving, the basic concepts that those terms describe have been around for a long time. Terminology may change, and the analytical tools may evolve, but the basic potential anticompetitive effects have not changed. Simply put, when we examine a merger we set out to determine 6 U.S. DEP'T OF JUSTICE & FED. TRADE COMM'N, 1992 HoluzoNTAL MERGER GUIDELINES (1992) § 0.1, n.6 [hereinafter MERGER GUIDELINES]. [Vol. 2005 THE SECTION 7 STANDARD whether the transaction is likely to lead to consumer harm- e.g., higher prices, reduced quality or quantity, or a reduced rate of innovation and product improvement-by significantly altering the balance of the market in favor of the combined firm, or of the combined firm and its competitors. A. Coordinated Effects As explained in the Guidelines, a merger "may diminish competition by enabling the firms selling in the relevant market more likely, more successfully, or more completely to engage in coordinated interaction that harms consumers. Coordinated interaction is comprised of actions by a group of firms that are profitable for each of them only as a result of the accommodating reactions of the others."7 This behavior can be express or tacit, and the behavior may or may not be lawful in and of itself. Successful coordinated interaction "entails reaching-terms of coordination that are profitable to the firms involved and-an ability to detect and punish [cheating] ." The Guidelines provide examples of "terms that are profitable," such as common pricing, fixed price differentials, stable market shares, and customer or territorial allocations and restrictions.9 Historically, the literature and the case law have been richer regarding coordinated effects than unilateral effects. Given the priority assigned to rooting out conspiracies under the Sherman Act, it should be expected that courts, enforcers, and scholars would carefully examine the extent to which a transaction might make express or tacit collusion more likely. At the federal enforcement level, when attempting to determine whether a merger-say, a reduction from four competitors to three-is likely to result in anticompetitive coordinated effects, we apply the analytical framework of the 7 Id.§ 2.1. 8 Id. 9 Id. § 2.11. No. 2:293] COLUMBIA BUSINESS LAW REVIEW [Vol. 2005 Horizontal Merger Guidelines to determine whether the requisite detection and enforcement "mechanism" for coordination exists. Put another way, we look for certain market characteristics that suggest that further concentration will increase the likelihood of express or tacit collusion among that market's remaining producers." Thus, because it is easier for a group of competitors to reach agreement on a single product than on a portfolio of products, we look to see whether the relevant products are homogeneous or differentiated. Because it is more difficult to rig a competition if you cannot predict when sales will occur, we examine whether sales in the market are regular or lumpy. Because successful tacit price coordination depends on competitor access to information about other competitors' prices, and because express agreements are difficult to enforce without knowing what others are doing, we examine the extent to which transactions and production decisions in the market are transparent. Because participants in a coordination scheme can hardly enforce illicit agreements in court, we look to see whether they will be able to enforce the agreement through their own actions. For example, is there some way for colluding suppliers to rein in a maverick? If suppliers in a market are not free to cut their short-term prices in order to bring the maverick back into line-e.g., because prices are set by long-term contracts-successful coordination in that market is less likely. B. Unilateral Effects A merger "may diminish competition even if it does not lead to increased likelihood of successful coordinated interaction, because merging firms may find it profitable to alter their behavior unilaterally following the acquisition by elevating price and suppressing output."11 For example, the merger of two horizontal competitors eliminates whatever 10 See Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 238 (1993). "1 MERGER GUIDELINES, supra note 6, § 2.2. COLUMBIA BUSINESS LA W REVIEW [Vol. 2005 No. 2:293] THE SECTION 7 STANDARD competitive constraints the two firms exerted on each other prior to the deal, so the acquiring firm might be able profitably to raise its prices, slow its innovation, or otherwise compete less aggressively post-merger because it no longer has to fear being undercut by the acquired firm. If sufficient substitutes are not available in the market to discipline the combined firm, consumers will be forced to pay more for the same goods in the post-merger world. Unilateral effects made their official Guidelines debut in 1992, and there is little case law expressly addressing unilateral effects in the merger context. However, as Judge Walker observed in the Oracle decision, the concept is hardly new. 2 If it were, we would not have condemned a merger to monopoly under the Clayton Act before the Guidelines were revised because a monopolist would have sufficient market power to raise prices unilaterally. This is not to ignore the proliferation of unilateral effects theories over the last decade, thanks largely to the work of many excellent economists. While some of these theories and analytical tools may still have an image that is slightly avant-garde (e.g., merger simulations and game theories), many of them have very old and established roots (e.g., Cournot theories). There are a wide variety of theories: output competition theories and price competition theories, auction theories and bargaining theories, and so on. At their heart, however, these are all attempts to examine and explain-and in the Clayton Act context, to predict-the unilateral conduct of firms post-transaction. Their variety is a result, not of novelty, but of a continuing effort by economists and others to refine the analysis of the competitive process. These evolving analytical tools are the reason we are able at times to rely less on structural analyses of market shares and market concentration statistics when evaluating the likely competitive effects of a 12 See United States v. Oracle Corp., 331 F. Supp. 2d 1098, 1113 (N.D. Cal. 2004) ('unilateral effects' is primarily a new term [used] to address antitrust issues that courts have in other contexts considered for quite some time"). No. 2:293] THE SECTION 7 STANDARD COLUMBIA BUSINESS LAW REVIEW merger. Although one can rely on these new techniques too much, as I will discuss further below, the movement away from "by the numbers" merger review that they have made possible is a positive development. While on the subject of unilateral effects, there are a few points I would like to make about presumptions. According to the Guidelines, there should be a presumption of unilateral effects where "the merging firms have a combined market share of at least thirty-five percent," and "data on product attributes and relative product appeal show that a significant share of purchasers of one merging firm's product regard the other as their second choice."13 The presumption is rebuttable, for example, by evidence that rival sellers likely would replace any competition lost through the merger by repositioning their product lines. In his recent decision in United States v. Oracle, Judge Walker suggested that "strong presumptions based on mere market concentration may be ill-advised in differentiated products unilateral effects cases," given the difficulty of properly defining the market in such cases." Judge Walker's decision, however, did not rest on the application of any presumptions, but rather on his rejection of the government's definition of the relevant market. Indeed, by the time Judge Walker was finished with the market definition question, it would have been difficult to establish that the merging parties had more than 35% of any market. His statements regarding the presumption are therefore pure dicta. Nevertheless, I want to state that I disagree with them, particularly any suggestion that a presumption should not apply in unilateral effects/differentiated products cases. If two competitors in a differentiated product market account for a large percentage of sales (i.e., more than 35%), and other information indicates that a significant number of their customers view the two firms as their top two choices, it follows as a matter of logic that the transaction, all else equal, is likely to harm a significant number of consumers. If the market realities of a 1 MERGER GUIDELINES, supra note 6, § 2.211. 14 Oracle, 331 F. Supp. 2d at 1121. [Vol. 2005 No. 2:293] THE SECTION 7 STANDARD particular case lead to a different result, the parties can provide evidence of those circumstances to rebut the presumption. I would also point out the following to all those who take comfort from any retreat from the use of presumptions: As reliance on presumptions is reduced, the antitrust enforcement agencies will have to conduct more extensive investigations and more fact-intensive analyses during their already-compressed HSR merger reviews. This need will require the agencies to collect more information earlier in the process, so that they can conduct the necessary analyses and prepare for possible litigation. The result will be increased burdens on the parties, including less flexibility by the agencies in negotiating modifications to second requests. Another related point warrants mention. Some who read the Guidelines mistakenly interpret the 35% presumption as creating a safe harbor for mergers where the combined shares are less than 35%.15 This reading of the Guidelines is not correct. Simply put, there is no such safe harbor. This is because market shares do not tell the whole story. Many mergers that cross the 35% threshold are not challenged because they are not anticompetitive. That is why the presumption is rebuttable. The limit of market share statistics cuts both ways, however. Mergers that fall below the 35% threshold are frequently not a competitive problem, but they are not simply given a pass. A proper analysis of such a transaction examines whether there could be a reason for concern despite a relatively small combined market share. This is what Judge Walker did in the Oracle decision. Notwithstanding the fact that under his definition of the market, Oracle and PeopleSoft combined would not have had a market share greater than 35%, the judge did not end his inquiry with a structural analysis. Instead, he proceeded to analyze the likely competitive effects of the transaction in what he saw as the relevant market. This approach was " See, e.g., the exchange between Gregory Werden and Joseph Kattan at the joint DOJ/FTC Merger Workshop, Feb. 2004 (transcript available at http://www.ftc.gov/bc/mergerenforce/040218ftctrans.pdf at p. 12 1-23. THE SECTION 7 STANDARDNo. 2:293] COLUMBIA BUSINESS LAW REVIEW appropriate, and further demonstrates the lack of a safe harbor in this context. IV. PROVING COMPETITIVE EFFECTS What happens when an investigation leads an agency to conclude that the likely anticompetitive effects of a proposed merger will constitute a substantial lessening of competition? Reviewing authorities in some jurisdictions have the power to block such mergers. In such jurisdictions, parties may be entitled to appeal the decision to a court, such as the European Community's Court of First Instance in Luxembourg, but market realities combined with the usually drawn-out appeals process mean that in most such cases the reviewing authority has the final word for all practical purposes on that transaction. By comparison, agencies in the United States have no legal authority to prohibit a transaction in the first instance:16 we can investigate a transaction, and if we conclude that the deal is anticompetitive we can file a complaint in federal court (or the FTC can initiate an administrative law judge proceeding) to block it, but we must persuade a judge to enjoin the deal. To be sure, a decision by the Assistant Attorney General or the Federal Trade Commission to file a complaint in a merger case is a significant event. The decision reflects the considered judgment of the agency that the transaction violates Section 7. More generally, the risks and costs of litigation and the uncertainty that it can cause in the market among customers and key employees of the acquired firm, and even in the mind of a seller who is eager to collect the purchase price, often force parties to the negotiating table. Crown jewels have been surrendered and whole deals abandoned in the face of enforcement action by the federal agencies. Nevertheless, the parties may require the agencies to put their case to a judge. 16 The FTC, of course, can prohibit a transaction after review of a decision on appeal from an administrative law judge proceeding. [Vol. 2005 No. 2:2931 THE SECTION 7 STANDARD How, then, does the government persuade a judge that a transaction is likely to result in a substantial lessening of competition? The Guidelines describe the rebuttable presumption that arises when the government can establish that the post-merger firm will have a large share of the relevant market. This has been a basic tenet of antitrust enforcement for decades. 7 The Guidelines expand on this concept in their discussion of various thresholds in the Herfindahl-Hirschman spectrum, which are intended to provide parties with some guidance as to when the agency might be inclined to scrutinize the transaction more closely. But as we saw in both the Oracle and the Arch Coal decisions, presumptions and HHI thresholds are not the end of the analysis. Although we at the Division disagreed with many of Judge Walker's factual and legal conclusions, we certainly concur with his decision to examine the competitive effects of the Oracle/PeopleSoft deal despite the fact that the combined firm would not have a share of the relevant market-as he defined it-that was as much as 35%. In a related point, faced with competing possible market definitions, Judge Bates adopted an interesting approach. Reasoning that a transaction that comes close to the threshold might be one that is too close to call on a technicality, he chose to apply a "slight presumption," which was enough for him to continue the analysis and examine the evidence of competitive effects. The question then remains: what kinds of evidence beyond market shares can be presented to prove that a 17 See, e.g., United States v. Philadelphia Natl Bank, 374 U.S. 321, 363 (1963), in which the Court held: [A] merger which produces a firm controlling an undue percentage share of the relevant market, and results in a significant increase in the concentration of firms in that market is so inherently likely to lessen competition substantially that it must be enjoined in the absence of evidence clearly showing that the merger is not likely to have such anticompetitive effects. No. 2:293"1 THE SECTION 7 STANDARD COLUMBIA BUSINESS L4W REVIEW [Vol. 2005 transaction is likely to result in a substantial lessening of competition? In light of recent decisions I would like to offer a few brief observations about three specific types of evidence: customer testimony, company documents, and economic analysis. A. Customer Testimony The Arch Coal and Oracle decisions have generated much discussion about the role and value of customer testimony. Judge Bates stated in his Arch Coal decision that he did not doubt "the sincerity of the anxiety expressed by [Southern Power River Basin coal] customers," who testified that increased consolidation in the market would lead to less competition and higher prices.18 However, Judge Bates summarized the customer testimony as stating "little more than a truism of economics: a decrease in the number of suppliers may lead to a decrease in the level of competition in the market."19 For his part, Judge Walker provided a lengthy, if selective, review of what the enterprise software customer witnesses said on the stand and in their affidavits and depositions. He characterized the customer witnesses as "knowledgeable," "well-informed," and "sincere." However, he ultimately "found the testimony of the customer witnesses largely unhelpful to plaintiffs' effort to define" the relevant market, and he "question[ed] the grounds upon which the [... witnesses offered their opinions on the definition of the product market and competition within that market.""° In light of these decisions, a number of commentators have questioned the government's reliance on customer testimony, both at trial and in merger investigations. I cannot speak to the customer testimony that was provided by the FTC in Arch Coal, as I was not involved in that case. However, I continue to believe that the customer testimony "' FTC v. Arch Coal, Inc., 329 F. Supp. 2d 109, 145-46. (D.D.C. 2004). 19 Id. at 146 (emphasis in original). '0 United States v. Oracle Corp., 331 F. Supp. 2d 1098, 1130-31 (N.D. Cal. 2004) COLUMBIA BUSINESS 1LA4W REVIEW [Vol. 2005 No. 2:293] THE SECTION 7 STANDARD the Division presented in the Oracle case was both compelling evidence and an extraordinarily important tool for understanding the structure of the market. As an initial matter, few dispute the value of customer testimony regarding their functional needs and past actions. Compared to the lawyers, the parties, and even the experts, customers are uniquely competent to testify about their functional needs, about their past practices, what suppliers they have used and why, what suppliers they have rejected and why, and the extent to which head-to-head competition between the merging parties is or is not significant. That kind of important foundational information is tremendously useful in our analysis, and we believe it is valuable to a court as well. The recent discussion has focused on customer testimony on predicted future activity and predicted future competitive effects. Some argue that customers are too biased to serve as reliable sources of predictive evidence. To that I would say that all witnesses can be biased, so it is entirely appropriate to probe their positions and delve underneath their testimony to ascertain their motivations. Nonetheless, of the potential witnesses out there, I respectfully suggest that it is the customers whose interests tend to be most in line with the goals of the antitrust laws, which are, after all, enforced to protect consumer welfare-i.e., their welfare. The Federal Rules of Evidence properly limit the testimony that a customer can offer."' However, although a customer's views about what might happen are not as reliable as his testimony about what has happened, the former kind of testimony is not without value. Rule 701 states that a witness who is not testifying as an expert-and only the exceptional customer witness could be classified as an expert-may still offer his or her opinions if such testimony is "(a) rationally based on the perception of the witness, (b) helpful to a clear understanding of the witness' testimony or the determination of a fact in issue, and (c) not based on scientific, technical, or other specialized knowledge 2 See, e.g., FED. R. EVID. 701, 702. THE SECTION 7 STANDARDNo. 2:293] COLUMBIA BUSINESS LAW REVIEW within the scope of Rule 702," which governs testimony by 22experts . Testimony regarding customers' specific experiences can shed light on the potential competitive effects of a transaction by, for example, demonstrating the extent to which a substitute supplier may be able to serve a particular class of customer. Suppose a customer faced with the recent idling of one of its main supplier's plants was unable to locate an acceptable new alternative source, causing prices to rise by 25%. Such a customer might have a rational basis from which to predict the likely price effects of a transaction combining his main suppliers. In my opinion, this kind of evidence can appropriately provide at least part of the support necessary for a conclusion that the transaction is likely to reduce competition. Of course, the weight given to customer testimony depends on the specific case. Take, for example, a market in which suppliers charge essentially the same price to all of their customers, i.e., a market devoid of price discrimination. If an inframarginal customer in such a market testifies that he would not switch from merging suppliers A and B to supplier C in response to a post-merger price increase, it remains possible that C will be able to replace the competition that will be lost as a result of the AB merger. Prices in such markets are determined by the behavior of marginal customers. If those customers can discipline the post-merger pricing decisions of AB by switching to C in response to a price increase, the concerns of inframarginal customers, even if sincerely held, are not sufficient to show harm to competition. On the other hand, if prices are separately negotiated for each customer such that the price paid by one has no direct effect on the prices paid by others, the testimony of any one customer carries substantially more weight. This was a significant flaw in Judge Walker's decision, in which he all but ignored the government's price discrimination evidence and arguments, and at points appeared to treat all of the testifying customers as if they 22 FED. R. EVID. 701. [Vol. 2005 THE SECTION 7 STANDARD were inframarginal customers whose views, though sincerely held and based on extensive experience, could be swept aside as insufficient. More generally, Judge Walker criticized the Division's customer witnesses for not having provided more detailed, internal, quantitative analyses of the alternative suppliers that Oracle claimed were in the relevant market, such as Lawson, AMS, and outsourcers.23 This criticism raises another important question regarding customer testimony: How much support must a customer's views have before they are given credence by the fact-finder? Put another way, how many spreadsheets must a customer produce in order to elevate his testimony above the level of "mere" opinion or speculation? The answer to that question depends on the circumstances of the case. In some cases, the lack of analyses of the type sought by Judge Walker might reinforce the customer's testimony. To illustrate, I submit that the fact that the knowledgeable, experienced, sophisticated, sincere customers that the Division put forward-and those are all Judge Walker's words-had not conducted extensive analyses of the costs of using Lawson or AMS, demonstrates something very important all by itself. I submit that it reflects that those other vendors most likely could not meet the customers' functional needs, and therefore could not take PeopleSoft's place as a constraining competitive force against Oracle in the post-merger market. If a contractor looking to haul granite from a quarry to a construction site fails to generate a quantitative analysis of the cost effectiveness of doing the job with a fleet of Ford Escorts instead of a truck, the decision not to conduct the analysis is evidence that the Escorts are not reasonable substitutes for the truck. In the world of enterprise software applications, the fact that customers do not produce spreadsheets analyzing the cost of replacing their enterprise software suite with thousands of accountants with adding machines similarly reflects that the 23 I disagree with a description of the Division's customer testimony as devoid of such analyses. There was testimony that these alternatives were far more expensive for many customers. No. 2:2931 COLUMBIA BUSINESS LAW REVIEW latter approach is not a price-constraining substitute for the former. While the Division did introduce customer testimony on the greater costs of other options, it also introduced extensive testimony on the inability of other options to meet enterprise customers' functional needs.2" Under such circumstances, sophisticated customers looked at them no further. However, in other cases, a customer's failure to provide some quantitative analysis could well reduce the weight given to his or her testimony. B. Company Documents The merging parties' own documents and testimony also provide important evidence in most cases. There is general agreement that the parties will have some explaining to do if, for example, the recommendation to the Board describes how the merger will enable the company to cement its dominant position in a particular geographic market; an e-mail from the vice president for strategic development to the CFO asks her to adjust the offer price to reflect the fact that the deal will enable the combined firm to charge "whatever we want"; or the PowerPoint presentation to institutional investors prominently features the logo for the old "Monopoly" game. Even aside from such high-level strategic planning documents, company documents that provide details of extensive head-to-head competition between the merging firms, for example, raise the possibility that consumers may suffer if the competition between the parties can not be replaced post-merger. Similarly, pricing strategy documents 2' While I have no wish to re-litigate the Oracle case, the Division disagreed strongly with Judge Walker's conclusion that the customer testimony he heard was merely "speculation." To cite just one example, there was at least one customer who was so concerned about a post-merger price increase that it purchased a superfluous license as an insurance policy to avoid expected higher prices after the transaction, believing that it would cost even more money if it waited until its options were reduced as a result of the merger. Judge Walker declined to address that testimony in his opinion, but I suggest that it is compelling evidence of the transaction's likely anticompetitive effects. [Vol. 2005 No. 2:293] THE SECTION 7 STANDARD can be persuasive evidence. If, for example, the acquiring firm only looks to the acquired firm when setting its prices, or if in a bidding market the merging parties are most often the lowest and second-lowest bidders, those facts raise a concern (albeit a rebuttable one) that the elimination of the acquired firm as an independent supplier might have an anticompetitive effect on the market. For that matter, any document that discusses the likely price effects of the transaction, whether a blatant prediction of a price increase found in a strategy document or a financial model of the transaction that includes variables for revenue or margin increases, has the potential to set off alarm bells at the agencies and in the courtroom. C. Economic Analyses There was a time in the not-too-distant past when a prosecutor could obtain a felony conviction if she had an eyewitness and proof of motive and opportunity. Today, a prosecutor who goes in front of a jury without DNA evidence, computer models, and psychoanalytical studies risks disappointing television-soaked jurors who have come to expect more than traditional, old-school detective work. We might call it the "CSI factor": the tendency of factfinders to confuse availability with necessity, and to find in favor of a defendant if presented only with the traditional and not the state-of-the-art. DNA evidence and the increasingly sophisticated economic tools that have been developed by the antitrust community, such as regressions, merger simulations, diversion ratios, and critical loss analyses, share some characteristics. Both are usually complicated, highly technical, and expensive. The quality of both depends enormously on the quality of the data that is analyzed. Although economic analyses are not as precise as DNA evidence, they can provide useful insight into whether a transaction is likely to lessen competition. But we still have to use such tools with caution, always recognizing their inherent limitations. A wise man-Mark Twain, according THE SECTION 7 STANDARDNo. 2:293] COLUMBIA BUSINESS LAW RE WE W [Vol. 2005 to some--once observed that "most people use statistics the way a drunkard uses a lamp post: more for support than illumination." When done properly, regression analyses and other types of sophisticated economic analyses can be very illuminating, but when done improperly-when based on faulty data or flawed assumptions, for example-such studies can be less useful.25 Furthermore, Section 7 only requires the government to show that a transaction may substantially reduce competition, not that the government prove that it will. We need to exercise caution so that greater emphasis on increasingly refined and complex economic analyses at the expense of more traditional forms of evidence does not edge us towards a standard of absolute or clear and convincing proof. That is not the standard established by the Act. For that matter, depending on the transaction or the industry, sophisticated merger simulations and other analyses simply may not be possible. For example, because of the vast amount of "scanner data" that is collected by retail stores in order to track sales, pricing, and inventory information, mergers involving consumer products lend themselves well to modeling exercises and cross-elasticity calculations, which if done correctly can be highly probative. But what of products that are not sold via such carefully monitored and robust channels? If the government does not or cannot prepare an economic model, should we ignore more "traditional" evidence like customer testimony and documentary evidence? Have merger cases been reduced to 5 See, e.g., Gregory Werden et al., A Daubert Discipline for Merger Simulation, 18 ANTITRUST 89 (2004). Werden and his co-authors write: If merger simulation had a well-established and consistent high degree of accuracy in predicting the effects of actual mergers, its reliability would be a settled question. Because that is not the case, the reliability of any particular application of merger simulation should be gauged by examining the modeling process, which is at least as much art as science. COLUMBIA BUSINESS LA W REVIEW [Vol. 2005 No. 2:293] THE SECTION 7 STANDARD mere battles of experts? My answer to all of those questions is a strong "No." Indeed, even when more sophisticated forms of evidence are available, I submit that traditional evidence such as testimony and documents from customers, competitors, and the merging parties remain valid and useful evidence that can support a finding that a transaction violates Section 7. Take the following graph as an example. Figure 2 is an example of a data tabulation that is not sophisticated, but one that is nevertheless highly illustrative. It is a graph that I created that shows the general pattern of the total of volume of mail sent out by Members of Congress under their old franking privilege, by month, around 1970: Figure 2 Mail Time Series 60 0 40 30 20 . 0 1970-1971 I suggest that, without doing any sophisticated analysis of the raw data, one can predict from this chart alone the one month out of twenty-four in which elections occur for Members of Congress. Notice that there seems to be an increase in the amount of mail going out from Members to their constituents every other October (e.g., October 1970 but not October 1971). Based only on this compelling graph, I would be willing to draw a link between the spikes in volume of mail and the elections. No. 2:293] THE SECTION 7 STANDARD COLUMBIA BUSINESS LAW REVIEW EVol. 2005 Here is another real-world example, this time from something more relevant to our topic of competitive effects. Professor Ken Elzinga, the government's economic expert in the Oracle case, was asked to review Oracle's "discount approval forms." These were forms that the company's sales people submitted to senior management in order to receive authorization to offer prices that were discounted off of the list price by more than a pre-authorized amount. Each contained a "Justification" section that often mentioned competition with a particular vendor as the reason behind the request. Professor Elzinga reviewed the discount approval forms in an effort to understand the extent of the head-to-head competition between Oracle and PeopleSoft, and to determine the sources of competitive pressure causing Oracle to reduce prices. For example, if Lawson were the significant player in the high-function enterprise software application market that Oracle claimed it to be, one would expect to see Oracle discounting against Lawson on a regular basis. Figure 3 is a chart that Professor Elzinga prepared as a summary of his findings. Professor Elzinga found that PeopleSoft is mentioned far more than any other company, followed by SAP and JD Edwards, the latter of which is now part of PeopleSoft. The list of vendors against whom Oracle's sales people feel the need to discount aggressively falls off dramatically after that point. Is this analysis as sophisticated as a merger simulation or regression analysis? No. Is it nevertheless illustrative? I believe that it is. Indeed, I believe that even Mark Twain might have found it illuminating. Thank you again for inviting me to participate in this review. [Vol. 2005COLUMBIA BUSINESS LAW REVIEW THE SECTION 7 STANDARD U U U U U I _ _ _ _ 0 0C 0 0 0) 0 0D v c~i 0 0 to v' c~ 1-4 1-4 1-4 pauaupl saml jo jaqu"N V-!A iou0o3I Ii00-jAOd uo0saed SON smpul -. 0 d~ap M 4 aaD ooneoa 4P.0IO .2upv uonios uluod a-vaH • mooja IWO " -u l!d. "ooelo~d 2Loq(o " pll! mjpuan atpw 10400 ouodXHl wam dayv Sw-v Ioq-u! .. p.r -2.1 I wVS I 5 sopa^dooj Uo0d- No. 2:293] 315 'o c a .5 CO o 0 . IIIIII