WHAT IS HARM TO COMPETITION? EXCLUSIONARY PRACTICES AND ANTICOMPETITIVE EFFECT Eleanor M. Fox* I. INTRODUCTION Antitrust, or competition law, is said to be comprised of two types of offenses—exploitative and exclusionary. The paradigmatic exploitative offense is a cartel, which raises prices to buyers and ultimately to consumers. The paradigmatic exclusionary offense is a boycott to enforce a cartel. The cartelmembersmust keep at bay outsiders who would destroy their enterprise. This means that such a boycott is also exploitative—a use of exclusion to achieve exploitation. A question that bedevils antitrust is: Does antitrust address anything more? Is there only one type of practice that is anticompetitive: that which is exploitative? Does antitrust exist only to keep firms from artificially reducing market output and raising price, as captured in the triangles and rectangles of the familiar economists’ diagrams?1 Or is there an * Eleanor M. Fox is Walter J. Derenberg Professor of Trade Regulation at New York University School of Law. The author thanks Jonathan Baker, Margaret Bloom, John Fingleton, Harry First, Jonathan Jacobson, Timothy Muris, and Robert Pitofsky for their helpful comments, and Tara Koslov for excellent editorial suggestions. She thanks, also, participants at workshops at Loyola Law School-Chicago and Max Planck Institute for helpful comments, questions, and dialogue; in particular, Joseph Bauer, Paul Brietzke, Peter Carstensen, Josef Drexl, Ulrich Ehricke, Andrew Gavil, David Gerber, Ulrich Immenga, Robert Lande, James Langenfeld, Christopher Leslie, and Spencer Weber Waller. Not all agree with the views I express. I am grateful, also, for the support of the Filomen D’Agostino and Max E. Greenberg Research Fund of New York University School of Law. 1 For simplicity, I use “price-raising” to include maintaining prices at a level higher than they would otherwise be. “Output-limiting” likewise includes maintaining artificially low output. Even exploitation may not be a sufficient basis for a single-firm violation. U.S. antitrust law does not proscribe single-firm exploitative pricing. It rejects an excessive pricing violation in order not to interfere with the risk-reward system that encourages each firm to strive to be the best, and also for administrative reasons. The question in this article is whether an exploitation scenario is a necessary condition for a violation. 371 [Vol. 70 Antitrust Law Journal 372 exclusionary-practices violation, apart from aggregate consumer (or total) wealth loss?2

At the heart of this dilemma is a question of meaning. What is an “anticompetitive” practice? Different understandings of “anticompetitive” in different jurisdictions may lead to different outcomes in the analysis of mergers and dominant firm conduct that may be exclusionary but do not necessarily change the shapes of the triangles and rectangles, at least not in knowable ways. This article describes alternative touchstones used by jurists and jurisdictions to conceptualize a practice or transaction as anticompetitive. One is the microeconomic model popularly used in the United States today, which counsels no antitrust intervention unless the transaction is likely to diminish aggregate consumer or total wealth (thus, the critical

importance of the welfare triangle to show output limitation). According to this methodology, there is no “exclusionary” violation; the violation is exploitation. The second methodology begins on a larger canvas. The analyst looks at the market structure and dynamics, and asks whether the practice interferes with and degrades the market mechanism. Freedom of trade (and competition and innovation) without artificial market obstruction is presumed to be in the public interest, especially the public’s economic interest.3 Barriers must be justified. By this metric, signifi- cant unjustified exclusionary practices are anticompetitive and should be prohibited. A description of frameworks would not be complete without acknowledging a third. Some nations expand the concept of harm to competition to include harm to the competitive dynamic among small and middlesized firms. This approach tends to protect small firms from efficient competition, such as sustainable low-price competition, and therefore is protectionist. Nonetheless, I mention the third framework for two 2 I do not distinguish between decreases in aggregate consumer welfare and decreases in total (producer plus consumer) welfare. In a jurisdiction that takes a total welfare approach, consumer loss may be offset by producer gain. In that case, consumer exploitation is still a necessary condition for a violation, but it is not a sufficient condition. Both tests look only to the outcome of a particular conduct or transaction and prescribe no enforcement without negative aggregate wealth effects. The question this article poses is whether and to what extent antitrust does more. 3 See generally E.M. Graham & J.D. Richardson, Issue Overview, at 3 et seq., in Global Competition Policy (E.M. Graham & J.D. Richardson eds., 1997); see also Giuliano Amato, Introduction and ch. 3, in Antitrust and the Bounds of Power: The Dilemma of Liberal Democracy in the History of the Market (1997). The larger canvas as starting point also provides a home for theorists who stress the “freedom” value of free trade, such as the Ordoliberals of the German Freiburg school. See David J. Gerber, Law and Competition in Twentieth Century Europe: Protecting Prometheus ch. 7 (1998); Amato, supra, at 40–43. 2002] What Is Harm to Competition? 373 reasons. First, it describes a part of some nations’ definition of harm to competition. Competition and “unfair” competition are not distinguished. Second, the third possibility stands as a cautionary story to jurisdictions that adopt the larger-canvas model: enforcement of the law under a “protection of the market” paradigm can spill over into protection of competitors, and jurisdictions that would abhor (or even just disfavor) that result must at least give attentive regard to defendants’ stories that their conduct helps the market work better and therefore should not be proscribed. Part II of this article is about U.S. law: how the law came to adopt the output paradigm, and how, nonetheless, in this author’s view, there is some discomfort among American jurists and policy makers in not reprehending significantly exclusionary and unjustified conduct by dominant firms. I analyze the Microsoft case as an example. Part III of this article treats European Community law, which begins analysis on the larger canvas (and currently may be under some pressure to retrench). The European Union is at a crossroads as to whether it will move to the dominant U.S. paradigm or will, as an alternative, take more seriously firms’ proof that their “competition-distorting” conduct helps the market work more effectively. Part IV places on the map the framework that

protects small (often indigenous) firms from the windstorms of large (often leveraged) efficient competitors. Part V restates on two charts the choices for conceptions of “harm to competition.” The first chart is descriptive. It recognizes all three conceptions of harm to competition— output limitation, dynamic harm to the market process, and protecting rivalry among small and middle-sized firms. The second chart is normative. It takes the view of antitrust minimalists or political libertarians (fear of antitrust intervention, not of private power) and acknowledges only two possibilities: reprehending conduct that limits output, and protecting competitors from competition. In a brief conclusion, the article reflects on the significance of divergent conceptions of harm to competition for antitrust in globalized markets. Some might assert that the real debate regarding the first two perspectives is a debate only about proof of output limitation, not whether output limitation is the touchstone of competitive harm. When, however, the standard required for proof of output limitation is so low that a court may condemn conduct even though reduced output is a mere remote possibility—as often occurs in exclusionary conduct cases—one must suspect that the ground for prohibition is something other than output limitation. Debate over the concept of competitive harm can be disguised as a debate over proof. A jurisdiction’s assignment of the burden of [Vol. 70 Antitrust Law Journal 374 justification to the defendant often appears to reflect the court’s understanding that the conduct proved by the plaintiff is anticompetitive, rather than the court’s division of the functional task of proving output limitation. If, upon the plaintiff’s satisfying its prima facie case, one can infer a reasonable likelihood of output limitation, either in view of the facts of the particular case or as a result of economic experience with similar facts, then the problem is about proof. But if the plaintiff can make a prima facie case by proving facts that do not support such an inference, the problem is about concept.4

II. THE UNITED STATES This section reviews the evolution of U.S. antitrust law. It includes older—and overruled—U.S. case law because, inmy view, while in earlier times policy makers and jurists failed to appreciate the possible negative effect of aggressive enforcement of the law on consumer well-being, the earlier U.S. antitrust law contained “truths” about antitrust that still resonate around the world. The trouble with pre-1980s U.S. antitrust was not necessarily that it cared about abuses of power unlinked to reductions of aggregate wealth or output, but that it offered no limiting principle that would restrain enforcement that harmed consumers. U.S. antitrust of the 1980s and forward cured that problem, not with a tailored limiting principle,5 but with a bold new model that, like the Washington Consensus, is especially beneficent to the well-endowed and the mobile.6 4 See, as to the debate regarding proof of probable output restraint, Timothy J. Muris, The FTC and the Law of Monopolization (Muris I ), 67 Antitrust L.J. 693 (2000); David A. Balto & Ernest A. Nagata, Proof of Competitive Effects in Monopolization Cases: A Response to Professor Muris, 68 Antitrust L.J. 309 (2000); Timothy J. Muris, Anticompetitive Effects in Monopolization Cases: Reply (Muris II ), 68 Antitrust L.J. 325 (2000).

Post-Chicago economics addresses the question: When do the facts prove probable output restraint? By relaxing theoretical assumptions about the robustness with which markets work and giving greater attention to actual behavior and context, a number of distinguished economists are offering evidence of probable or possible price rises in areas in which robust-market assumptions reveal no concerns. See, e.g., Jonathan B. Baker, Mavericks, Mergers, and Exclusion: Proving Coordinated Competitive Effects Under the Antitrust Laws, 77 N.Y.U. L. Rev. 135 (2002); Jay Pil Choi & Sang-Seung Yi, Vertical Foreclosure with the Choice of Input Specifications, 31 Rand J. Econ. 717 (2000); Thomas G. Krattenmaker & Steven C. Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve Power over Price, 96 Yale L.J. 209 (1986); Barry Nalebuff, Competition Against Bundles (Yale School of Management Working Paper #7 (2000); Michael Whinston, Tying, Foreclosure, and Exclusion, 80 Am. Econ. Rev. 837 (1990). 5 For example, antitrust illegality for exclusionary conduct might be trumped by a defense that the conduct was a means of responding to the market or of offering buyers a product or service they would not otherwise obtain; thus, that enforcement condemning the conduct would be inefficient. See also infra text following note 58. 6 The Washington Consensus is the understanding that deregulation, privatization, and in general economic liberalization and reliance on free markets is the right prescription 2002] What Is Harm to Competition? 375 In the United States, I observe a de facto but unacknowledged builtin steam valve—lip service to the output limitation standard but creativity in its use, as developed below. In the context of exclusionary restraints, the steam valve allows protection against unjustified but not necessarily output-limiting exclusions, in the name of output limitation. In Europe no hidden steamvalve has been necessary because the European competition system turns at least as much on preserving competitive structure and open market values as on prohibiting conduct because it has exploitative outcomes. A. A Brief History of Antitrust and Exclusionary Practices Under U.S. Law In the United States twenty-five years ago and for many years prior, “anticompetitive” or “lessening competition” was a wide term that connoted harm to themarket processes through, among other things, depriving powerless market actors of a fair right to compete. Typical statements appeared in the famous Standard Stations case,7 involving Standard Oil’s one-year exclusive contracts with 16 percent of gas stations in the western United States, against a background of similar contracts of its six leading competitors who together accounted for 40 percent of gas stations in the area: “Standard’s use of the contracts creates just such a potential clog on competition as it was the purpose of § 3 [of the Clayton Act] to remove wherever, were it to become actual, it would impede a substantial amount of competitive activity.”8 The practice, said the Court, improperly “excludes suppliers from access to the outlets controlled [by the dealers].” 9 The rule became known as the rule of “quantitative substantiality.” If competitors were blocked from substantial competitive activity or foreclosed from a substantial percentage of the market, the practice was illegal. Tying clauses were condemned even more readily than exclusive dealing contracts. The railroads in the American West sold tracts of land on the condition that the buyers ship their goods on the seller’s railroad. The Supreme Court held the contracts illegal because: “So far as the Railroad was concerned its purpose obviously was to fence out competitors, to stifle competition. While this may have been exceedingly benefi-

for economies worldwide. See, as to the Washington Consensus and its effects, Joseph E. Stiglitz, Globalization and its Discontents 20 (2002). 7 Standard Oil Co. of Cal. v. United States, 337 U.S. 293 (1949). 8 Id. at 314. 9 Id. at 298. [Vol. 70 Antitrust Law Journal 376 cial to its business, it is the very type of thing the Sherman Act condemns.”10

As for mergers under Section 7 of the Clayton Act, the Court condemned Consolidated Foods’ acquisition of Gentry (producer of dehydrated onions and garlic) because it gave Consolidated Foods, which had enforced reciprocal dealing in the past, “the advantage of a mixed threat and lure of reciprocal buying”; thus the merger gave it “the power to foreclose competition from a substantial share of the markets.”11 The Court said: We hold at the outset that the “reciprocity” made possible by such an acquisition is one of the congeries of anticompetitive practices at which the antitrust laws are aimed. The practice results in “an irrelevant and alien factor,” . . . intruding into the choice among competing products, creating at least “a priority on the business at equal prices.”12

In Fashion Originator’s Guild of America v. FTC,13 the Court prohibited a pervasive combination of makers and sellers of designs, designer fabrics, and designer garments, which was intended to prevent the designs from falling into the hands of “pirates” (style copiers). Defendants used boycotts and threats to boycott, as well as other strong-arm tactics, against any enterprise that “pirated” designs or sold copied designs. The FTC did not find that the combination raised prices, limited production, or degraded quality in the market; but FOGA got no benefit from this fact. The Court said: “[A]ction falling into these three categories does not exhaust the types of conduct banned by the Sherman and Clayton Acts.”14

The evil was that the combination exercised coercive control and power to exclude. It was equivalent to a private government. Why did the U.S. Supreme Court of the 1940s, ’50s and ’60s (from 1953 forward, the Warren Court) attempt to safeguard markets from exclusionary practices, even without proof that the conduct harmed consumers? As any close reader of the large body of Supreme Court cases will know, it was not because the shift of market share to dominant firms could cause prices to rise and that the law was precautionary in this respect. It was because antitrust law was a mechanism to preserve the competitive functioning of the market, to minimize privilege and power, 10 Northern Pac. Ry. Co. v. United States, 356 U.S. 1, 8 (1958). 11 FTC v. Consolidated Foods Corp., 380 U.S. 592, 593 (1965) (quoting from the FTC’s administrative opinion). 12 Id. at 594. 13 312 U.S. 457 (1941). 14 Id. at 466. Output would not have been limited if the discounters did not need to sell copied designs. 2002] What Is Harm to Competition? 377 and to safeguard competition on the basis of merit. The law sought to promote openness, opportunity, and freedom from coercion by firms with power.15 At mid-century and in the two decades thereafter, it was not generally believed, or was not a worry, that conscious regard for

openness and opportunity might handicap efficient businesses, hurt American consumers, and hold back American firms in the (yet to be) global economy. Rather, openness and protection against power were equated with market efficiency. But much of the Warren Court jurisprudence has long since been overruled.16 Beginning particularly in 1980–81, there emerged a new paradigm for American antitrust: a rule of non-intervention in the absence of conduct that is likely to reduce consumer surplus. Loss of competitive opportunity on themerits is no longer deemed “anticompetitive” underU.S. law. Law against “mere” exclusionary harms is denigrated as law that protects inefficient competitors and harms consumers17 (even though this is not necessarily the case).18

How did the narrow U.S. perspective win the competition for the “market”; i.e., the competition to be the U.S. antitrust paradigm? And what are the implications of this view for a world in which most nations still reserve a place at the center of antitrust for exclusionary and coercive practices by firms with power? B. One View of Chicago By the end of the 1970s, the U.S. antitrust laws were robust—to many observers, too robust. They prohibited many “normal” business transactions; they had over-expanded, in favor of helping the underdog and dispersing power. The same critique was leveled against other bodies of U.S. law, such as civil rights law—there was too much law and it handicapped American business. The 1980s ushered in an era of conservatism, under the leadership of President Ronald Reagan. The Reagan Administration set about to cut back the law that regulated business. 15 See Eleanor M. Fox, The Modernization of Antitrust—A New Equilibrium , 66 Cornell L. Rev. 1140 (1981). 16 Fashion Originators’ Guild , 312 U.S. 457 (decided before the Warren era), has not been overruled. Because the case concerns a combination of competitors not to compete, it gets the benefit of the per se rule. 17 See Department of Justice (DOJ) Press Release, Statement by Assistant Attorney General Charles A. James on the EU’s Decision Regarding the GE/Honeywell Acquisition ( July 3, 2001), available at http://www.usdoj.gov/atr/public/press releases/2001/8510.htm. 18 For example, a rule of law might, without inefficiency, ban monopolists’ restraints that fence out competition on the merits and are not efficiency-justified. [Vol. 70 Antitrust Law Journal 378 But how to cut back the antitrust laws without a repealer vote of Congress (which would not succeed)? How far, and by what rhetoric and what concept? There was a concept that nicely fit the mission— a concept that would minimize antitrust as far as possible while still acknowledging the existence of the antitrust statutes. The solution was a rule of non-intervention unlessmarket conduct was provably inefficient, and “inefficient” was to be given the following narrowest-possible meaning: the conduct must confer market power that would be used to limit output of the product or service, and the conduct must not be justifiable as an attempt to serve the market. This was the approach associated with the Chicago School. Cartels would, naturally, be the one clear target of the law. Even this target was a concession to the existence of the antitrust statutes. The most libertarian wing of the Chicago School would have preferred no antitrust law at all, unless it prohibited only government

interference with the freedom of business. Adherents believed that cartels would self-destruct faster than government intervention could catch them. And the most conservative wing of Chicago School believed that competitor collaboration was usually reasonably necessary to sustain a healthy and productive economy.19

In the new era, cartels, and little else, would be prohibited. In particular, it was explained that practices that had been regarded as exclusionary and foreclosing were probably neither, at least not in any sense meaningful to competition. First, efficient firms would find a way to maneuver around exclusionary restraints. Second, even if a practice tended to exclude rivals from access to inputs or outlets, the practice was probably efficient and was seldom output-limiting. Thus, Assistant Attorney General Baxter, in engineering the AT&T break-up (which efficiently freed the AT&T long distance function from regulation), was unconcerned by the continued freedom of Western Electric (the hardware subsidiary) to be the exclusive supplier to AT&T and the new Baby Bells of their hardware needs; themarket would work.20 Complaints about the foreclosing effects of exclusive dealing were regarded as simply the cries of competitors who wanted the government to protect them from competition.21 19 See, e.g., John S. McGee, In Defense of Industrial Concentration (1971). Indeed, the word “cartel” was barely known to U.S. antitrust jurisprudence until 1981. Before 1981, price fixing was called just that; and it was one important part but not the sole core of antitrust law. 20 See United States v. AT&T, Competitive Impact Statement, 47 Fed. Reg. 7170, 7178–79 (Feb. 17, 1982). 21 There was more than a germ of truth to the criticisms. Antitrust of the 1960s had so over-expanded in favor of the underdog that the law tended to protect inefficient competitors. Antitrust plaintiffs sometimes won lawsuits waged against competition itself , and in 2002] What Is Harm to Competition? 379 The 1980s victory of the Chicago School was more a victory of economic libertarianism and political conservatism than of maximization of a microeconomic welfare function. “Consumer welfare” was the label given for the raison d’eˆtre of the new regime. Consumer welfare calculated as aggregate consumer surplus was a limiting principle on antitrust enforcement;22 it stood for the admonition that the law must not be invoked unless a challenged practice by a particular firm decreased aggregate consumer welfare. (Given the presumption of business effi- ciency, it seldom did.) In its own right, however, a consumer welfare paradigm is not necessarily a rule of non-intervention. “Consumer welfare” and “output limitation” became words that anchored the conversation of antitrust; they became necessary to the antitrust discourse. Eventually they took on an elasticity. Thus, Judge Diane Wood was able to proclaim in Toys “R” Us that when the popular toy retailer pressured the big toy makers to shift the supply of the most coveted toys from no-frills warehouse clubs to the popular retailer alone, Toys “R” Us had effected an output limitation to the warehouse clubs,23 even while Justice Antonin Scalia could proclaim that a producer’s cutoff of a well-performing discount distributor did no antitrust harm at all because “just cutting off a discounter” does not provide a price signal around which producers can cartelize, and

without a cartel there could not have been an output limitation.24

By the end of the 20th century there was thus a considerable range for maneuver by enforcers and jurists in claiming that exclusionary concountless private cases, defendants settled for significant sums of money because it was the better part of wisdom in view of the risks. The critics of antitrust—those who would have voted against any antitrust law—capitalized on this phenomenon to push the pendulum well beyond a proportional response to the law’s overbreadth. 22 See Eleanor M. Fox & Lawrence A. Sullivan, Antitrust—Retrospective and Prospective: Where Are We Coming From? Where Are We Going?, 62 N.Y.U. L. Rev. 936 (1987); Eleanor M. Fox, Chairman Miller, The Federal Trade Commission, Economics and Rashomon, 50 L. & Contemp. Probs., Autumn 1987, at 33; see also Robert Bork, The Antitrust Paradox: A Policy at War with Itself ch. 6 (1978) (consumer welfare stands for the sum of consumer and producer surplus; antitrust must limit itself to conduct that limits output, thus decreasing the sum). “Consumer welfare” was a sweeter pill than a transparent attack on antitrust. 23 Toys “R” Us, Inc. v. FTC, 221 F.3d 928, 936 (7th Cir. 2000) (vertical aspect). Toys “R” Us provided a full line of toys, helped to pioneer “hot” toys, and priced low, but its prices were not as low as the no-display warehouse clubs. One might see the shift in business pattern as efficiently and permissibly keeping the Toys “R” Us price higher than the warehouse price so that Toys “R” Us could continue its merchandising services. See also United States v. Visa U.S.A. Inc., 163 F. Supp. 2d 322, 342 (S.D.N.Y. 2001) (MasterCard/ Visa’s policy of exclusivity with banks “limited the output” of American Express) (appeal pending). 24 Business Elecs. Corp. v. Sharp Elecs. Corp., 485 U.S. 717, 726–27 (1988) (manufacturer’s cut-off of discounter pursuant to agreement with higher priced retailer). [Vol. 70 Antitrust Law Journal 380 duct did or did not meet the test of harming consumer welfare and therefore did or did not deserve to be called “anticompetitive.” At one end of the spectrum stood the U.S. Federal Trade Commission under the leadership of Robert Pitofsky during the Clinton administration. Serious market exclusions, especially coercive exclusions, concerned the FTC. A paradigmatic case was the proceeding against Intel, the dominant supplier of the microprocessing chip that is the nervous system of most personal computers. When sued by certain of its customers for infringing their intellectual property, Intel had cut them off from the flow of technical information they needed to incorporate the Intel chip into their hardware. The FTC prevailed upon Intel to cease and desist from discriminatory cutoffs.25 Also typical was the FTC proceeding and order against Toys “R” Us, noted above, which enjoined the retailer from coercing toy makers to limit their patronage of the warehouse clubs, as well as FTC decrees in telecommunications and media mergers and alliances, which imposed obligations to give nondiscriminatory market access to competitors.26

At the other end of the spectrum were the antitrust minimalists, who would withhold antitrust intervention in the absence of credible proof that conduct would increase market power, raise price, and limit output, and those, somewhat less minimal, who would challenge, also, conduct that reasonably threatened either to maintain or create artificially low output.27 The following section explores the minimalist perspective in greater detail, using the writings of one prominent antitrust scholar as an example. C. Wrong on the Law, Wrong on the Facts, Wrong on Policy Timothy Muris, while a professor at George Mason University School

of Law and before his appointment in 2001 as Chairman of the Federal Trade Commission, wrote an article entitled, “The FTC and the Law of 25 See Intel Corp., FTC Docket No. 9288 (consent order to cease and desist, Aug. 3, 1999), summarized at Trade Reg. Rep. (CCH) [Transfer Binder 1997–2001] ¶ 24,575. But compare Intergraph Corp. v. Intel Corp., 195 F.3d 1346 (Fed. Cir. 1999). See infra note 31. 26 Toys “R” Us, Inc., FTC Docket No. 9278 (cease and desist order), summarized at Trade Reg. Rep. (CCH) [Transfer Binder 1997–2001] ¶ 24,516, aff’d, 221 F.3d 928 (7th Cir. 2000); America Online, Inc. and Time Warner, Inc., FTC Docket No. C-3989, (consent order to cease and desist, Apr. 17, 2001), summarized at Trade Reg. Rep. (CCH) [Transfer Binder 1997–2001] ¶ 24,835; Time Warner, Inc., FTC Docket No. C-3709 (consent order to cease and desist, Feb. 3, 1997), summarized at Trade Reg. Rep. (CCH) [Transfer Binder 1993–1997] ¶ 24,104; see also Byron E. Fox & Eleanor M. Fox, Mergers That Impair Market Access, in Corporate Acquisitions and Mergers ch. 11 (2002). 27 See Part C infra; Muris II, supra note 4. 2002] What Is Harm to Competition? 381 Monopolization.”28 The article critiqued the Pitofsky FTC’s initiatives in monopolization cases. Setting the stage, Muris stated that antitrust concern about competitors’ cartels is sound and laudable, but that single firms, even monopolists, only rarely harm competition. The FTC, he said, “proposes to alter what many believe to be the basis for [single- firm] liability. . . . The agency appears to believe that in monopolization cases government proof of anticompetitive effect is unnecessary.”29

Muris helpfully stated precisely what he meant by “anticompetitive effect” and “harm to competition,” phrases that he used interchangeably with one another and with “harm to consumers.” Exclusionary behavior, to qualify as anticompetitive, must be, he said: “reasonably . . . capable of making a significant contribution to creating or maintaining monopoly power.” . . . Monopoly power is a concept that requires analysis of competitive effect. In both law and economics, such power is defined as the ability to raise price and restrict output in an industry. . . . [I]ndirect evidence . . . [such as market share and entry conditions,] is merely a proxy for actual proof of anticompetitive effects—namely, the ability to raise price and restrict output.30

Muris proceeded to argue that the Pitofsky FTC, by not putting itself to the burden of proving that challenged conduct significantly contributed to creating, maintaining, or enhancing monopoly power, was wrong on the law and wrong on policy, and that in the Intel case it was also wrong on the facts, for a dominant firm’s termination of a pre-existing collaboration in a network industry is not output-limiting.31

I focus here particularly on the wrong-on-the-law argument, for in this section of his article Muris essentially argues that the Supreme Court has adopted his definition of “anticompetitive.” If he is correct, there 28 Muris I, supra note 4. 29 Id. at 694. 30 Id. at 696–97 (footnotes omitted) (quoting 3 Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 650c, at 69 (rev. ed. 1996). See Bork, supra note 22, at 123 (“Antitrust must content itself with the identification of attempts to restrict output and let all other decisions, right or wrong, be made by the millions of private decision centers that make up the American economy.”). 31 Muris I, supra note 4, at 716 et seq. But see Balto & Nagata, supra note 4. Intel’s cutoff of Intergraph from the know-how it needed to incorporate Intel’s chip was coercive and discriminatory—factors on which the FTC relied. Moreover, the cut-off set back Intergraph’s attempts to invent around the Intel chip and thus become a competitor of Intel. However, the cut-off by Intel apparently was not intended to discipline Intergraph as a potential challenger (in contra-distinction to Microsoft’s strategies to “cut the air off”

from Netscape in marketing its browser, see Part II.D infra), but was probably an act of hostility to and intimidation of Intergraph because Intergraph had sued Intel for patent infringement and because Intel wanted to coerce Intergraph to license its innovations to Intel. See Intergraph Corp. v. Intel Corp., 195 F.3d 1346 (Fed. Cir. 1999). [Vol. 70 Antitrust Law Journal 382 is no exclusionary offense as such; there is only an exploitative offense, although exploitation can sometimes be enabled by market exclusions. Muris correctly states that the modern U.S. antitrust cases seem to read “unfairness” out of Section 2 of the Sherman Act, and they signal that the Supreme Court would probably jettison “fairness” even in cases brought under Section 5 of the Federal Trade Commission Act (the statutory basis of the Intel proceedings), even though Section 5 expressly prohibits “unfair methods of competition.” In the last decade the Supreme Court has said that antitrust cases must be based on market harm, not unfair exclusion, and not even malicious exclusion.32 Moreover, some modern cases support Muris’s view of market harm (i.e., the conduct must limit output or maintain artificially limited output). Indeed, in its most recent antitrust decision, California Dental Association v. FTC,33 the Supreme Court refused to draw any inference of market harm from the California dentists’ trade association’s rule against advertising “reasonable prices,” “low fees,” “10% discount for seniors,” etc., because (in its view) the rule might not have been output-limiting. Advertisements of the sort prohibited could be puffery. Puffery can make patients skeptical of dentists and depress demand, the Court said. Therefore (said the Court) the dental association rules could have been output-increasing. In fact, under law that mandates a serious application of the output standard and places on the plaintiff the burden of proof, many of the old staples of antitrust jurisprudence would shift to the bin of bad law. The full impact of Muris’s perspective is revealed by his treatment of the U.S. antitrust classic, Lorain Journal.34 Lorain Journal, the only daily newspaper in an Ohio town in 1950, considered itself threatened by the opening of the only nearby radio station, WEOL. Lorain Journal refused to accept ads from its advertising customers (the small businesses in Lorain) if they placed any ads with WEOL. The local advertisers needed to advertise in Lorain Journal, and few could be expected to stray. The Supreme Court easily found a violation of the Sherman Act and enjoined Lorain Journal’s conduct, even thoughWEOL had not been snuffed out. Professor Muris states that Lorain Journal has “recently been questioned.” The radio station apparently remained profitable and was never in danger of bankruptcy.35 The market for the placement of local ads was never artificially limited. The fact that Lorain Journal “can reasonably 32 See, e.g., Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209 (1993); NYNEX Corp. v. Discon, Inc., 525 U.S. 128 (1998). 33 526 U.S. 756 (1999). 34 342 U.S. 143 (1951). 35 Muris I, supra note 4, at 715. 2002] What Is Harm to Competition? 383 be questioned undercuts the FTC’s attempt” to rely on case law to avoid proof of consumer harm, Muris writes. “These criticisms [of Lorain Journal and other cases on the basis that they condemned exclusionary

acts that did not turn out to limit output], at a minimum, reveal that there is serious question whether allegedly exclusionary practices, such as RRC [raising rivals’ costs], are in fact exclusionary [meaning, anticompetitively so].”36

A review of the cases analyzed by Professor Muris reveals that there is virtually no single-firm exclusionary practices case not involving government- imposed protections that would pass his test for proof of output limitation. That, of course, is what the reader was led to expect in the first paragraphs of his article when Muris declared that anticompetitive single-firm exclusionary conduct was indeed “rare.”37

A number of contemporary cases on exclusionary practices tend to be noncommittal if not obfuscatory in their usage of “anticompetitive.”38

Yet others openly aver that the antitrust laws protect competition, not efficiency, and that the absence of consumer harm is no obstacle to a judgment for the plaintiff.39 36 Id. 37 Id. at 693. 38 See, e.g., Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 451 (1992) (refusal to dismiss independent service operators’ (ISOs’) claim that Kodak illegally tied repair parts and aftermarket service to imaging machine sales because the ISOs had introduced credible evidence that Kodak could and did exploit buyers of repair services in the aftermarket; but see dictum id. at n.29); Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985) (monopolist’s unjustif ied exclusion of competitor by refusal to deal held illegal); Conwood Co. v. United States Tobacco Co., 2002-1 Trade Cas. (CCH) ¶ 73,675 (6th Cir. 2002) (dominant snuff manufacturer illegally excluded competitor by dirty tricks; although new products were introduced and output increased, plaintiff proved that output would have increased more in the absence of defendant’s bad behavior). 39 See Fishman v. Estate of Wirtz, 807 F.2d 530 (7th Cir. 1986). The case concerned a competition to purchase the Chicago Bulls, a professional basketball team. Illinois Basketball, Inc. (IBI) (with Marvin Fishman) had won a bid for the Chicago Bulls, but the bid was lost when a consortium of defendants, who controlled the only stadium in town, refused to lease the arena to IBI because members of their group wanted to buy (and eventually bought) the Bulls for themselves. Sued for Sherman Act violations, defendants argued that the Bulls and the stadium were natural monopolies, and that consumers— the fans—would face a monopolist owner of the team in any event, and were indifferent as to who managed the Bulls. The conduct, said defendants, involved only substitution of one monopolist for another; therefore, there could be no harm to competition. A panel of the appellate court, by Judge Cudahy, held for plaintiffs: the antitrust laws protect competition and the competition process, not results. It was no defense that consumers were not hurt. Thus: [Defendants] assert that “the antitrust laws do not apply where there is no consumer interest to protect.” . . . The dissent makes the same argument about consumer effect: “Antitrust law condemns results harmful to consumers. . . .” We agree that the enhancement of consumer welfare is an important policy— probably the paramount policy—informing the antitrust laws. . . . Some Supreme [Vol. 70 Antitrust Law Journal 384 Microsoft fits the category of noncommittal if not obfuscatory analysis of exclusionary acts. In the district court, Judge Thomas Penfield Jackson was torn between the two conceptions of exclusionary restraints (i.e., with and without a price rise), and the Court of Appeals for the D.C. Circuit, struggling successfully for consensus, used conflicting premises for determining when an exclusionary restraint is anticompetitive.40

I turn to the Microsoft case in some detail. While it is sui generis in many respects, it is typical in its ambivalence regarding seriously exclusionary practices that may not have output effects.

D. Microsoft, American Antitrust, and the New Economy If proof of output limitation is a necessary condition to a successful antitrust case, can there ever be a single-firm violation in the new economy? In the new economy, the major competition takes the form of innovation competition, and it is nearly impossible to prove the counterfactual of what technologies, if any, would have succeeded had they not been stunted by exclusionary practices. Moreover, in the new economy, where often there are network effects, the competition may be for the market, not in the market; one firm might replace another. This article asserts that the findings of violations in Microsoft are consistent not with output theory—unless in a most attenuated way—but with a theory of defense of the market against significant, unjustified distortions. We do not and could not know whether Microsoft’s exclusionary contracts and threats made consumers worse off in terms of price or innovation. The court preferred to leave the future course of the market not to Microsoft but to the more impersonal forces of competition. By now we know the story well. Microsoft was the dominant firm in the market of Intel-compatible personal computer (PC) operating systems. Microsoft Windows occupied 95 percent of the market. It enjoyed (and enjoys) the benefits of network effects; it is the standard operating system for PC software applications. Applications makers who Court cases indicate that effect on ultimate consumers is, in an appropriate context, a significant consideration in analyzing a business practice to see whether there has been an antitrust violation. . . . The antitrust laws are concerned with the competitive process, and their application does not depend in each particular case upon the ultimate demonstrable consumer effect. A healthy and unimpaired competitive process is presumed to be in the consumer interest. Id. at 563. Judge Easterbrook wrote a long, vigorous dissent, disagreeing on law, economics, and policy; arguing that ownership of complements by a single group—the defendants— made the preclusion of competition [not his words] good for consumers; and calling all unjustified exclusionary acts that do not result in specific harm to consumers just torts. 40 United States v. Microsoft Corp., 87 F. Supp. 2d 30 (D.D.C. 2000), rev’d in part, aff’d in part, 253 F.3d 34 (D.C. Cir.), cert. denied, 122 S. Ct. 350 (2001). 2002] What Is Harm to Competition? 385 write forWindows realize the benefits of the built-in widespread audience through the huge installed base. If they write for any other operating system, they face a high risk of insufficient sales. Microsoft, the operating system monopolist, perceived a threat from middleware—platformsoftware, which, if successfully designed and commercialized, could interoperate with multiple PC operating system platforms. Applications makers might then write their applications to the middleware, and the distinctive qualities of Microsoft Windows would become entirely unimportant. This development would have “commoditized” Windows. Microsoft was alarmed. The threat to Microsoft came from two main sources—Netscape, through its browser, Navigator, and Sun Microsystems, which had developed the Java language with the intent to make it a cross-platform language. Perhaps being paranoid, as it claims it was, Microsoft believed that Netscape and Java were on the brink of developing and commercializing the feared middleware innovation. (In fact they were probably far from the brink.) Themiddleware innovation depended on, among other

things, Netscape’s Navigator’s achieving a critical mass of operating system users, and Java’s cross-platform development. Microsoft set about to remove the threat. Microsoft developed its own browser, Internet Explorer (IE), and bundled its browser with its operating system. Through various contracts with PC original equipment manufacturers (OEMs), Internet service providers, Internet software vendors, and others, and through threats and deceptions, it closed off the most efficient channels for Netscape to disseminate its browser and it sabotaged the development of crossplatform Java. In the district court, Judge Jackson declared that the browser/operating system package was a tie-in illegal per se under Section 1 of the Sherman Act.He dismissed the exclusionary contracts claims on grounds that Section 1 requires total exclusion41 and Netscape was not totally 41 Microsoft, 87 F. Supp. 2d at 53–54 (validity of the holding that illegality requires total exclusion is subject to doubt). Perhaps more important to an appreciation of the narrow or schizophrenic law on exclusionary practices, Judge Jackson, in a 1998 decision, dismissed the states’ claim that Microsoft’s use of its monopoly power in operating systems to gain advantages not on the merits in the browser market violated § 2. Judge Jackson held that there is no “mere leveraging” violation. Applications (the browser) and the operating system are complements, and buyers will pay only so much for the package; there is only one available monopoly profit. Therefore, mere leveraging can only hurt competitors but cannot hurt consumers. United States v. Microsoft Corp., 1998-2 Trade Cas. (CCH) ¶ 72,261 (D.D.C. 1998). [Vol. 70 Antitrust Law Journal 386 excluded; it had other avenues to reach users. He held that Microsoft attempted to monopolize the browser market. He held numerous acts and agreements by Microsoft violative of Section 2 of the Sherman Act as part of a monopoly maintenance violation; but even those acts and agreements he deemed not sufficiently exclusionary to violate Section 1. Famously, he ordered that Microsoft be broken in two. The Court of Appeals for the D.C. Circuit sat en banc. Per curiam, the court reversed and remanded the tie-in violation,42 reversed the finding that Microsoft had attempted to monopolize the browser market on grounds that the plaintiffs had not proved a browser market and had not proved barriers to entry into browser making, and it dismissed the attempt claim; it reversed the break-up remedy; and it remanded the tying claim and the question of appropriate remedy. The court upheld most, but not all, of the monopoly maintenance claims, testing in each instance whether the plaintiffs had proved a prima facie case that the act was anticompetitive, whether Microsoft had offered a procompetitive or efficiency justification, and, if so, whether the plaintiffs had satisfied their burden to show a net anticompetitive effect.43

What did the court mean by “anticompetitive,” and did the court use that word consistently? Of particular importance to this inquiry is the court’s finding that the plaintiffs had proved no browser market. Therefore, the anticompetitive quality of a restraint could not be based on a price rise44 (or technology decline) in browsers. Most of the violations found by the court of appeals involved acts of one of two kinds: (1) those that foreclosed Netscape, the only significant

browser competitor, from efficient access to a critical mass of users (the critical mass being necessary to shift applications makers away from Microsoft’s operating system to the middleware); and (2) threats to challengers to stop developing the technologies that might have caused the feared paradigm shift to a middleware platform to which applications could be written that would work on any operating system, thus commoditizing Windows. The two most cost-effective methods of distributing browser software were pre-installation by OEMs, such as Compaq, and packaging the browser with Internet access software distributed by 42 Microsoft, 253 F.3d at 92–95. According to the court, the underlying facts involving platform software and functionalities were too complex, and the packaging of applications with operating system software too likely to respond to consumers’ needs, to merit a per se rule. Id. 43 Id. at 58–78. 44 Again, “price rise” includes maintenance of an artificially high price. 2002] What Is Harm to Competition? 387 Internet access providers (IAPs), such as America Online (AOL). Microsoft targeted and sabotaged both lines of distribution. The appellate court concluded that each of several sets of conduct was illegal. In each instance, the court asked whether the plaintiffs had made a prima facie case of anticompetitive exclusionary conduct, and if so whether Microsoft had proved out-weighing procompetitive effects. The focus here is on the nature of the exclusionary conduct that was found to be sufficient for the prima facie case. Was this exclusionary conduct anticompetitive in a consumer-harm sense, or was it “merely” a significant foreclosure and not justified as a form of vigorous competition? In general, it is clear from the court’s description of each set of exclusionary acts, as distinct from certain of its generalizations such as quoted below, that output limitation and the resulting consumer harm (including harm from stifling competitors’ innovation) was not a necessary condition for characterization of conduct as anticompetitive. In fact the court did not find that Microsoft’s conduct had the actual or probable effect of increasing or maintaining its operating system monopoly. The court seemed at first to be adopting the definition of anticompetitive effect offered by Professor Muris. It said: Whether any particular act of a monopolist is exclusionary [meaning anticompetitively so], rather than merely a form of vigorous competition, can be difficult to discern: the means of illicit exclusion, like the means of legitimate competition, are myriad. The challenge for an antitrust court lies in stating a general rule for distinguishing between exclusionary acts, which reduce social welfare, and competitive acts, which increase it.45

The court proceeded to analyze and characterize each incident, to determine whether plaintiffs had established their prima facie case and if so whether Microsoft had rebutted it. At this point the court shifted to a loose analysis wherein foreclosure became the touchstone for “anticompetitive.” Foreclosure of unspecified dimensions fromone important route of access to the browser market (although plaintiffs had failed to prove a browser market) was accepted as “anticompetitive” and thus sufficient for the governments’ prima facie case.46

Regarding Microsoft’s licenses to OEMs, such as Compaq, which prohibited the OEMs from removing desktop icons and Start menu entries 45 Microsoft, 253 F.3d at 59. 46 Ironically, just such proof is not sufficient to make a prima facie case if the charge is “mere” leveraging. See Judge Jackson’s decision dismissing the leveraging claim, discussed [Vol. 70 Antitrust Law Journal 388 from Windows, the court said that the provision “prevents many OEMs from pre-installing a rival browser and, therefore, protects Microsoft’s monopoly from the competition that middleware might otherwise present. Therefore, we conclude that the license restriction at issue is anticompetitive.”47

As to the provision that prohibitedOEMs from modifying the sequence of screens that appear when the computer is turned on (boot sequence), the court said that this provision prevented OEMs from altering the boot sequence to promote Internet access providers, many of which used Navigator. “Because this prohibition has a substantial effect in protecting Microsoft’s market power, and does so through a means other than competition on the merits, it is anticompetitive.”48

As to the restriction that OEMs could not cause any interface but Windows to launch automatically, the court said: “this type of license restriction, like the first two restrictions, is anticompetitive: Microsoft reduced rival browsers’ usage share not by improving its own product but, rather, by preventing OEMs from taking actions that could increase rivals’ share of usage.”49 (Microsoft successfully justified this restriction, however, by proving that it was necessary to prevent undermining the essence of its copyright.) As to Microsoft’s commingling the code of its own browser (IE) with its operating system code and excluding IE from the Windows’ Add/ Remove function, the court had a similar assessment: “Because Microsoft’s conduct, through something other than competition on themerits, has the effect of significantly reducing usage of rivals’ products and supra at note 41. A link between the exclusion and increase in market power is necessary; but as discussed below, see infra text accompanying note 57, that link was never proved. 47 Microsoft, 253 F.3d at 61. Because plaintiffs had failed to prove a browser market, Microsoft’s exclusion of Netscape’s browser from the efficient channels of distribution was not relevant to any claim of monopolizing a browser market. It was relevant only because Netscape needed critical-mass use of its browser if it was to have any chance to develop middleware, which, if successfully commercialized, could have eroded Microsoft’s monopoly power in the operating system market. Therefore, the degree of Netscape’s foreclosure from browser distribution channels has a different, and lesser, meaning than in a case in which browsers are the market. Even 100% foreclosure of Netscape from channels of browser distribution, let alone 100% foreclosure from channels that Microsoft had the power to clog, was not a sufficient condition for Microsoft to gain or protect power in the operating system market. The middleware innovation still would have been necessary. As to the significance of foreclosure in analyzing competitive harm, see Jonathan M. Jacobson, Exclusive Dealing, “Foreclosure,” and Consumer Harm , 70Antitrust L.J. 311 (2002). 48 Microsoft, 253 F.3d at 61–62. 49 Id. 2002] What Is Harm to Competition? 389 hence protecting its own operating system monopoly, it is anticompetitive . . . .”50

As to Microsoft’s partially exclusive contracts with Internet access providers,

which include service providers and online services such as AOL, the court found that the IAP contracts closed off a substantial percentage (no number stated) of the available opportunities for browser distribution and was therefore prima facie anticompetitive.51 A similar finding was made regarding the Internet software vendors (ISVs), although no foreclosed share (of the browser non-market) was specified and the ISV channel was much less efficient than the IAP channel. As to Microsoft’s agreements with major ISVs requiring them to promote Microsoft’s version of the Java language exclusively, which undermined the Netscape/Sun enterprise to diffuse Sun-standard Java along with Netscape’s Navigator, the court said: “Because Microsoft’s agreements foreclosed a substantial portion of the field for [Java] distribution and because, in so doing, they protected Microsoft’s monopoly from a middleware threat, they are anticompetitive.”52 Moreover, Microsoft deceived Java developers, who believed that the language as used by Microsoft would operate across platforms. This conduct “served to protect its monopoly of the operating system in a manner not attributable either to the superiority of the operating system or to the acumen of its makers, and therefore was anticompetitive.”53

Finally, the threats to Intel, pressuring it to stop supporting crossplatform Java, were also found anticompetitively exclusionary.54

On the other hand, low pricing (giving away IE “free” by bundling it with the operating system) was simply “offering a customer an attractive deal[;] . . . the hallmark of competition.”55 Also, developing a product (a high-performance Java), even though it was developed to be incompatible with rivals, was not a violation.56 The appellate court effectively held that the creation of an incompatible product is not anticompetitive, regardless of exclusionary intent and effect, because it is creation. Critically, the court acknowledged there was no finding that, but for Microsoft’s conduct, “[Netscape’s] Navigator and Java would have 50 Id. at 65. 51 Id. at 67–71. 52 Id. at 76. 53 Id. at 77. 54 Id. at 77–78. 55 Id. at 68. The plaintiffs had not appealed this ruling to the court of appeals. 56 Id. at 75 (reversing district court). [Vol. 70 Antitrust Law Journal 390 developed into serious enough cross-platform threats to erode the applications barrier to entry [into the market for PC operating systems]” and thus constrain Microsoft’s power.57 The court held that such a finding was not necessary, at least in a government enforcement action. Navigator and Java were nascent threats to Microsoft. The court said it was enough that Microsoft’s exclusionary conduct, at the time it was undertaken, reasonably appeared capable of making a significant contribution to maintaining Microsoft’s monopoly power; but no finding was made that this was so. Thus, the link between the exclusionary conduct and possible output limitation was never forged. How, then, can we generalize the D.C. Circuit’s usage of “anticompetitive”? The conduct was all exclusionary, not exploitative (althoughMicrosoft, paranoid as it said it was, may have thought that it was protecting its

monopoly). The critical ingredient in the court’s characterization of each conduct set was the quality of the conduct: Did the conduct serve the market or merely exclude? Was it “competition on the merits” or exclusion for the sake of exclusion? It was perhaps a misnomer for the court to say, at numerous points, “this conduct had a substantial effect in protecting Microsoft’s market power”—for, finally, we are told that the court did not know, and that it is fine to be agnostic about this unproved proposition. Indeed, if Microsoft had proffered compelling evidence (it thinks it did) that Netscape and Java were not real middleware threats and, therefore, that Microsoft’s conduct could have had no effect in maintaining its market power in the operating system market, it appears that the court would have dismissed the proffer as not relevant. We might thus interpret the Microsoft holding as follows: Conduct that intentionally, significantly, and without business justification excludes a potential competitor from outlets (even though not in the relevant market), where access to those outlets is a necessary though not sufficient condition to waging a challenge to amonopolist and fear of the challenge prompts the conduct, is “anticompetitive.” At first blush, the court’s articulated principles may seem to adopt Professor Muris’s test, with differences relating to the level of required proof that the exclusion caused or probably would cause maintenance of too-low a level of output of operating system software.58 However, a 57 Id. at 79. 58 But perhaps the dispute on causation is no quibble. See discussion of Intel, supra text accompanying note 25, text following note 30, and note 31. If Intel had cut off Intergraph’s supply of information because it feared that Intergraph would develop a competing technology, but there was no proof that Intergraph was likely to have succeeded or that 2002] What Is Harm to Competition? 391 reading of the court’s treatment of each set of exclusionary practices suggests that unreasonable and unjustified exclusionary conduct by a dominant firm is anticompetitive59 and, if sufficiently exclusionary, it is illegal unless justified by pro-competitive, pro-efficiency offsets; except that low pricing and product design change are always presumed procompetitive even if they have significant exclusionary effects. I do not claim that the Microsoft test (or my interpretation of it) is the test for exclusionary practices under U.S. law. Rather, I claim that it is one test in fact applied by courts to conduct that significantly forecloses competition on the merits and is not efficiency justified. When this test is applied, it may be dressed up with a story of probable lower output and higher prices.60

To say that courts are applying a truncated rule that allows us to infer output limitation does not satisfy the inquiry, for if there is no good reason to predict output limitation in the particular case or like cases, the “inference” is a legal presumption, not a factual inference. The conclusion of harm must therefore be based on another notion. For example, it may be based on the assumption or perception that markets are more likely to reward merit if they are not clogged by substantial unjustified exclusions. Moreover, although we may not know the direction in which “open” competition will take us, we may prefer to let the

chances of competition—rather than the strategies of the dominant firm—take us there. The competing test is Professor Muris’s test; and that is the one that commands greater support in U.S. law today. Indeed, American adherence to the output test, and the fear that any other test is rudderless and will quickly spill over into protection of inefficient competitors at the expense of consumers, formed the platform for the U.S. antitrust enforcers’ challenge to the European Commission after its prohibition of the General Electric Company’s acquisition of Honeywell.61 European jurisprudence does not reflect this fear. no other firms posed a similar or greater threat, the case would have been quite similar to Microsoft, but Professor Muris’s test presumably would not have supported an inference of causation of consumer harm. 59 One might wish to add: as long as the target is a possible potential competitor of the defendant, the defendant was worried about its threat, and the exclusion made the threat more remote. This interpretation, however, puts more weight on intent and perception, as distinguished from probable effect, than contemporary courts normally allow. Moreover, if the innovation competition was for the market—that is, competition to be the new monopolist—even successful competition would not change output levels. This was Judge Easterbrook’s point in Fishman v. Estate of Wirtz, discussed supra note 39. 60 See supra cases cited in notes 26 and 38. 61 See infra note 83 and text thereafter. [Vol. 70 Antitrust Law Journal 392 III. THE EUROPEAN UNION A. Introduction The European Union’s treatment of exclusionary practices is sympathetic with the one theme of the 1960s–1970s’ American jurisprudence that had a lasting resonance. That is: Competition laws protect the competitive structure and dynamic of the market.62 They protect openness of and access to markets, and the right of market actors not to be fenced out by dominant firm strategies that are not based on competitive merits.63 Protection of this competition process is believed likely to preserve incentives to compete and to serve both consumers and efficient market actors. Thus, Competition Commissioner Mario Monti writes: [E]nshrined in the Treaty . . . [is] “an open market economy with free competition.” Since its adoption more than 40 years ago, the Treaty acknowledges the fundamental role of the market and of competition in guaranteeing consumer welfare, encouraging the optimal allocation of resources and granting to economic agents the appropriate incentives to pursue productive efficiency, quality and innovation. Personally I believe that this principle of an open market economy does not imply an attitude of unconditional faith with respect to the operation of market mechanisms. On the contrary, it requires a serious commitment—as well as self-restraint—by public powers, aimed at preserving those mechanisms.64

The Treaty establishing the European Communities (Treaty of Rome, 1957) contained, from the start, two antitrust articles: Article 85 (now 62 See Doris Hildebrand, The European School in EC Competition Law, 25 World Competition L. & Econ. Rev. 3, 7 (2002). 63 See supra Part II.A. The EC law is sympathetic with cases such as United States v. United Shoe Machinery Co., 110 F. Supp. 295 (D. Mass. 1953) (Wyzanski, J.), aff’d per curiam, 347 U.S. 521 (1954). It is not sympathetic with Brown Shoe, Vons, or Albrecht, which protected competitors and other market actors at the expense of consumers. Brown Shoe Co. v. United States, 370 U.S. 294 (1962); United States v. Von’s Grocery Co., 384 U.S. 270

(1966); Albrecht v. Herald Co., 390 U.S. 145 (1968), overruled by State Oil Co. v. Khan, 522 U.S. 3 (1997). 64 Mario Monti, European Competition Policy for the 21st Century , in International Antitrust Law & Policy, 2000 Fordham Corp. L. Inst. ch. 15 at 257 (Barry Hawk ed., 2001). Other competition commissioners have taken a more eclectic view of the basis for the open market principle. Thus, Commissioner Monti’s predecessor, Karel Van Miert, wrote: The aims of the European Commission’s competition policy are economic, political and social. The policy is concerned not only with promoting efficient production but also achieving the aims of the European treaties. . . . To this must be added the need to safeguard a pluralistic democracy, which could not survive a strong concentration of economic power. Frontier-Free Europe, May 5, 1993 (quoted in Per Jebsen & Robert Stevens, Assumptions, Goals, and Dominant Undertakings: The Regulation of Competition Under Article 86 of the Euro- 2002] What Is Harm to Competition? 393 81) and Article 86 (now 82). Article (ex) 86, which prohibits abuse of dominance, was intended to regulate the behavior of dominant firms so that they would not take undue advantage of other market players, including buyers, sellers, and competitors. Indeed, the Treaty itself, in Article 3(1)(g), requires “a system ensuring that competition in the internal market is not distorted”—a mandate that is held to condemn unjustified exclusionary practices because they are exclusionary and thus distort the normal functioning of the market on competitive merits. B. Abuse of Dominance In the European Union, under Article 82, dominant firms have special responsibilities. The origin of this duty lies in the fact that, at the time the original six states (France, Germany, Italy, Belgium, the Netherlands, and Luxembourg) formed the European Communities, statism pervaded the states’ economies. State-owned enterprises controlled the mostly national markets. The duty, however, never was limited to state-owned enterprises, and is well embedded as a general principle into the law. As the Court of Justice said in Michelin v. Commission, the dominant firm “has a special responsibility not to allow its conduct to impair undistorted competition on the common market.”65 Moreover, it is clear from the wording of Article 82 that it was intended to regulate the conduct of dominant firms, to prevent them from unfairly using their power, not merely to prevent them from expanding or protecting their power.66

Exclusionary contracts and practices are a major form of abuse of dominance under the Treaty of Rome. In Hoffmann-La Roche,67 Roche, the dominant vitamin maker, decided to increase its manufacturing capacity. To hedge its risks, it procured an agreement with its competitor Merck whereby Merck agreed to buy fromRoche its vitamins needs above its own manufacturing capacity; in turn, Roche gave Merck a favorable price. The Court of Justice held the underlying contracts illegal because they “are designed to deprive the purchaser of or restrict his possible choices of sources of supply and to deny other producers access to the market.”68

pean Union, 64 Antitrust L.J. 443, 450 (1996)); see also Jebsen & Stevens, supra, at 443–51, 458–61. 65 Case 322/81, [1983] E.C.R. 3461, ¶ 57. 66 See Rene´ Joliet, Monopolisation and Abuse of a Dominant Position (1970); Jebsen & Stevens, supra note 64; Amato, supra note 3, ch. 5. 67 Case 85/76, [1979] E.C.R. 461. 68 Id. ¶ 90.

[Vol. 70 Antitrust Law Journal 394 Recent cases continue the theme. In Tetra Pak,69 Tetra Pak was the dominant firm in the manufacture of aseptic cartons for packaging milk and juice, and the machines that make them. Some contracts required customers to buy non-aseptic machines and cartons also from Tetra Pak (tying contracts), and some required exclusive dealing. The Court of First Instance held that the contracts were illegal. The Court of Justice affirmed. Tetra Pak’s dominant position in the related aseptic market “gave Tetra Pak freedom of conduct compared with other economic operators on the non-aseptic market, such as to impose on it a special responsibility under art. 86 tomaintain genuine undistorted competition on those markets.”70 As the Court of First Instance said: The Court of Justice has in particular ruled that, where an undertaking in a dominant position directly or indirectly ties its customers by an exclusive supply obligation, that constitutes an abuse since it deprives the customer of the ability to choose his sources of supply and denies other producers access to the market.71

Me´tropole Te´le´vision (M6) v. Commission (Me´tropole)72 illustrates precisely how the Treaty takes a broadmarket-access approach to “harm to competition” under Article 81(1), and requires a separate consideration of justifications for exemption under Article 81(3). (The enterprise may prove in justification that its conduct improves production, distribution or technological progress and gives consumers a fair share of benefits.) Recall that, if an agreement does not restrict or distort competition, it is entitled to a negative clearance; but if it does restrict or distort competition, to be valid, the agreement must be exempted. Exemptions are of limited duration and they customarily impose conditions on the parties. 69 Case T-83/91, [1994] E.C.R. II-755 (Ct. First Instance), aff’d, C-333/94P, [1996] E.C.R. I-5951 (Nov. 14, 1996). The Court of Justice substantially adopted the Court of First Instance’s judgment and reasoning. 70 Id., Court of First Instance judgment, [1995] E.C.R. II-762, ¶ 122. See also British Airways (Virgin), Case IV/D2/34.780, Commission decision of July 14, 1999, O.J. (L 30) (Feb. 4, 2000) 1, prohibiting loyalty rebates because such schemes foreclose the market and thus foreclose access by competitors of the dominant firm. Therefore, they are anticompetitive. A U.S. court took the opposite position in Virgin Atlantic Airways Ltd. v. British Airways PLC, 257 F.3d 256 (2d Cir. 2001), viewing the loyalty rebates as price competition even if they delayed and deterred market entry, and declaring that plaintiff had not alleged or presented facts sufficient to support a claim of predatory pricing. 71 Tetra Pak, supra note 70, ¶ 137. Tetra Pak, supra, and Michelin, supra note 65, are cited as “settled case-law” in Tetra Laval BV v. Commission, Case T-502 (Ct. First Instance Oct. 25, 2002), available at http://curia.eu.int.jurisp., ¶ 157 (annulling Commission prohibition for, inter alia, lack of proof that the merged firm would exercise its leverage, e.g., by tying, bundling, forcing, or loyalty rebates). 72 Case T-112/99, [2001] E.C.R. II-2459. 2002] What Is Harm to Competition? 395 In Me´tropole, six major producers of television programming created a partnership, Television par Satellite (TPS). The partnership agreement gave TPS a ten-year exclusive right to broadcast the four general-interest channels produced by the partners, and a right of priority on the specialinterest channels that the partners produced. The EuropeanCommission granted an exemption for various clauses, including the exclusivity clause and the right of priority, cutting back their duration to three years. The partnership sought to annul the decision granting the exemption on

grounds that any negative effects were outbalanced by positive effects; therefore the joint venture did not harm competition and all of the clauses were entitled to a negative clearance. TPS cited, among other things, the fact that two new entrants—Canal Satellite and AB Sat—were launched on the TV satellite market without need to broadcast the channels promised exclusively to TPS. The Court of First Instance rejected TPS’s argument. It said: 47 The applicants submit that, in reaching its conclusion in the contested decision that the exclusivity clause and the clause relating to the special-interest channels constitute restrictions of competition within the meaning of Article 85(1) of the Treaty, the Commission relied on incorrect assessments and misapplied that provision. * * * 64 It is in fact manifest that, as only TPS is authorised to transmit the general-interest channels owing to the exclusive rights which it enjoys, the competitors of TPS are denied access to the programmes which are considered attractive by numerous French television viewers. * * * 66 In the light of the foregoing, the applicants have not showed that the Commission relied on erroneous assessments in concluding that the exclusivity clause restricted competition within the meaning of Article 85(1) of the Treaty.73

The principle by which the European Court condemns exclusionary practices by dominant firms, unless justified, is often phrased as a dynamic one: the right of market actors to enjoy access to the market on the merits. It is a principle of freedom of non-dominant firms to trade without artificial obstacles constructed by dominant firms, and carries an assumption that preserving this freedom is important to the legitimacy of the competition process and is likely to inure to the benefit of all market players, competitors and consumers. The difference of focal point in the EU and the United States has the potential to produce different outcomes in the pending cases against 73 Id. [Vol. 70 Antitrust Law Journal 396 Microsoft, especially regarding the issues of bundling and duty to disclose technical information to facilitate interoperability. In the U.S. Microsoft case, the bundling issue arose under the rubric of tying. Judge Jackson held that Microsoft’s conduct in tying its browser to its operating system was illegal per se. The appellate court reversed this holding, noting that packaging applications with platform software might serve consumer welfare, and it remanded the claim to the district court under a rule of reason.74 The U.S. Department of Justice withdrew the claim rather than retry it,75 possibly because it did not wish to win. No claim was made, nor could it have been under current interpretations of U.S. law, that, simply because Microsoft was a monopolist controlling an industry standard, it had a general duty under Section 2 of the Sherman Act to disclose sufficient proprietary information to facilitate interoperability of applications software with Microsoft’s operating system. Indeed, we will recall, the appellate court held that even purposeful creation of incompatibilities through product design (i.e. the Java language, which Microsoft altered for Windows to defeat Sun Microsystems’ plans for a cross-platform language) does not run afoul of Section 2 because it falls

into the prophylactically-protected category of innovation.76

Meanwhile, the European Commission filed its own statement of objections and opened proceedings against Microsoft. In the statement of objections, the Commission charged that Microsoft had abused its duty to facilitate interoperability by (as reported in the press) “withholding technical information that rivals needed to allow their software to run smoothly with Microsoft’s industry-standard Windows operating system” and that it “had illegally bundled its media-playing software with Windows to undermine competition in the fast-growing new market for online music and video software.”77 On both points, unless the conduct is justi- fied, EC law may support the finding of an abuse of dominance because 74 One must distinguish the conduct by which Microsoft purposely and unjustifiedly commingled browser code with operating system code so that the browser could not be removed without degrading the system. This was one of the several illegal acts. See supra text accompanying note 42. 75 The plaintiff states followed suit. 76 Microsoft, 253 F.3d at 75. See United States v. Microsoft Corp., 147 F.3d 935, 949–50 (D.C. Cir. 1998); see also Independent Serv. Orgs. Antitrust Litig., 203 F.3d 1322 (Fed. Cir. 2000), cert. denied, 121 S. Ct. 1077 (2001); compare Image Technical Servs., Inc. v. Eastman Kodak Co., 125 F.3d 1195 (9th Cir. 1997), cert. denied, 523 U.S. 1094 (1998). 77 See Steve Lohr with Paul Meller, Microsoft Move May Hasten Settlement of European Cases, N.Y. Times, Nov. 28, 2001, at C1 (settlement was in fact not hastened); European Commission Press Release IP/01/1232, Commission Initiates Additional Proceedings Against Microsoft, Aug. 30, 2001, available at http://europa.eu.int/rapid/. . .=gt&doc=IP/ 01/1232 0 AGED&lg=EN&display=. 2002] What Is Harm to Competition? 397 of the foreclosures and the violation of the principle of openness,78 while U.S. law supports the freedom of Microsoft’s action.79

C. Mergers Mergers, too, may be price-raising, exclusionary, or both. The European Merger Regulation apparently imports the spirit of Article 82 case law intomerger jurisprudence in cases of threatened exclusionary effects. When a merger “creates or strengthens a dominant position as a result of which effective competition would be significantly impeded,” the merger runs afoul of the Merger Regulation.80 When a merger creates a market structure that offers leveraging opportunities likely to inflate the share of a dominant or near-dominant firm by empowering the firm to preempt significant opportunities of competitors, the merger may be seen as creating or strengthening a dominant position, i.e., creating a situation that facilitates abuse of dominance. De Havilland,81 Boeing/ McDonnell Douglas,82 and GE/Honeywell83 are all examples of this principle. Such a merger may sometimes, in addition, be seen as price-raising. GE’s proposed acquisition of Honeywell was seen as both. 1. The Saga of GE/Honeywell General Electric Company is the world’s largest producer of large and small jet engines for commercial and military aircraft. It and a 50-50 joint venture with SNECMA supply more than half of all engines for large commercial jets. The engine market is concentrated, with Pratt & Whitney and Rolls-Royce being GE’s principal competitors. GE Commer- 78 See supra text accompanying and following note 66; Eleanor M. Fox, Monopolization and Dominance in the United States and the European Community—Efficiency, Opportunity, and Fairness, 61 Notre Dame L. Rev. 981, 1014–15 (1986) (description of U.S./EC clash in the IBM case).

79 See supra text accompanying note 76. Noting the possibility that the Commission might take a different tack in the European Microsoft case, Charles James, U.S. Assistant Attorney General for Antitrust, “issued a veiled warning to the European Commission” against use of legal arguments that U.S. courts have rejected. Global Competition Review Briefing, May 20, 2002, available at http://www.global-competition.com/headlnes/archive/2002/ apr may/hdln mnu.htm. 80 Council Regulation 4064/89, 1989 O.J. (L 395) 1, corrected version, O.J. (L 257) 14, amended, Council Regulation 1310/97, 1997 O.J. (L 180) 1, corrected version, 1998 O.J. (L 40) 17, effective March 1, 1998. The Merger Regulation is printed in G. Bermann, R. Goebel, W. Davey & E. Fox, European Union Law, Selected Documents (2002), as Competition Doc. No. 7, p. 531, and is available on the Web site for the Competition Directorate at http://europa.eu.int/comm/competition/mergers/legislation/. 81 Commission decision of Oct. 2, 1991, [1992] 4 C.M.L.R. M2. 82 Commission decision of July 30, 1997, 1997 O.J. (L 336) 16. 83 Commission decision of July 3, 2001, Case COMP/M2220, not yet reported; appeal pending, available at http://www.europa.eu.int/comm/competition/mergers/cases/deci sions/m2220 en.pdf. [Vol. 70 Antitrust Law Journal 398 cial Aviation Services (GECAS) is one of the world’s largest aircraftleasing companies and one of the largest buyers of planes. It buys about 10 percent of aircraft, it and a sister corporation finance the purchase of airplanes, and it is an important launch customer for airplanes. Once an aircraft manufacturer chooses to incorporate a particular supplier’s engine and other elements, it normally prefers to continue purchasing the same brand because of efficiencies, such as acquired knowledge and training, as well as replaceability across a fleet. GECAS had in the past exercised its power to cause aircraft makers to incorporate GE engines.84

Honeywell International is a leading firm in the production of avionics including navigating equipment, certain nonavionic products, engines for corporate jets, and engine starters. GE and Honeywell agreed to merge, in what would have been the largest industrial merger in history. They filed their merger notifications with the U.S. authorities, who cleared the deal after requiring a spinoff of competitively overlapping engine assets.85 The parties also made premerger filings with the European Commission, among other jurisdictions. The European Commission expressed several concerns. First, the merged firm, having a large line of complementary products, would probably engage in product bundling. It was likely to lower the price of the bundle, while charging high prices for parts of the bundle offered separately. The competitors would be unable to lower the prices of their products to the same extent and would eventually abandon the market or market segment, at which time themerged firm would be in a position to raise its prices. Second, GECAS would use its buying and launching platform leverage to cause aircraft makers to shift their business to Honeywell as well as to GE products, causing a significant shift of business to GE, a weakening of the deserted competitors, and their eventual exit from the market or market segment. On July 3, 2001, the European Commission blocked the merger. The Commission declared: The combination of the two companies’ activities would have resulted in the creation of dominant positions in the markets for the supply of avionics, non-avionics and corporate jet engines, as well as [in] the

strengthening of GE’s existing dominant positions in jet engines for 84 Id. ¶¶ 132 & 452. 85 DOJ Press Release, Justice Department Requires Divestitures in Merger Between General Electric and Honeywell (May 2, 2001), available at http://www.usdoj.gov/atr/public/ press releases/2001/8140.htm. 2002] What Is Harm to Competition? 399 large commercial and large regional jets. The dominance would have been created or strengthened as a result of horizontal overlaps in some markets as well as through the extension of GE’s financial power and vertical integration to Honeywell activities and of the combination of their respective complementary products. Such integration would enable the merged entity to leverage the respective market power of the two companies into the products of one another. This would have the effect of foreclosing competitors, thereby eliminating competition in these markets, ultimately affecting adversely product quality, service and consumers’ prices.86

A typical passage of the decision, quoted below, explains how the transaction was expected to confer dominance on Honeywell as a seller of supplier-furnished equipment (SFE) through functional integration. Namely, GE was expected to make strategic use of its position as a launch platform, financer, and significant purchaser of aircraft, in a context in which airframe manufacturers were indifferent towards component selection and had an efficiency incentive to gravitate towards the use of single brand avionics and non-avionics for most of their fleet. 344 Following the proposed merger, Honeywell will immediately benefit from GE Capital’s ability to secure the exclusive selection of its SFE products on new platforms. By leveraging its financial power and vertical integration on the launch of new platforms (for example, through financing and/or through orders placed by GECAS), the merged entity will be able to promote the selection of Honeywell’s SFE products, thereby denying competitors the possibility to place their products on such new platforms. That would delay the cash inception of Honeywell’s competitors and deprive them of the necessary return to fund future investments and innovation. Honeywell’s products will, in particular, benefit from GECAS’s role as a significant purchaser of aircraft. Post-merger, GECAS will extend its GE-only policy to Honeywell products to the detriment of competitors such as Collins, Thales and Hamilton Sundstrand and ultimately of customers. Indeed, given the relative indifference of airlines towards component selection, the benefits of a non-GE offer for airframe manufacturers would become less significant than the benefits they could achieve in the form of additional aircraft purchase[s] by GECAS. . . . 346 Accordingly, GE’s strategic use of GECAS’s market access and GE Capital’s financial strength to favour Honeywell’s products will position Honeywell as the dominant supplier on the markets for SFE avionics and non-avionics products where it already enjoys leading positions. 347 The effect on rival avionics and non-avionics manufacturers will be to deprive them of the future revenue streams generated by the 86 European Commission Press Release IP/01/939, The Commission Prohibits GE’s Acquisition of Honeywell ( July 3, 2001), available at http://europa.eu.int/rapid/start/ cgi/guesten.ksh?p action.gettxt=gt&doc=IP/01/939\0\AGED&lg=EN&display=. [Vol. 70 Antitrust Law Journal 400

sales of the original equipment and spare parts. Future revenues are needed to fund development expenditures for future products, foster innovation and allow for a potential leapfrogging effect. By being progressively marginalised as a result of the integration of Honeywell into GE, Honeywell’s competitors will be deprived of a vital source of revenue and see their ability to invest for the future and develop the next generation of aircraft systems eventually eliminated. 348 Indeed, given the fact that Honeywell’s avionics and non-avionics competitors are unable to reproduce GE’s financial strength and vertical integration to any appreciable degree . . . , their limited size and financial strength would probably lead to a reduction of their competitive strength in those markets where the extension of GE’s business practices to Honeywell’s products would reduce seriously their chances to win future competitions.87

As noted, the Commission also predicted foreclosure of competitors though GE’s sale of product bundles at low and sometimes subsidized prices, the competitors’ consequent withdrawal from market segments, and GE’s eventual elevation of prices.88

U.S. Assistant Attorney General Charles James greeted the decision with disapprobation. He said that Europe prohibited the merger because it “would have been procompetitive and beneficial to consumers,” and that the Commission “apparently concluded that a more diversified, and thus more competitive GE, could somehow disadvantage other market participants.”89 Other U.S. antitrust officials elaborated on an argument that because GE’s engines and Honeywell’s avionics were complements (aircraft makers need both), the merger would be price-lowering and procompetitive,90 and that the Commission was protecting the competitors of Honeywell and GE rather than protecting competition.91 87 GE/Honeywell, Commission Decision, supra note 83. 88 Id. ¶¶ 350–411. 89 Mergers and Acquisitions: Antitrust Division Chief Reacts to EU Decision to Prohibit GE/ Honeywell Deal, 81 Antitrust & Trade Reg. Rep. (BNA) 15 ( July 6, 2001). 90 This argument depends on both GE and Honeywell being dominant firms in their markets—a factual circumstance that the U.S. officials believed did not exist. See infra note 91. 91 See, e.g., William J. Kolasky, Conglomerate Mergers and Range Effects: It’s a Long Way from Chicago to Brussels, Address Before George Mason University Symposium (Nov. 9, 2001), available at http://www.usdoj.gov/atr/public/speeches/9536.htm.; Timothy J. Muris, Merger Enforcement in a World of Multiple Arbiters, Address Before Brookings Institution Roundtable on Trade and Investment Policy (Dec. 21, 2001), available at http:// www.ftc.gov/speeches/muris/brookings.pdf. For a different point of view, see Go¨tz Drauz, Unbundling GE/Honeywell: The Assessment of Conglomerate Mergers Under EC Competition Law, in International Antitrust Law & Policy, 2001 Fordham Corp. L. Inst. ch. 9 at 183 (Barry Hawk ed., 2002); Eleanor M. Fox, U.S. and European Merger Policy—Fault Lines 2002] What Is Harm to Competition? 401 Although the U.S. antitrust authorities championed the merger as efficient, GE had never argued to the European Commission that the merger was efficient; merely that, because it was only conglomerate, it could not harm competition. The Commission saw the merger as driven by GE’s expectations to use its leverage (as it had done with GE products) to make Honeywell dominant. Based on different inferences from the facts, the two agencies drew different conclusions. The U.S. Department of Justice inferred that prices would go down and stay down. The European Commission inferred both that use of leverage would shift share

to the merged firm and confer dominance, and that, after a period of low pricing of bundles, prices would rise. What was anticompetitive to the European Commission was procompetitive to the U.S. Department of Justice.92

Thereafter, in Tetra Laval BV v. Commission,93 the Court of First Instance annulled a Commission decision prohibiting a conglomerate merger. The CFI accepted the proposition that a merger may enable a firm to leverage its way into dominance.94 It identified uses of leverage that may constitute an abuse of dominance: tying, bundling, forced sales, and loyalty rebates. It observed, however, that effects of conglomerate mergers are normally “neutral, or even beneficial” for competition; therefore “the proof of anticompetitive conglomerate effects of such a merger calls for a precise examination, supported by confincing evidence, of the circumstances which allegedly produce those effects. . . .”95 The Commission had not offered such proof. First, it had failed to consider that the merged firm’s incentives to engage in any practice that constituted an abuse of dominance were reduced or eliminated due to the illegality of the practice, and that they would have been reduced or eliminated by the commitments that the merger parties had offered to the Commission.96

Second, since the merged firm could not be presumed to engage in illegal practices, it was necessary to consider whether the firm would have engaged in leveraging strategies that were not illegal, e.g., loyalty rebates or low prices that were objectively justified; but the Commission and Bridges: Mergers that Create Incentives for Exclusionary Practices, 10 Geo. Mason L. Rev. (forthcoming 2002). 92 General Electric is seeking annulment of the decision in the Court of First Instance. Questions of proof—as well as concept—are involved in the appeal. See infra text accompanying notes 112–115. 93 Supra note 71. 94 Id. ¶¶ 146, 151. 95 Id. ¶¶ 155, 156. 96 Id. ¶¶ 159–161, 218. [Vol. 70 Antitrust Law Journal 402 had offered no proof that this would be the case.97 Therefore the Commission had failed to prove its leveraging case (and had failed as well to prove its other allegations). D. Modernization and Reform in the European Union The above descriptions and analyses are based largely on judgments of the European Court of Justice and decisions of the European Commission. Some Court of Justice judgments on abuse of dominance are not of recent origin. Just as U.S. law and practice has changed, the law and practice of the European Community (EC) may change. Indeed several changes are already in progress. This section examines certain traditional baselines, and recent and expected changes. Until the entry of the United Kingdom and Ireland into the European Community in 1973, the law of the EC was derived from the civil law system, not the common law, system. Not only were the original six Member States civil law countries, but they included among them statist, protectionist, and prescriptive traditions. Article 81 itself embeds a regulatory mode of procedure by “catching” (making void in inception) all agreements that “distort” competition in a broad sense, and providing

for exemption of such agreements that satisfy specified conditions— essentially, agreements that enhance production or distribution while allowing consumers a fair share of benefits. Accordingly, significantly exclusionary, trade-restraining, and rivalry-limiting agreements (even intrabrand) were void unless justified. Facing an unmanageable workload of requests for individual exemptions, the Commission adopted a practice of legislating prescriptive measures known as “block exemptions,” which authorized particular kinds of agreements without the usual noti- fication and individual exemption as long as the agreement contained certain enumerated clauses (the white list) and did not contain other enumerated clauses (the black list). Within the last decade, two things have happened, partially in response to criticism that the system was too rigid and out of tune with contemporary economics, and partly in an effort to address a work overload. Problems of lesser or local importance were diverting the Commission from major, Community-wide competition problems, such as cartels. First, at the level of the Competition Directorate and the Commission, the Competition Directorate has spearheaded major reforms to simplify and liberalize the block exemptions. The new-generation block exemp- 97 Id. ¶¶ 219, 308–309. The court did not question whether strategies that were objectively justified could be anticompetitive. 2002] What Is Harm to Competition? 403 tions contain no mandated clauses and minimal prohibited clauses. Moreover, they are more sensitive to the possibility that the prohibitory regulations could impair efficiencies. For vertical agreements, a new block exemption provides a safe harbor when not more than 30 percent of the market is affected and no hard-core prohibitions are included;98

and no pre-exemption notification is required for individual exemptions. 99 In addition, a “modernisation” proposal, expected to be adopted by the end of 2002, empowers Member States and their antitrust authorities to grant individual exemptions from Article 81.100 By including the Member States in the EC enforcement network, current economic thinking, if not already adopted, is likely to make its way into the intellectual discussion that influences enforcement.101

Second, at the level of the courts, there have been changes in jurisprudence in selected areas. In Bronner v. Mediaprint,102 the Court of Justice circumscribed the essential facilities doctrine. The major newspaper in Austria, with the only pervasive distribution system in the country, was not required to give access to a small newspaper where the latter could survive by its own devices, although at higher cost and lower circulation. In NDC Health/IMS,103 the Court of First Instance vacated an interim order that gave a competitor access to IMS’s copyrighted zone design for collecting sales data useful to the pharmaceutical companies; and, pending decision on the merits, the CFI expressed skepticism regarding plaintiff’s theory of defendant’s duty to license its intellectual property. Perhaps most dramatically, in June and October 2002, the Court of First Instance overturned three Commission decisions prohibiting mergers. In two cases, the CFI found that the Commission had failed to accord 98 Commission Regulation No. 2790/1999 of 22 December 1999 on the Application of Article 81(3) of the Treaty to Categories of Vertical Agreements and Concerted Practices,

1999 O.J. (L 336) 21, reprinted in Selected Documents, supra note 80, Competition Doc. No. 4, p. 504. See Alexander Schaub, Vertical Restraints: Key Points and Issues Under the New EC Block Exemption Regulation, in International Antitrust Law&Policy, 2000 Fordham Corp. L. Inst. ch. 13 at 201 (Barry Hawk ed., 2001). 99 Council Regulation No. 1215/1999 of 10 June 1999 amending Regulation No. 19/ 65/EEC and Council Regulation No. 1216/1999 of 10 June 1999 amending Regulation No. 17, 1999 O.J. (L 148) 1, 5. See also Schaub, supra note 98. Amended Regulation 17 is printed in Selected Documents, supra note 80, Competition Doc. No. 1, p. 474. 100 See Alexander Schaub, Continued Focus on Reform: Recent Developments in EC Competition Policy , in International Antitrust Law & Policy ch. 2, 2001 Fordham Corp. L. Inst. 31 (Barry Hawk ed., 2002). 101 See id. for a description of the network. 102 Oscar Bronner GmbH & Co. KG v. Mediaprint Zeitungs und Zeitschriftenverlag GmbH & Co., Case C-7/97, [1998] E.C.R. I-7791. 103 IMS Health Inc. v. Commission, Case T-184/01 R, [2001] E.C.R. II-2349, and, further suspending proceedings for interim relief pending final judgment, order of CFI [2001] E.C.R. II-3193. [Vol. 70 Antitrust Law Journal 404 basic procedural rights to the parties. In all three, the CFI found that the Commission had made serious and pervasive errors of economic analysis and evidence; principally, the Commission had not proved what it said it had proved.104

Finally, in the aftermath of GE/Honeywell, Competition Commissioner Monti has stated that EC competition law is intended to protect consumers, not competitors; he has refuted the charge that EC competition law does otherwise.105 He has stated, also, that there is no efficiency “offense”—meaning that transactions are not illegal because they are efficient.106 As reported in the press, after ameeting with the Competition Commissioner, the staff of the Competition Directorate reversed course on an emerging recommendation to prohibit amerger between Carnival cruise group and P&O Princess; the evidence of collective dominance may not have withstood the scrutiny of the court.107

What do these developments suggest regarding the EC law on exclusionary practices that are neither justified nor price-raising? One could argue that the Community now accepts U.S. law and Chicago School economic analysis; but this would be wrong.108 The European analysis involves more microeconomics than it previously did. The Competition Directorate analysts are likely to be more rigorous about their proof. They are increasingly likely to credit efficiency aspects of conduct and transactions. But EC law has not become economics, and the economics incorporated into EC law is not solely Chicago School or even post- Chicago economics.109 New vertical guidelines for agreements involving more than 30 percent of the market are still concerned about foreclosure of market actors, as one of four independent negative effects of exclu- 104 Airtours v. Commission, Case T-342/99 (Ct. First Instance June 6, 2002); Schneider Electric SA v. Commission, Cases T-310/01 and T-7702 (Ct. First Instance Oct. 22, 2002); Tetra Laval BV v. Commission, Case T-5/02 (Ct. First Instance Oct. 25, 2002). The cases are available at http://curia.eu.int.jurisp. In Airtours and Schneider, the court found violations of the parties’ rights of defense. 105 See also Commissioner Mario Monti, The Future for Competition Policy in the European Union (Extracts), Address at Merchant Taylor’sHall, London ( July 9, 2001), available at http://europa.eu.int/comm/competition/speeches/index 2001.html. 106 See Commissioner Mario Monti, Review of the EC Merger Regulation—Roadmap for the Reform Project, Conference, British Chamber of Commerce, Brussels ( June 4, 2002), available at http://europa.eu.int/comm/competition/speeches/index 2002.html.

107 See Francisco Guerrera & Tobias Buck, Brussels U-Turn on Carnival Merger May Signal Policy Shift, Fin. Times, July 17, 2002, at 4. 108 See Statement of Commissioner Mario Monti, supra note 64; Hildebrand, supra note 62. 109 Post-Chicago economics, by relaxing classical assumptions, is more likely to find price and output effects than is Chicago School economics. See supra note 4. 2002] What Is Harm to Competition? 405 sionary agreements (higher prices being an aspect of the others).110 As one author recently commented, “If the supplier is dominant, exclusive dealing is generally prohibited by Article 82 where it is capable of affecting trade between EC Member States.”111 Article 82 still prohibits abuse of dominance, not just its creation and maintenance. I find no reason to believe that EC law will abandon its concern that dominant firms may use their power to appropriate advantages for themselves at the expense of competitors, nor to abandon its vision of harm to competition that regards open markets, access on the merits, and safeguarding of the market mechanism as mainstays of healthy competition. The first point of the divergence of U.S. and EC law regards rigor of proof of output limitation and price rise. Airtours augurs a period of greater rigor in European analysis in proving what the Commission purports to prove, including price rise. The second point is a question of concept. In abuse of dominance cases, so many of which have been vetted in the Court of Justice, it is clear that “harm to competition” is a wider concept than result-oriented output limitation. Use of dominant power to procure significant advantages not on competitive merits, thereby preempting competitors’ opportunities, is a harm to competition under Article 82. The European Court would have had no difficulty condemningMicrosoft’s purposeful exclusions of Netscape fromall efficient channels of browser distribution, without concern as to whether there are barriers to the browser market or whether browsers may be used in fixed proportions to operating systems, yielding only one monopoly profit for the package. The future of European merger law is less clear, since only collective dominance (a price-raising theory) has yet been vetted before the Court of First Instance and the Court of Justice.112

110 Guidelines on Vertical Restraints, ¶ 107(1). The guidelines may be found at 2000 O.J. (C 291) 1, and on the Competition Directorate’s Web site, www.europa.eu.int/comm/ competition/antitrust/legislation/entente3 en.html#iii 1. 111 S.O. Spinks, Exclusive Dealing, Discrimination, and Discounts Under EC Competition Law, 67 Antitrust L.J. 641, 650 (2000). A more nuanced statement would be: Exclusive dealing by a dominant firm is presumptively prohibited. The practice may be objectively justified; i.e. as necessary or important for more efficient distribution, where consumers get a fair share of the benefits. 112 See France v. Commission (Kali + Salz), Cases C-68/94, 30/95, [1998] E.C.R. I-1375. This was the first European Court case to overturn a Commission decision that a merger created collective dominance. The Court held that collective dominance is an admissible theory in a proper case, but it annulled the decision for failure of proof. [Vol. 70 Antitrust Law Journal 406 If the GE/Honeywell prohibition is defended only on the basis of a predicted price-rise,113 then it, too, will give the courts little basis to explore the ground of exclusionary but not necessarily exploitativemergers. In a price-rise scenario, the exclusion would be inextricably linked

to exploitation. If, on the other hand, the ruling is defended independently on the ground that GE would predictably use its leverage to shift significant market share to Honeywell, conferring on Honeywell a presumptively dominant market share, weakening incentives of competitors, 114 then the court will, for the first time, have the opportunity to rule on whether significantly exclusionary mergers creating dominance, unjustified by the merging parties,115 are anticompetitive under the Merger Regulation. IV. A DEVELOPMENT PERSPECTIVE The United States, the European Union, and the many nations that adopt their models, do not represent the entire world. What is “harm to competition” in the rest of the world? Some jurisdictions define “anticompetitive” yet more broadly than the European Union and the United States. Some define “anticompetitive” to embrace methods of competition that they perceive to be unfair. An unfair competition component of competition policy may be anathema to policy makers in mature market jurisdictions, especially Western jurisdictions, because such a conception can protect competitors fromcompetition itself and application of this point of view has (in the developed world) no payoff except to the protected competitors. In Indomaret,116 the Indonesian competition commission enjoined a large supermarket fromexpanding into venues of traditional small stores, to protect against destruction of traditional local communities. The Commission may have perceived that the social costs to the people as citizens of the local communities were greater than the gains from low prices and variety realized by the people in their role as consumers, especially in this time of transition, social unrest, and recent memory of riots.117

113 See, for support of the leveraging theory, Robert J. Reynolds & Janusz A. Ordover, Archimedean Leveraging and the GE/Honeywell Transaction, 70 Antitrust L.J. 171 (2002). 114 Under EC law, a market share of 50% or more is deemed to create dominance, unless respondents refute the inference. See AKZO Chemie BV v. Commission, Case C-62/ 86, [1991] E.C.R. I-3359. 115 The merging firms have the burden to prove that the merger would create efficiencies or serve other consumer interests. GE did not offer such proof. 116 P.T. Indomarco Prismatama, 03/KPPU-L-1/2000. 117 The Business Supervisory Commission (the KPPU) found that Indomaret “does not observe the principle of balance in accordance with the principle of economic democracy 2002] What Is Harm to Competition? 407 In an illuminating essay presented at the Japan Fair Trade Commission’s 50th Anniversary Competition Symposium in December 1997, Kyu- Uck Lee stated the case for a fairness component in competition law:118

Competition is the basic rule of the game in the economy. Nevertheless, if the outcome of competition is to be accepted by the society at large, the process of competition itself must not only be free but also conform to a social norm, explicit or implicit. In other words, it must also be fair. Otherwise, the freedom to compete loses its intrinsic value. Fair competition must go in tandem with free competition. These two concepts embody one and the same value. This may be the reason that competition laws of several countries such as Korea and Japan clearly specify ‘fair and free competition’ as their crown objective. . . . I believe that the abstract notion of fairness rests, inter alia, on

equitable opportunities, impartial application of rules and redemption of past undue losses. . . . Fairness, then, does not imply absolute libertarianism but instead takes the form of socially redefined freedoms. Viewed from this perspective, the polemic whether competition laws should aim only at enhancing economic efficiency rather than at promoting some social policy goals such as fairness may appear to be irrelevant. After all, efficiency is intrinsically not a value-free concept.119

Drawing upon his Korean experience, Mr. Lee continued: [I]n a developing economy where, incipiently, economic power is not fairly distributed, competition policy must play the dual role of raising the power, within reasonable bounds, of underprivileged economic agents to become viable participants in the process of competition on the one hand, and of establishing the rules of fair and free competition on the other. If these two objectives are not met, unfettered competition will simply help a handful of privileged big firms to monopolize domestic markets that are usually protected through import restrictions. This will then give rise to public dissatisfaction since the game itself has not been played in a socially acceptable, fair manner.120

in promoting healthy competition between the interests of business enactors and public interests,” and ordered it “to cease its expansion in traditional markets, in which it is directly facing small-scale retailers, in the context of realizing balance in the competition between large-scale, medium-scale and small-scale business enactors. . . .” Id., relief, ¶ 2. 118 Kyu-Uck Lee, A “Fairness” Interpretation of Competition Policy with Special Reference to Korea’s Laws, in The Symposium in Commemoration of the 50th Anniversary of the Founding of the Fair Trade Commission in Japan, Competition Policy for the 21st Century at 61 (KFTC 1997) (on file with author). Professor Lee was then president of the Korea Institute for Industrial Economics & Trade and Chairman of the Competition Advisory Board for the Korean Fair Trades. 119 Id. at 61–62. 120 Id. at 62. In Korea, according to Mr. Lee, a subcontracting act and a small business act complement the competition law by bolstering the bargaining power of small and medium-sized firms and protecting small firms from entry of large firms for a reasonable time while they gained sufficient strength. Id. at 63. [Vol. 70 Antitrust Law Journal 408 Mr. Lee concluded: “[F]air and free” competition should be one and the same concept. Without fairness, freedom alone may not achieve the desirable outcomes expected from competition, especially in developing economies where unfair elements can be exacerbated by competition. . . . Despite the practical importance of fairness in competition policy, however, it is all the more difficult to have a practical yet socially agreed upon concept of fairness due to diverse individual value judgments. Competition authorities must bear this in mind in implementing competition laws. Similarly, the notion of competition itself differs in countries with different social and cultural traditions and conditions. Therefore, it is essential that competition authorities in various countries be able to better understand each other’s stance and policy environment in searching for the global rules of the economic game.121

Since 1997, when Korea faced a financial crisis, Korea undertookmajor liberalizing economic reforms. It has reduced government intervention, reinforced its market system, lowered tariffs, and exposed its firms to domestic and international competition. Korea has apparently risen from the category of developing countries to the category of developed countries. Nonetheless, reflecting on the challenges facing developing countries, Nam-Kee Lee, Chairman of the Korea Fair Trade Commission, has

observed that “developing countries cannot avoid concerns about the competitiveness of domestic businesses. In this context, it would not be well advised to suggest that developing countries adopt the same level of competition policy as developed countries, when their markets are not as mature and businesses not as competitive.”122

The Indonesian and Korean perspectives need not alter one’s preferred definition of “anticompetitive.” We might continue to label rules that protect firms from competition itself “anticompetitive” (as I do) even if wemay appreciate their justice and even their long-term contribution to efficiency in countries that must develop competition. Mr. Lee’s comments remind us that definitions have cultural and normative content. What is harm to competition is not pure, scientific, and absolute. 121 Id. at 64–65. Other formulations by developing countries include legislation against restrictive business practices, such as outlined in the 1980 UNCTAD Code. The restricted business practices enumerated therein are either exploitative or exclusionary of firms without power. The code-like prohibition of categories of conduct, such as exclusive dealing and tying, even tempered as it is by a reasonableness defense as a nod to the developed countries, may be simpler and easier to administer in a world of scarce enforcement resources. See Model Law on Competition, UNCTAD, TD/RBP/CONF. 5/7 (UN 2000). To many developing nations, restrictive business practices, now called anticompetitive practices, are harms to competition. 122 Nam-Kee Lee, Korean Economic Development Policy Lessons—The Shift from Industrial to Competition Policy, Keynote Speech at the Intergovernmental Group of Experts on Competition Law and Policy (Fourth Session) of UNCTAD, July 3, 2002, available at http://ftc.go.kr/data/hwp/200207.doc. 2002] What Is Harm to Competition? 409 V. HARMS TO COMPETITION—A GRAPHIC SUMMARY We return, then, to the initial perspective defined largely by U.S. and EU perceptions: there are two principal views as to whether serious, unjustified exclusions by dominant firms may be “anticompetitive” even if they have no negative output effects. The categories of harm from exclusionary conduct may be depicted thus. Table A Harm to Competition Harm Only to Competitors 1 2 3 Limitation of output Blocking of competition Foreclosure of firms on the merits; no without power by apparent procompetitive efficient competition; effects of the conduct thus, condemnation of the conduct protects competitors Mostly cartels and For U.S., includes many Indomaret government-privileged truncated-rule and monopolies; post- burden-shifting cases Brown Shoe Chicago School analysis where there is no reason (vertical aspects) widens the range.123 to believe output will be limited; e.g. Indiana Von’s Grocery Federation of Dentists ;124

Microsoft. For EU, includes many cases of abuse of dominance; e.g. Hoffmann-La Roche, Tetra Pak. From the point of view of welfare economists trained to predict

whether particular conduct or transactions will increase or decrease 123 Few cases of price predation, including predation by fidelity rebates, could confidently be placed within column 1. The conditions under which price predation will succeed in increasing market power and limiting output are demanding, and consumers are likely to gain from failed predation, at least if the target remains a viable player or the market is realistically contestable. In theory, an alleged price predation case could come within any of the three columns, depending on the facts. But because costs of error in condemning low prices are high, U.S. courts generally perceive the problem of price predation along lines of Table B, infra. They adopt a similar perspective for predation by product design. See, regarding Microsoft, text supra following note 59. They are less concerned, however, by unjustified foreclosing practices (which by definition do not have intrinsic pro-consumer qualities); thus, the not insignificant number of U.S. exclusionary cases not involving price or product design predation that fall within column 2. 124 FTC v. Indiana Fed’n of Dentists, 476 U.S. 447 (1986) (dentists’ combination to withhold x-rays from insurers held illegal). [Vol. 70 Antitrust Law Journal 410 aggregate consumer or total welfare, there is no column 2. Either conduct harms consumers or its prohibition harms consumers. From the point of view of policymakers who preferminimal antitrust intervention, column2 is so small and the danger of slippage into column 3 is so great that the better part of wisdom is to ignore column 2 and to label as “protection of competitors” enforcement against any conduct that does not qualify for column 1.125 The categories of exclusionary conduct may then be depicted thus. Table B Harm to Competition Harm to Competitors from Efficient (or (Harm to Consumers) Even “Unfair”) Competition 1 2 Harm to competition means no more Enforcement protects competitors and than: this transaction or conduct is harms consumers. inefficient because it limits output and raises prices. Enforcement is justified only in case of negative output effects. As developed in this article, jurisdictions do not agree on which is the wiser perspective, Table A or Table B. The contemporary case law of the United States largely adopts the perspective of Table B, at least in words if not in actions, ostensibly allowing enforcement only against conduct within column 1 (on either table). Even within jurisdictions, however, there is healthy disagreement. There may be “constructive opacity” within jurisdictions that proclaim only to protect consumers from bad outcomes (that is, there may be non-transparency as to whether 125 Thus, Robert Bork, when a jurist, said in Rothery Storage & Van Co. v. Atlas Van Lines, Inc., 792 F.2d 210, 221 (D.C. Cir.), cert. denied, 479 U.S. 1033 (1987): “If it is clear that Atlas and its agents by eliminating competition among themselves are not attempting to restrict industry output, then their agreement must be designed to make the conduct of their business more effective. No third possibility suggests itself.” Similarly, Professor Bork said in The Antitrust Paradox: “Improper exclusion” is always deliberately predatory and inefficient, which is rare, or the exclusion is the product of superior efficiency. “There is no ‘intermediate case’ of exclusion . . . .” Bork, supra note 22, at 160; see also supra note 30. While denial of the middle column of Table A may be thought to prevent the error of protecting inefficiencies, this strategy also has costs of error. Combined with a general policy of deregulation based on trust in business, not government, extreme or prophylactic non-interventionism could tend to insulate inefficiencies of, and even deceptions by, large merged firms and enhance their power to abuse competitors, consumers, stockholders, and employees. See Kurt Eichenwald & Simon Romero, The Latest Corporate Scandal Is

Sudden, Vast and Simple, N.Y. Times, June 27, 2002, at A 1, detailing the widespread frauds and in some cases ensuing failures of Enron, Tyco, Adelphia, Dynergy, Global Crossing, and WorldCom. 2002] What Is Harm to Competition? 411 the case law turns on proof or principle) that operates as a safety value to preserve a modicum of flexibility to catch exclusionary practices that distort the market mechanism.126

VI. CONCLUSION Nations have the right to adopt their own lexicons and to choose their own law. Nations’ choices are an internal matter unless and until application of a rule has negative effects outside of the jurisdiction or is so opaque or discriminatory as to harm foreign actors within the jurisdiction. What are the implications of the conclusions of this article for global antitrust? Can we ever effect a reasonably seamless world competition system if we disagree on that most basic of questions, what is harm to competition? Happily, yes.127 Healthy diversity tends to sharpen the dialogue and facilitate adjustment and readjustment to whatever the context demands—just as it has done and continues to do within both the United States and the European Union. 126 See text accompanying notes 23–27. One may question whether opacity is ever constructive and simply observe that it operates as a safety valve. 127 Nations might usefully anchor their most stable agreements, as may be done in elucidating the competition principles suggested in the Doha Declaration. Declaration at WTO Ministerial Meeting at Doha, Qatar (Nov. 14, 2001), ¶¶ 23–25 (anti-cartel; protransparency, due process, and national treatment). Moreover, nations might agree to rules of permissible and impermissible extraterritoriality, and to paths for dispute settlement in the event of systems clashes.