Efficiencies and Antitrust: A Story of Ongoing Evolution

Prepared Remarks before ABA Section of Antitrust Law, 2002 Fall Forum

Washington, D.C.

Date:
By: 
Thomas B. Leary, Former Commissioner

One of my recent articles(2) advanced the proposition that, contrary to popular belief, merger policy in the last 25 years has not shifted back and forth with changes of administrations. The evidence suggests instead that there was a revolutionary change, which began in the late 1970s, followed by a period of gradual evolution continuing to the present day.

This article will further develop that theme. In one sense, the article is narrower than its predecessor because the focus will be on a single issue, namely, the treatment of "efficiencies." In another sense, it is broader because it will pay attention to developments in non-merger law to a greater extent than the earlier article did. It will also range beyond the historical record and include an extensive discussion of present issues in the antitrust analysis of efficiencies.

Both the history and the policy issues are complex, and do not lend themselves to easy generalizations. It will become apparent, however, that, during the last quarter century, an increased recognition of the importance of efficiencies has co-existed with selective reluctance to recognize those efficiencies that are hard to quantify.

I. The Early History

Before the dramatic shift in the last 25 years, antitrust either consciously ignored the existence of efficiencies or was overtly hostile to them. Efficiencies of scale were easy to recognize but subordinated to other concerns; more subtle efficiencies resulting from transaction cost savings or improved incentives were not recognized at all.

A. Hostility or Conscious Disregard

Notwithstanding an occasional acknowledgment of the fact that economic efficiency is one of the main goals of antitrust,(3) efficiencies tended to be ignored or viewed as an aggravating rather than a mitigating factor. A good example of the former is provided by Judge Learned Hand's opinion in the Aluminum monopolization case,(4) perhaps the most influential antitrust opinion in the middle years of the last century. The opinion states that the Sherman Act is "based upon the belief that great industrial consolidations are inherently undesirable, regardless of their economic results"(5) and, referring to more recent statutes as well, adds:

"Throughout the history of these statutes it has been consistently assumed that one of their purposes was to perpetuate and preserve, for its own sake and in spite of possible cost, an organization of industry in small units which can effectively compete with each other."(6)

In the later Topco case,(7) involving horizontal territorial restraints, the Supreme Court expressed a different rationale for its refusal to consider efficiencies but the effect was the same:

"The fact is that courts are of limited utility in examining difficult economic problems. Our inability to weigh in any meaningful sense, destruction of competition in one sector of the economy against promotion of competition in another sector is one important reason we have formulated per se rules."(8)

Perhaps the most familiar expression of disregard for possible efficiencies is from Brown Shoe,(9) where the court cited the legislative history of the amended merger law(10) and said:

"[W]e cannot fail to recognize Congress' desire to promote competition through the protection of viable, small, locally owned businesses. Congress appreciated that occasional higher costs and prices might result from the maintenance of fragmented industries and markets. It resolved these competing considerations in favor of decentralization. We must give effect to that decision."(11)

This conscious disregard of possible efficiencies was transmuted into conscious hostility by the Federal Trade Commission. The most conspicuous example is the decision in Foremost Dairies.(12) Foremost began as a modest local operation and expanded nationally, primarily by geographic extension acquisitions. The Commission found that "[t]he resultant disparity and size and type of operations permits the large conglomerate to strike down its smaller rivals with relatively little effort or loss in overall profits," and then went on to conclude that a violation of Section 7 could be established by proof that "the acquiring firm possesses significant power in some markets or that its overall organization gives it a decisive advantage in efficiency over small rivals."(13)

Perhaps even more notable than this remarkable statement is the fact that counsel for Foremost did not see fit to attack the idea head on; the main thrust of the defense was that multiple acquisitions would have only a modest impact on efficiency.(14) In hindsight, this seems incredible. At the time, however, it was a realistic assessment of the Commission's likely views on merger law. A retrospective study done years later found that the Commission itself or its administrative law judges had consistently either cited efficiencies as justification for finding that a merger was illegal or the absence of efficiencies as support for finding that a merger was benign.(15) Even in those cases where efficiencies (or lack of them) were not decisive, the complaints almost uniformly asserted that efficiency creation should be a factor that weighed against legality. The most common claim was that the merger would yield "decisive competitive advantages."

B. Inability to Recognize Efficiencies

If the courts or the Commission were not affirmatively hostile to some kinds of efficiency, they seemed unable to recognize the existence of others. An interesting illustration of this distinction is provided by Philadelphia National Bank.(16) In this merger case, the defendants did not argue that a larger bank would yield economies of scale or other cost savings.(17) Presumably, it would have been futile to do so, in light of the Brown Shoe case decided in the previous year. Rather, they claimed that a larger bank would get more business, which would be better for the local economy - - an argument that was likely vague by design. The Supreme Court gave short shrift to it anyhow:

"[A] merger the effect of which 'may be substantially to lessen competition' is not saved because, on some ultimate reckoning of social or economic debits and credits, it may be deemed beneficial. A value choice of such magnitude is beyond the ordinary limits of judicial competence, and in any event has been made for us already, by Congress . . . ."(18)

Forty years ago, economies of scale could be understood as efficiencies, for better or worse, but other kinds of efficiency were hard to recognize. Today, the ability to attract customers, perhaps through the added convenience of multiple locations, could well be asserted as an efficiency in its own right. However, as outlined in a later discussion,(19) non-quantifiable efficiencies are still hard to deal with in merger cases.

This inability to recognize more subtle efficiencies was not confined to merger cases. Tying arrangements, for example, were condemned per se because they "serve hardly any purpose beyond the suppression of competition."(20) Exclusive dealing arrangements were evaluated under a test that focused directly on the magnitude of the commerce "foreclosed,"(21) irrespective of underlying rationale for the transactions. An atypical signal in White Motor(22) that the Supreme Court might be open to consideration of the business rationale for vertical territorial restrictions was limited to the vanishing point only four years later in Schwinn.(23)

It is, perhaps, not surprising that the courts (and the antitrust bar) were so oblivious to the potential for efficiencies other than those directly related to scale because the economic literature was sparse. Perhaps the most significant early contribution was the publication of Ronald Coase's article on The Nature of the Firm.(24) Coase examined the reasons why firms choose to produce goods or services themselves or, alternatively, purchase them from outside suppliers. He stated that the key discriminating factors are so-called "transaction costs," and this basic insight ultimately led to a vastly expanded understanding of the nature of efficiencies. But, it took time.

In 1968, Oliver Williamson published his influential article, Economics as an Antitrust Defense: The Welfare Tradeoffs,(25) which elaborated on the connection between Coase's ideas about efficiency and the resolution of antitrust cases. He wrote:

"[I]f neither the courts nor the enforcement agencies are sensitive to these [efficiency] considerations, the system fails to meet a basic test of economic rationality. And without this the whole enforcement system lacks defensible standards and becomes suspect."(26)

After Coase and Williamson, it was no longer appropriate to assume that a merger is simply designed to eliminate a competitor or that a vertical restraint is simply designed to "foreclose" a rival from a customer or a source of supply. It is apparent that the decision to merge rather than, say, to enter into a supply contract may be prompted by the need to reduce significant transaction costs associated with a supply contract. Similarly, the decision to sell through independent dealers may be prompted by higher costs of internal forward integration, but the vertical restrictions in the dealer contracts may more closely align dealer interests with those of the supplier. But, for a long time, these economic theorists were talking largely to themselves. Both mainstream economists(27) and enforcement officials(28) continued to be suspicious of unfamiliar things that they did not understand.

C. The Convergence of Law and Economics

The milestones in the translation of the new economic learning into principles of antitrust law are now familiar. Some were highlighted in my earlier article on the evolution of merger policy,(29) and will not be repeated here. What is remarkable in retrospect is the speed with which this new learning was accepted in the antitrust legal community. Perhaps the infamous trio of Utah Pie, Schwinn andAlbrecht(30) represent the highwater expressions old-fashioned formalistic antitrust. These decisions are separated by only a decade from Sylvania, Brunswick and Fortner II cases(31) that arguably mark the transition to the economics-oriented antitrust that we observe today. There is now a general consensus that policy should, in William Baxter's words, be "based on whatever it is we know at any particular moment about the economics of industrial organization."(32)

This convergence of law and economics does not mean that the battles are now over and nothing more needs to be learned. On the contrary, as I have outlined elsewhere,(33) and will again in the last part of this article, there are still a lot of things we do not know and questions to be resolved. One of the most difficult is the nagging question of how to deal with efficiencies that may be most important but cannot be readily quantified as cost reductions.

The distinction between scale efficiencies that can be quantified and so-called "managerial" efficiencies that are equally real but more tenuous still persists. The distinction is most evident in the different standards for weighing possible efficiencies in merger cases and non-merger cases. We will illustrate the persistence of these distinctions in the section that follows, and then go on to evaluate the reasons for them in the final policy-oriented discussion.

II. Recent History: Consideration of Efficiencies During the Last Twenty Five Years.

A. Merger Cases and the Emphasis on Cost Reductions

The Supreme Court has referred to the role of efficiencies in merger cases only once since antitrust analysis was transformed about twenty five years ago. In the Cargill case,(34) a competitor brought an action to enjoin the merger of two leading beef packers. Plaintiff argued that the merger was anticompetitive because it would create "multiplant efficiencies" that would enable the merged entity to reduce prices and increase its market share. The Supreme Court emphasized that reducing prices in order to increase business is often "the very essence of competition" and that "mistaken inferences . . . are especially costly because they chill the very conduct the antitrust laws are designed to protect,"(35) and stated:

"To hold that the antitrust laws protect competitors from the loss of profits due to such price competition would, in effect, render illegal any decision by a firm to cut prices in order to increase market share. The antitrust laws require no such perverse result, for '[i]t is in the interest of competition to permit dominant firms to engage in vigorous price competition, including price competition.'"(36)

Since the plaintiff in Cargill was asserting that post-merger prices would be too low,(37) the case did not squarely address the question whether prospective efficiencies could be used as a "defense" to overcome an inference that increased concentration would lead to higher prices. The question has been considered in four Court of Appeals decisions, however, and the court in each case has demonstrated a willingness to consider an efficiencies defense.(38)

All four cases involved challenges by the Federal Trade Commission. In two cases, the courts permitted proposed hospital mergers to proceed based on claims of enhanced efficiency(39) and, in a third hospital merger case, the court found that proof of efficiencies was relevant but not sufficient to overcome the FTC's prima facie case.(40) The fourth case involved the proposed merger of two of the three baby food manufacturers in the United States. The court noted that "the trend among lower courts is to recognize the [efficiency] defense,"(41) but held that the parties had failed to demonstrate "extraordinary efficiencies" sufficient to rebut the strong inference of anticompetitive effects in a 3-2 merger.

The pattern of decisions in the District Courts is similar. Courts are willing to entertain efficiency claims but subject them to close scrutiny. In three recent cases,(42) Staples, Cardinal Health and Swedish Match, the District Court for the District of Columbia applied the analytical framework of the DOJ/FTC Merger Guidelines, and found that the efficiency claims were insufficient. They were found to be overstated, or not merger specific, or not sufficiently likely to be passed on to customers.

B. Agency Views of Merger Efficiencies

Since very few merger cases are actually litigated and since (with the exception of hospital mergers) the prosecutors tend to prevail in court, the internal treatment of merger efficiencies by the agencies is of critical importance. Evidence here is harder to come by.

One source is the authoritative statements of the agencies themselves in their merger guidelines. In the last 20 years, the agencies have issued guidelines on four separate occasions: 1982, 1984, 1992 (horizontal mergers only) and 1997 (discussion of efficiencies only).(43) Although the 1982 Guidelines were not particularly receptive to efficiency arguments in merger cases, each successive iteration has been more hospitable.

The DOJ 1982 Merger Guidelines did explicitly recognize efficiencies as a defense. The Guidelines added, however, that efficiencies would be considered only in "extraordinary cases," and that the defense required proof of "substantial cost savings" by "clear and convincing evidence."(44) The FTC's own separate contemporaneous Statement Concerning Horizontal Mergers was even more chilly; it specifically rejected efficiencies as a legally cognizable defense and stated that they would be considered only in the exercise of "prosecutorial discretion at the pre-complaint stage."(45)

The DOJ's 1984 Guidelines were consciously more hospitable to efficiency claims. The introductory statement to these Guidelines stated that the discussion of efficiencies in the predecessor version "had a restrictive, somewhat misleading tone." The Guidelines went on to say that "the primary benefit of mergers to the economy is their efficiency-enhancing potential." They stated that the Department "never ignores efficiency claims" and accords them "appropriate weight." Nevertheless, they retained the requirement that efficiencies be demonstrated by "clear and convincing evidence." Moreover, the 1984 Guidelines warned that the Department would reject efficiencies that could "reasonably be achieved" through other means, and pointed out that the more significant the competitive risks, the greater the magnitude of efficiencies needed.(46)

The discussion of efficiencies in the 1992 Guidelines, issued jointly by the DOJ and the FTC, was unchanged with one significant exception - - the omission of the earlier requirement that efficiencies be proven by clear and convincing evidence.(47) The 1997 Revisions retained the introductory language from the 1984 and the 1992 Guidelines declaring that "the primary benefit of mergers to the economy is their potential to generate . . . efficiencies." The revision elaborated on the mechanism by which efficiencies could increase the competitiveness of firms, and it expanded the list of efficiency benefits to include "improved quality, enhanced service, or new products" in addition to lower prices.(48)

This overt willingness to entertain efficiency claims beyond those that can be quantified as cost reductions is potentially of immense importance. It was actually foreshadowed eighteen years ago in the GM-Toyota matter,(49) where the Commission - - by a 3-2 vote - - gave qualified approval to a production joint venture between the first and the third largest automobile companies in the world. (This was a joint venture, not a merger, but the antitrust analyses are in many ways similar.) The rationale for the transaction was not the potential for scale economies but rather a mutual education process in which one partner (Toyota) would gain experience producing vehicles with a U.S. labor force and the other partner (GM) would become familiar with Japanese lean production techniques. The experiment was apparently successful for both partners, as Toyota subsequently built its own assembly plant in Georgetown, Kentucky and General Motors applied the lessons learned in the venture to the design of its own Saturn division.(50)

The Commission's decision on the GM-Toyota joint venture seemed daring at the time, and was met with some vigorous opposition. For example, one Commissioner dissented with unusually harsh rhetoric:

"In this decision, the Commission has swept another set of generally recognized antitrust principles into the dustbin, using again the incorporeal economic rhetoric that now dominates Commission decision-making. In this case, the decision results in the blessing of a business proposal that is both breathtaking in its audacity and mind-numbing in its implications for future joint ventures leading U.S. firms and major foreign competitors that seek to lend a helping hand."(51)

GM-Toyota would undoubtedly be less controversial today.(52) Congress has, in the meantime, passed statutes specifically designed to encourage the formation of efficiency-enhancing research and/or production joint ventures.(53) In the absence of anticompetitive ancillary restraints, rule-of-reason treatment is assured. Moreover, ventures that are notified to the antitrust agencies are subject to single-damage lawsuits only in the event of subsequent antitrust litigation. To my knowledge, only one antitrust lawsuit has ever been brought against a notified joint venture, and the plaintiff was unsuccessful.(54)

Since GM-Toyota, the antitrust agencies have not only issued increasingly pro-efficiency merger guidelines but they also have expressed support for expanded recognition of efficiencies. Robert Pitofsky, who chaired the Commission from 1995 into 2001, mildly criticized his Republican predecessors in the 1980s for failure to adequately take account of efficiency claims. In 1992, when still a professor at Georgetown University Law Center, he wrote:

"Experience in capital markets in the 1980s demonstrates the U.S. antitrust policy concerning mergers was (and continues to be) on the wrong track . . . . [F]ew would argue that the failure of the United States enforcement agencies and courts to take in account efficiency . . . considerations in merger analyses was the principal cause of American firm's difficulties in international trade. Nevertheless, in some market situations, consideration of such [a] factor[] could result in permitting otherwise illegal mergers and could make a significant difference in the ability of firms to compete in international trade."(55)

When Pitofsky became FTC Chairman in 1995, he inaugurated his administration with comprehensive hearings on antitrust policy, including the appropriate role of efficiencies. The hearings culminated in a FTC staff report that endorsed further integration of efficiency concerns into competitive effects analysis and argued that efficiencies should be treated as "a rebuttal [to a share-based prima faciecase], not an affirmative defense."(56)

This change would affect the burden of proof.

Note, however, that the emphasis is still on production efficiencies that promise cost savings to be passed on to consumers. The current Chairman of the Commission, Timothy Muris, has in turn been critical of his Democratic predecessors in the 1990s for taking an unduly narrow view about the kinds of efficiencies that count.(57) In an article published while still a professor at George Mason University School of Law, he included among the neglected efficiencies some that can be quantified, like lower costs of capital or economies in product promotion. In addition, however, he states that "there have been well-known instances in which the managerial and administrative expertise of an acquiring company has had a substantial positive impact on an acquired firm" - - and refers to the GM-Toyota joint venture as an example of this general proposition.(58)

In the policy discussion that follows in Part III, we will discuss some reasons for the comparative neglect in merger cases of efficiencies not directly related to production. It is noteworthy, however, that even more intangible efficiencies than those cited by Muris are routinely recognized in non-merger cases.

C. Consideration of Efficiencies in Non-merger Cases

Landmark vertical cases like Sylvania,(59) Sharp,(60) and Kahn(61) are normally not considered to be "efficiency" cases but their rationale can be traced back directly to Ronald Coase's pioneering work on transaction costs in the 1930s.(62) In fact, almost all vertical cases today seem to turn on the perceived efficiency of a particular restraint.(63)

The Sylvania decision, which overruled Schwinn and applied the rule of reason to a vertical non-price restriction is best remembered for its recognition that the economic welfare of consumers is the pre-eminent goal of antitrust and its overt acceptance of new economic theories that provided a pro-competitive explanation for behavior previously regarded as suspect. However, when the Court accepted the argument that increased inter-modal competition was worth the sacrifice of some intra-modal competition, it was implicitly weighing relative efficiencies. These are not scale efficiencies but rather efficiencies related to methods of distributing products. And, perhaps most remarkable was the Court's willingness to recognize the superiority of Sylvania's distribution system on the basis of improved sales performance rather than detailed cost analyses.

The Sharp decision, which held that it was not per se unlawful to terminate one customer after it was demanded by another customer, also depends on an implicit recognition that the efficiencies of a distribution system can be adversely affected if discount sellers "free ride" on the promotional activities undertaken by full-service dealers, who thereby necessarily incur higher costs. The Kahn decision, which held that maximum resale price restrictions were not per se illegal, depends on an implicit recognition that these arrangements may be a more efficient way to protect consumers from price-gouging dealers than the introduction of dealer competition.

Efficiencies can even be important when considering the legality of tying arrangements, which are still nominally subject to the per serule, under Jefferson Parish.(64) An essential element of the offense is that two separate products or services have been tied together, and the answer can turn on whether it is or is not efficient for a seller to offer them separately.(65) Another element of the offense is the existence of a "tie," and this may involve a subtle distinction between coercion and persuasion - - which is another way of asking whether the transaction is efficient from the buyer's point of view.(66) And, finally, some courts have considered other business justifications for the arrangements.(67)

It is one thing, of course, for the Supreme Court to recognize the possible or probable existence of various efficiencies and to mandate a further factual inquiry, it is another thing actually to decide ultimate liability. But, we are confident a survey would show that defendants prevail in the lower courts without a need actually to quantify the efficiencies of their challenged distribution systems. It is enough that they appear plausible.(68)

One area where the Supreme Court stubbornly resists consideration of efficiency claims is the continued per se condemnation of minimum resale price maintenance, on the basis of a precedent from the turn of the last century.(69) This is probably just a historic anomaly - - the exception that tests the rule - - but the Court's reluctance may also be based in part on some lingering scholarship that argues the same efficiencies can be achieved in a less restrictive way.(70) The role of "less restrictive alternatives" in non-merger cases corresponds to the role of "merger specificity" in merger cases.(71)

Implicit recognition of efficiencies is not confined to vertical restraint cases that involve distribution strategies. Innovation is an important efficiency as well, and even harder to quantify. Monopolization cases like Berkey Photo(72) demonstrate an extreme reluctance to interfere with innovative efficiency, even if the innovator already has monopoly power and the competitive impact of the innovation is serious.

D. Conclusory Comments on the Historical Survey

The discussion up to this point demonstrates that efficiencies in the abstract are universally recognized or pro-consumer and consistent with the goals of antitrust. It would be unthinkable for a court or a commentator to conclude as the Court did in Brown Shoe forty years ago, that antitrust's primary objective is "the protection of viable small, locally owned businesses."(73) However, it is also evident that efficiencies of different kinds are treated in different ways in different kinds of cases. The next section will attempt to analyze these distinctions in more detail and set out some challenges for the future.

III. Outstanding Policy Issues

A. Merger Matters

The preceding discussion suggests that decisionmakers are willing to take a broader view of efficiencies in non-merger cases than they are in merger cases. Accordingly, this section will focus on policy issues in the merger field. There are, however, some interesting non-merger issues that will be discussed at the end.

1. The Distinction Between Formal Decisions and Informal Agency Practice

As outlined above, judicial precedent and agency arguments in litigated cases suggest that an efficiencies "defense" in a merger case requires proof of quantifiable efficiencies that will be passed on to consumers to a degree that will be sufficient to outweigh some assumed upward price effect arising from increased concentration. In short, will prices be increased or will they not? In addition, the parties have the burden of showing that the claimed efficiencies are "merger specific" - - i.e., that they cannot be achieved in a less restrictive way like a production joint venture or a requirements contract.

Experienced counsel know that in actual practice the vast majority of efficiency claims are resolved internally in the agencies in a much less formal way. For mergers that do not involve extreme concentration levels, efficiencies are not analyzed as a "defense," and rigorously quantified, but rather included in the description of the overall business rationale for the transition. Because there is no need to strike a mathematical balance, the presentation can include not only scale effects on production but also things like improved potential for innovation and managerial efficiencies.

Two recent examples may be identified because in each case the Commission or individual commissioners issued public statements explaining why the transitions were not challenged. In AmeriSource/Bergen, the Commission statement referred not only to potential scale economies but also to the likelihood that improvements will be made "more rapidly than either [party] could do individually."(74)Individual statements on Synopsys/Avant! refer to the innovative potential of vertical integration.(75)

These examples are not atypical. Although the 1992 Guidelines refer to a presumption of illegality for mergers in the "highly concentrated" range (HHI over 1800), the reality is that they are often approved. A recent study of enforcement patterns in the Federal Trade Commission found that 81 of 109 non-challenged horizontal merger cases from 1983-1996 had HHI levels above 1800.(76) It is evident that the parties were able to provide persuasive arguments that the transactions would be pro-competitive or benign. This informal, internal practice can itself be efficient, but it is not entirely satisfactory for obvious reasons - - so, it is worthwhile continually to work on making the process more transparent and predictable. The balance of this Part III will discuss the various challenges involved in that effort.

2. Problems of Quantifying and Comparing Predictions

In merger cases (or any other case that requires predictions of the future),(77) rigorous proof necessarily involves the use of various forecasting techniques that are unfamiliar to most lawyers or judges. This factor alone produces some skepticism or resistance to the evidence. A good example is the opinion in the Staples case,(78) where voluminous econometric evidence was presented to, but largely ignored in the opinion of a sophisticated District Judge, who relied instead on internal documents and his own visual impressions. Triers of fact or those with discretion to prosecute seem reluctant to deal with complex statistics if there are other grounds for decision.(79) The statistics will serve to reinforce (or raise questions about) tentative conclusions based on other criteria, but I am not aware of a single external or internal decision where econometric evidence or engineering and accounting studies, standing alone, were decisive.

This reluctance to depend on econometric evidence is not solely based on fear of the unfamiliar; there really are serious problems associated with any effort to demonstrate rigorously that potential efficiencies outweigh potential anticompetitive effects. Estimates on both sides of the balance are inherently uncertain, even in apparently simple cases.

Assume a hypothetical horizontal merger,(80) where it is possible to identify competitors in a clearly-defined, homogenous geographic and product market, that there are no "fringe" players of competitive significance, that the product is relatively fungible and that comprehensive data are available on volumes and prices. (These are very generous assumptions that avoid complexities present in most cases.) In order to estimate the effects of combining the two parties into one, it is necessary to feed the sales data into an economic model.

Competing models are available,(81) but it is necessary to make further simplifying assumptions about matters like elasticities of demand and the ways in which companies will react to the strategies of their competitors. It is further necessary to make assumptions about the extent to which the merger itself will or will not change the relevant parameters.(82) Given all these assumptions, there are invariably differences in expert estimates about the price effects of increased concentration, standing alone.

The calculation of offsetting efficiencies is similarly imprecise. Neither agency staff nor courts have the company-specific knowledge even to calculate scale efficiencies from the ground up - - they have to rely on internal company estimates, to the extent that the estimates are facially plausible and consistent. Yet, anyone who has worked in the corporate world or studied it closely knows that even the most careful and objective estimates are likely to be wrong.(83) There is an extensive and growing business consulting literature that demonstrates a substantial number of merger transactions (some say more than half) do not achieve the shareholder benefits that were predicted.(84) A particularly interesting finding is that predicted "synergy" gains may occur only 50% of the time, but that there were other gains that were not predicted.(85) (These findings will be discussed in a different context later on.)

Another problem with efficiencies evidence is that there really is no adversarial process to test it. Possible adverse effects from the disappearance of a competitor may be confirmed by statements of customers or even of other competitors,(86) but it is not easy for any outsider to second-guess a company's own internal efficiency estimates. Moreover, there does not seem to be any rigorous learning as yet on the objective characteristics of companies that are likely or unlikely to achieve predicted merger efficiencies.(87)

In other words, both sides of the concentration/efficiency balance involve assumptions that are difficult to test and, in this situation, calculations that purport to predict future price movements with precision are unlikely to be conclusive.(88) I once heard a speaker(89)say that the situation is comparable to a football game. Imagine a runner moving up the field and then hit by a defending player. Thereafter, there may be a lot of forward and backward movement, as players on both sides push in opposing directions, before the runner and ball disappear in a pile. An official will peel off the opposing behemoths, and move the ball to a place on the field where he estimates it was when forward motion stopped moments before. If the gain appears to be close to a first down, the chains are then brought in from the sidelines and the outcome may be decided by inches!

It is obvious, of course, that the outcome-determinative event is not the precise measurement itself but the judgement-call that determined where the ball was placed. And, so it is with econometric calculations of price effects; the outcome-determinative events are not the calculations but the underlying assumptions on both sides of the balance.

This is not to say that econometric measurements are or should be ignored altogether. They can help to inform the debate but, for a long time to come, decisionmakers are likely to look for other confirming evidence as well.

3. The "Pass-On' Issue

The agencies in their guidelines, in individual speeches, and in their briefs assert that efficiencies count only to be the extent they are likely to be passed on to consumers.(90) This requirement is obviously intended to stake out a position in the theoretical debate about the fundamental objectives of antitrust. If the goal is to maximize "total welfare" or "overall welfare," it does not matter whether consumers or producers benefit from the efficiency improvements. If the goal is to maximize "consumer" welfare alone, the distribution of the gains is important and pass-on issues become important.(91) This issue emerged as more than a matter of theoretical debate recently in Canada - - the Competition Tribunal approved a merger that created a near monopoly by applying an "overall welfare" standard, thus sanctioning a likely wealth transfer from customers to the seller, only to be reversed on appeal.(92)

I do not believe that this is a fruitful policy debate for the simple reason that no endorsement of an "overall" welfare standard is politically viable in this country; the assumption that sellers are already much richer than buyers is just too deeply entrenched, even though it obviously is not always true. At the same time, however, I personally do not favor any separate requirement that pass-on of efficiency savings be shown.(93)

These are a number of reasons for skepticism about a pass-on proviso. It would assume that efficiencies can be neatly segregated into "variable" cost savings that presumably will be passed on and "fixed" cost savings that presumably will not. I am not confident that companies generally make these distinctions when they price in the real world. Moreover, these fixed-cost/variable-cost allocations superimposed on the already equivocal calculations discussed in the subsection immediately above simply add a further layer of imprecision to an already impressive exercise.(94) Finally, and perhaps most important, the antitrust agencies are never going to approve mergers that create monopoly power anyway and, in a competitive environment, efficiency savings are likely eventually to yield consumer benefits of some kind - - if not reductions in price, perhaps increased innovation and quality improvements.

Put another way, I do not believe we should be overly fixated on immediate changes in the dimensions of "rectangles" and "triangles;"(95) efficiency effects are much more subtle and longer lasting. In the next subsection, we will consider some issues surrounding these more subtle effects.

4. Significant Efficiencies that Tend to be Ignored

President Carter once appointed a prestigious Commission to study major revisions of the antitrust laws.(96) The Commission included Senators and Representatives, as well as prominent scholars and practitioners. Among the proposals considered was a so-called "no-fault" monopolization statute, under which a company that had maintained "persistent monopoly power" for an extended period of time could be dismembered, unless it could prove that its longstanding dominance was traceable to efficiencies.(97)Persistent monopoly power was undefined, and that was part of the problem,(98) but an even greater problem was pointed out in the separate dissenting opinion of Senator Orrin Hatch:

"In the absence of culpable conduct or some government-conferred (and hence presumably immune) advantage, market leadership can only be the result of relative efficiencies of some kind. Even if they cannot be identified by simple time-and-motion studies, they must be there.

Consider a simple analogy from the field of sports. Suppose a football team "persistently" wins most of its games. Under the majority's theory, proof of this domination would be taken as presumptive evidence that something is wrong - - and then the team would be forced to explain its success by reference to data on the speed of its backs, the heft of its linemen or, perhaps, the intelligence of its coaches. Absent a showing that the team has been violating the rules, however, the best proof of its efficiency would be the record of success which precipitated the proceeding in the first place."(99)

In short, the most significant efficiencies may be hard to quantify, or even identify, but they must be there somewhere if a law-abiding company maintains a leading market position over an extended period of time.(100)

I once used the market experience of General Motors Corp. to make a similar point about this "no fault" monopolization proposal of the Commission.(101) Immediately after it had acquired all of the companies that evolved into its five major automotive divisions,(102) each offering models in separate price ranges, GM's sales were still dwarfed by the dominant seller Ford, which then offered only one model. Ford presumably had vastly greater scale economies. Yet, within a period of less than twenty years, the market shares of the two major rivals had substantially flipped. Where was the efficiency? I suggested at the time that the best explanation was that GM was more efficient in predicting and producing the kinds of cars that people wanted to buy.(103)

Efficiencies of this kind, whether they are called innovation or managerial economies, are probably the most significant variable in determining whether companies succeed or fail - - or in determining whether certain more specific merger efficiencies are achieved or not. Yet, we do not overtly take them into account when deciding merger cases.

In the present state of our knowledge, there is some reason for this neglect. It is easy to look at the sales and profitability of a company like General Motors over a period of over 40 years and conclude that it must be doing something efficient.(104) Similarly, a court could look at the sales performance of a Sylvania and, with the support of economic theory, conclude that its distribution strategies are efficient.(105) But, how do you look at the business plan of an acquirer and decide whether or not it is likely to increase innovation or to achieve the managerial economies that are forecast, when you cannot even be confident that it will manage the combined entity well enough to achieve predicted efficiencies of scale or scope. We tend to ignore the less-tangible economies in the formal decision process because we simply do not know how to weigh them.

One possible approach to the dilemma would be to look at the previous track record of the acquiring company, or of those acquiree-employees who will transfer, and try to decide whether they are likely to manage efficiently in the future. In the GM-Toyota matter, for example, there was substantial evidence that the Japanese partner knew how to build cars much more cheaply. Unfortunately, the past is not always prologue, and people who manage effectively in one era or in one company do not necessarily manage effectively in another time or place, under different conditions. What impact, for better or worse, will the merger itself have on the effectiveness of these management teams?

Our difficulty with these questions does not mean that the less tangible efficiencies should be ignored. The agencies routinely and overtly consider less tangible efficiencies - - including perceived managerial competence - - when they review the qualifications of third parties who come forward to buy assets that merger partners have agreed to divest or carve out of a proposed transaction.(106) In fact, a non-public appendix to the consent agreement may contain the names of key employees who will transfer. It is hard to explain why potential efficiencies that are hard to quantify should be relevant at one stage of a merger proceeding but not at another. Moreover, we cannot be all that precise about the effects of concentration either, and should not impose a higher standard of proof on respondents than on prosecutors.

Fortunately, the situation is not as bad as it looks from the outside. As already stated, the agencies today regularly clear mergers without requiring that plausible efficiencies be quantified, absent very high concentration levels or other factors that raise particular concerns. And, likely efficiencies also can serve as a plus factor to help resolve close issues unrelated to concentration. In Synopsys/Avant!, for example, the potential for a higher level of innovation was one factor that helped me to pass on a vertical merger. In another non-public situation, the potential for improved management of very substantial acquired assets helped me to be comfortable with management plans for the relatively small assets to be divested.(107) I personally reject the notion that it is always improper to balance efficiencies in one area against potential harm in another area, particularly when the former appear to be much more significant than the latter. What we are talking about, after all, are degrees of risk - - not certainties - - and it would be perverse to place arbitrary restrictions on the factors considered.

Issues of this kind are also related to the issue of "merger specificity," which will be addressed immediately below. But, even beyond that issue, it seems clear that we do not deal in a transparent and rigorous way with the less tangible efficiencies, which nonetheless may be most important. We do consider them internally and informally but discount them altogether in a contested transaction because they are often difficult to quantify. We should do more to reconcile our public and our non-public practice.

5. The Requirement that Efficiencies Be Merger Specific

The requirement that efficiencies be merger specific(108) does have respectable antecedents. With considerable fanfare, Congress in 1984 and 1993, overwhelmingly approved two statutes, now known collectively as the National Cooperative Research and Production Act of 1993.(109) As mentioned above, these statutes were overtly intended to encourage the formation of collaborative joint ventures, in order to attain efficiencies in research and production.(110)

At first, it was expected that most of these ventures would be notified to the antitrust agencies, in order to qualify for detrebling, and a substantial number of notifications were filed in the early years.(111) Generally, filings have declined in the recent years. One likely explanation is that counsel have concluded the risks of litigation are minimal. However, it is also possible that some companies have decided total integration by merger is a lot more efficient than partial integration by joint venture.

It may well be true that mergers are more efficient, but they also can have a far greater competitive impact and are much harder to undo if there are problems. Accordingly, they are appropriately subject to a higher level of scrutiny. A proposed joint venture like GM-Toyota would probably be cleared without much fuss today but a proposal to merge General Motors and Toyota outright would likely face an uphill battle.

A focus on the extent to which savings are merger specific might also help to illuminate the issues in a controversial case likeGE/Honeywell.(112) In this case, the defenders of the merger asserted that the combination of General Electric, a leading supplier of aircraft engines, and Honeywell, a leading supplier of aircraft control systems or "avionics," would facilitate the offer of low-priced package deals that would provide immediate consumer benefits. The EU Commissioners, who rejected the merger, agreed that package deals were likely but feared that they would be anticompetitive in the long run. One way to characterize the dispute is simply as a difference of opinion about the proper way to balance prompt and likely short-term effects against speculative long-term effects.(113) If you focus on merger-specificity, however, it becomes clear that the issues are more subtle.

The first step would be to ask why package pricing is more likely if complementary products like engines and avionics are folded into one company. It is not obvious that a merger is a necessary pre-condition for a package deal. Companies can and do offer package deals on complementary products without a merger - - through ad hoc joint bidding arrangements(114) or long-term supply contracts. These arrangements may be much harder to negotiate and administer, however, and the reduction in the transaction costs would be a merger-related efficiency. Package deals may also be more likely because competitive pricing incentives may change as a result of common ownership. The changed pricing incentives of the combined company are not efficiencies, (any more than the changed pricing incentives of the remaining competitors would be) but they clearly are market facts that need to be taken into account.(115) It is not unusual for a transaction to have mixed motivations and effects, but the ultimate competitive impact of a transaction will be better understood if the merger-specific efficiencies can be isolated and considered separately.

Since efficiencies only become an issue when a merger facially appears to raise competitive problems, it is both reasonable and consistent with Congressional intent to inquire whether the efficiencies could be achieved in a less restrictive way, perhaps through a joint venture. If the answer is that the less restrictive alternative is not quite as efficient, then it would seem to me that this differential - - rather than the overall efficiency improvement - - is the appropriate number to weigh in the balance.(116) It is easier to say this, of course, than it is to do it.

6. The Significance of Evidence on Failed Mergers

With all of these variables that are difficult to resolve, it is only appropriate to acknowledge that the evaluation of efficiencies is still a highly subjective exercise. As we have seen, business executives are not very good at it themselves - - a substantial number of transactions do not achieve the shareholder benefits that were predicted.(117) It would be wrong, however, to conclude that a more proactive merger policy is appropriate simply because a lot of mergers do not deliver the expected benefits.

The vast bulk of mergers, even big ones, are competitively benign, and of no concern to competition authorities. It is not our job to protect shareholders against mergers that turn out to be harmful to their interests, even if it were generously assumed that we were competent to do so. Even an efficient merger may produce disappointing financial results (if, for example, too high a price was paid). Competitively benign transactions will succeed or they will fail, and the market will sort out the consequences.

Moreover, there may be good reasons to believe that a particular merger has a better-than-average chance of delivering efficiencies. Suppose, for example, the acquiring company has a history of successfully implementing similar transactions.(118) We might well expect this superior performance to continue. (Compare the track-record on productive efficiencies that was present in GM-Toyota.)

In the absence of a company-specific track record, however, it does seem appropriate to take account of the extant literature in the extent to which mergers generally do or do not work out as well as the parties predicted. For the reasons discussed above, specific efficiency claims need to be approached with some skepticism and, at the same time, the most significant efficiencies probably are neglected because they cannot be quantified at all. We may be able to narrow the zone of uncertainty by noting the relevant factors identified in studies of merger outcomes.

General studies may also properly affect our general approach to risks. There are some cases that involve a delicate balance between risks of over-enforcement and under-enforcement.(119) Obviously, scholarship that improves our understanding of merger efficiencies - - or of other relevant competitive effects - - can contribute to the balancing process.

B. Non-merger Matters

1. Vertical Restraints

As stated above, the Supreme Court recognized in Sylvania that non-price vertical restraints can produce pro-consumer efficiencies.(120) In fact, the Court went further and stated that intrabrand competition is the primary concern of the antitrust laws,(121) and the laterSharp decision went further still and stated that the adverse effects of non-price vertical restraints on intrabrand competition will generally be outweighed by the beneficial effects on interbrand competition.(122) In effect, the efficiencies balance was settled by a judicial presumption. Since that time, not surprisingly, courts have tended to uphold these restraints under the rule of reason(123) and controversy over judicial decisions has largely subsided.(124) The lack of interest may be related to the recent explosive growth of discount retailers, notwithstanding some dire predictions that discounters would be crippled by the Supreme Court's permissive attitude toward vertical restraints.(125)

The issue has recently come alive again, however, in another context. The Internet has made it possible for producers to communicate directly with ultimate consumers more efficiently than ever before, and this potential presents an obvious threat to intermediaries in the distribution chain. Predictably, these middlemen have responded both with private arrangements and with successful appeals for state legislation, designed to preserve the status quo. Some of the issues raised by these defensive measures(126) relate directly to efficiencies.

Those who seek to preserve the viability of existing middlemen may argue, for example, that the potential efficiencies of Internet sales need to be adjusted downward to account for potential adverse effects in quality and services delivered. (They also may raise health and safety concerns which are difficult to weigh in an efficiencies calculus, but somehow the serious concerns have to be distinguished from the trivial.) These matters were considered extensively last month at a workshop sponsored by the Federal Trade Commission.(127) The record developed at these hearings is likely to have a spillover effect on the treatment of efficiencies in other contexts.

2. Other Forms of Collaboration

The potential for efficiencies plays an important role when it is necessary to draw a line between horizontal "cartels" that are subject toper se condemnation and legitimate "joint ventures" that will be judged under the rules of reason. Consider, for example, a group of health-care professionals who want to improve their bargaining power by negotiating as a group with payors. If they can make a plausible case for potential group efficiencies, they are likely to survive agency scrutiny.(128) If they cannot, the group will be sued.(129)

The initial characterization, either as a cartel or a joint venture, is likely to be outcome determinative because it is so much harder to prosecute a rule of reason case against a joint venture. However, there is very little learning on the efficiency evidence that will be sufficient at this preliminary stage. Further elaboration of the evidentiary standards for this critical initial characterization may require a fresh look at the reasons why cartels and joint venture are treated differently in the first place. Antitrust makes a fundamental distinction between the collective action of a cartel and the individual action of a venture that is integrated enough to be treated as a single entity. But, historically, the distinction rests on formalistic rather than pragmatic grounds. (In that respect, it is analogous to the obsession with the locus of "title" that used to be so important in the law of vertical restraints.(130)) There are obvious differences in the efficiency-creating potential of truly integrated joint ventures and so-called "naked" cartels, but the disparate legal consequences may seem arbitrary as you move from these extremes toward a hazy borderline. This anomaly contributes to the political pressure behind special antitrust exemptions, which could also have serious spillover consequences.(131) The matter is worthy of further discussion.

An ex post inquiry to determine whether the promised efficiencies of the medical group have actually been achieved would also raise difficult issues. Efficiency in the delivery of medical services is not easily measured because there is such a significant quality component. Higher-priced quality services may produce better outcomes, for example, and it may be hard to decide whether a reduction in the "output" of services has resulted from market power or from improved efficiency.(132) Similar, if less dramatic, quality issues are present in other sectors of growing importance, which are also characterized by high levels of differentiation.(133)

The analysis of efficiencies in merger cases has stimulated much learned commentary and is the major focus of this article. It would be wrong, however, to treat the subject of merger efficiencies in isolation because potentially controversial issues are popping up again in the non-merger area as well. In fact, it even may be helpful to re-examine the reasons behind the law's preference for internal growth over growth by merger. The rationale is obvious if the merger eliminates a significant competitor, but in other situations it is not.

Conclusion

The formidable array of challenging issues described in Part III may create the erroneous impression that antitrust law's treatment of efficiencies is in bad shape. This is not so. The historical review demonstrates that there has been an immense and continuing improvement over the last 25 years, in both merger and non-merger law. A lot of silly ideas have been discarded and there is near-universal agreement that efficiencies are important. The fact that this change has been primarily driven by scholarship is a remarkable demonstration of the power of ideas.

The most frequently debated problems arise in an area like pre-merger review, where it is necessary to make predictions based on an assumed state that has not existed before.(134) Inside estimates may be unreliable and external estimates may be extraordinarily difficult. These problems seem to be satisfactorily resolved informally most of the time, even if it is difficult to deal with them publicly and formally on the rare occasions when it is necessary.(135) (Whether these contested cases were therefore wrongly decided is another question for which we do not yet have good answers.) The "transparency" problem, resulting from the informal disposition of most cases, should not be exaggerated. The rate of agency challenges has been extraordinarily low throughout the period considered,(136)and this tends to show that agency policies on efficiencies and the other mix of relevant issues are somehow effectively communicated to counselors who need to know.

This article, then, should not be understood as a cry of alarm. The much more modest objective is to isolate issues and hopefully clarify what we know and what we still do not know, in an effort to stimulate further study and debate.

Endnotes:

1. Commissioner, Federal Trade Commission. I wish to acknowledge the assistance of Dara J. Diomande and Thomas J. Klotz in the preparation of this article. As always, the views expressed are my own and not necessarily shared by any other Commissioner.

2. Thomas B. Leary, The Essential Stability of Merger Policy in the United States, 70 Antitrust L.J. 105 (2002).

3. Northern Pac. Ry. v. United States, 356 U.S. 1, 4 (1958)("[The Sherman Act] rests on the premise that the unrestrained interaction of competitive forces will yield the best allocation of our economic resource.").

4. United States v. Aluminum Co. of America, 148 F.2d 416 (2d Cir. 1945).

5. Id. at 428.

6. Id. at 429.

7. United States v. Topco Associates, Inc., 405 U.S. 596 (1972).

8. Id. at 610.

9. Brown Shoe Co. v. United States, 370 U.S. 294 (1962).

10. 15 U.S.C. § 18. Section 7 of the Clayton Act prohibits mergers when "the effect of such acquisition may be substantially to lessen competition, or tend to create a monopoly." In the years since Brown Shoe, legal scholars have expressed widely differing conclusions about the legislative history of the antitrust laws. See, e.g., Thomas B. Leary, Freedom as the Core Value of Antitrust, 68 Antitrust L.J. 545, 546 n.3 (2000)(citing representative views).

11. Brown Shoe Co., 370 U.S. at 344.

12. Foremost Dairies, 60 F.T.C. 944 (1962), modified 67 F.T.C. 282 (1965).

13. Id. at 1084, 1087 (emphasis supplied).

14. I was a junior member of Foremost's defense team.

15. Wesley J. Liebeler, Bureau of Competition: Antitrust Enforcement Activities, in The Federal Trade Commission Since 1970: Economic Regulation and Bureaucratic Behavior 65, 98 (Kenneth W. Clarkson & Timothy J. Muris eds. 1981).

16. United States v. Philadelphia Nat'l Bank, 374 U.S. 321 (1963).

17. See Federal Trade Comm'n, Anticipating the 21st Century: Competition Policy in the New High-Tech, Global Marketplace: A Report by Federal Trade Commission Staff, Ch. 2 (1996)("FTC Global Competition Staff Report").

18. Philadelphia Nat'l Bank, supra n.15 at 371.

19. See pp. 27-32, infra.

20. Standard Oil Co. of California v. United States, 337 U.S. 293, 305 (1949).

21. Id. at 314 ("We conclude, therefore, that the qualifying clause of Sec. 3 is satisfied by proof that competition has been foreclosed in a substantial share of the live commerce affected.")

22. White Motor Co, v. United States, 372 U.S. 253 (1963).

23. United States v. Arnold Schwinn & Co., 388 U.S. 365 (1967). Schwinn was itself reversed ten years later in Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977).

24. R.H. Coase, The Nature of the Firm, 4 Economica 386 (1937). In 1997, Coase won the Nobel Memorial Prize in Economic Science.

25. Oliver E. Williamson, Economics as an Antitrust Defense: The Welfare Tradeoffs, 58 Am. Econ. Rev. 18 (1968).

26. Id. at 34.

27. Ronald H. Coase, Industrial Organization: A Proposal for Research, in Industrial Organization (V. Fuchs, ed. 1972)("If an economist finds something - - a business practice of one sort or another - - that he does not understand he looks for a monopoly explanation.").

28. Oliver E. Williamson, Review of Bowman's Patent and Antitrust Law, 83 Yale L.J. 647, 661 (1974)("Officials charged with enforcing antitrust laws are even more inclined [than economists] to find monopoly purposes lurking in unfamiliar or unconventional business practices.").

29. See Leary, Essential Stability, supra n.1, at 108-11.

30. Utah Pie Co. v. Continental Baking Co., 386 U.S. 685 (1967), 388 U.S. 365 (1967)(finding that defendant's precipitation of a price war violated Section 2(a) of the Robinson-Patman Act, even though the plaintiff remained a profitable market leader); United States v. Arnold, Schwinn & Co., 388 U.S. 365 (1967)(holding territorial and customer restrictions per se illegal); Albrecht v. Herald Co., 390 U.S. 145 (1968)(finding maximum price fixing per se illegal).

31. Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36 (1977)(adopting a rule of reason analysis for non-price vertical restraints);Brunswick Corp. v. Pueblo Bowl-O-Mat Inc., 429 U.S. 477 (1977)(imposing an antitrust injury requirement); United States Steel Corp v. Fortner Enters., 429 U.S. 720 (1977)(requiring "sufficient economic power" over the tying product in tying arrangements).

32. William F. Baxter, Antitrust: A Policy in Search of Itself, 54 Antitrust L.J. 15, 16 (1985).

33. See, e.g., Thomas B. Leary, The Significance of Variety in Antitrust Analysis, 68 Antitrust L.J. 1007 (2001); Leary, Essential Stability,supra n.1; Thomas B. Leary, The Dialogue Between Students of Business & Students of Antitrust, N.Y.L. Sch. L. Rev. (forthcoming).

34. Cargill, Inc. v. Monfort of Colo., Inc., 479 U.S. 104 (1986).

35. Id. at 122 n.17 (quoting Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 594 (1986)); Lawrence A. Sullivan & Warren S. Grimes, The Law of Antitrust: An Integrated Handbook 145 (2000).

36. Id. at 116 (quoting Arthur S. Langenderfer, Inc. v. S.E. Johnson Co., 729 F.2d 1050, 1057 (6th Cir. 1984)).

37. Cargill really turns on the absence of "antitrust injury" from threatened lower prices, in the spirit of Brunswick, supra n.30. In fact, the Justice Department's amicus brief argued that competitors should not have standing to attack mergers at all.

38. William Kolasky and Andrew Dick closely examined judicial recognition of efficiencies (post-Brown Shoe). See William J. Kolasky & Andrew Dick, The Merger Guidelines and the Integration of Efficiencies into Antitrust Reviews of Horizontal Mergers, for the 20th Anniversary of the 1982 Merger Guidelines (June 10, 2002).

39. FTC v. Tenet Healthcare Corp., 186 F.3d 1045 (8th Cir. 1999); FTC v. Butterworth Health Corp., 121 F.3d 708 (6th Cir. 1997).

40. FTC v. University Health, Inc., 938 F.2d 1206 (11th Cir. 1991).

41. FTC v. H.J. Heinz, Co., 246 F.3d 708, 720 (D.C. Cir. 2001).

42. FTC v. Staples, Inc., 970 F. Supp. 1066 (D.D.C. 1997); FTC v. Cardinal Health, Inc., 12 F. Supp.2d 34 (D.D.C. 1998); FTC v. Swedish Match, 131 F. Supp.2d 151 (D.D.C. 2000). See also Kolasky, supra n.37.

43. U.S. Dep't of Justice, Merger Guidelines (1982), reprinted in 4 Trade Reg. Rep. (CCH) ¶ 13,102 [hereinafter 1982 DOJ Guidelines]; Fed'l Trade Comm'n, Statement Concerning Horizontal Mergers (1982), reprinted in 4 Trade Reg. Rep. (CCH) ¶ 13,200; U.S. Dep't of Justice, Merger Guidelines (1984), reprinted in 4 Trade Reg. Rep. (CCH) ¶ 13,103 [hereinafter 1984 Guidelines]; U.S. Dep't of Justice and Fed'l Trade Comm'n, Merger Guidelines (1992), reprinted in 4 Trade Reg. Rep. (CCH) ¶ 13,104 [hereinafter 1992 Guidelines]; U.S. Dep't of Justice and Fed'l Trade Comm'n, Merger Guidelines, 4 Trade Reg. Rep. (CCH) ¶ 13,104 (1992)(with April 8, 1997 revisions to § 4)[hereinafter 1997 Revision to Sec. 4]. The overall trends of these guidelines have already been outlined in my previous article on merger policy. See Leary, Essential Stability, supra n.1, at 114-21.

44. 1982 DOJ Guidelines, § V.A.

45. 1982 FTC Statement, § IV.

46. 1984 Guidelines, § 3.5.

47. 1992 Guidelines, § 4.

48. 1997 Revision to Sec. 4.

49. General Motors Corp., 103 F.T.C. 374 (1984).

50. See David Roos, "Automobile Competitiveness: A Report of the MIT International Motor Vehicle Program" (Statement prepared for the Capitol Hill Forum on Industrial Competitiveness in September 1995 and submitted as part of Professor Roos' presentation at the FTC Hearings on Global and Innovation-Driven Competition, Oct. 18, 1995)(describing significance of GM-Toyota joint venture for GM's adoption of production and management innovations).

51. General Motors Corp., supra n.48, at 397 (dissenting statement of Commissioner Patricia P. Bailey).

52. General Motors Corp., 116 F.T.C. 1276, 1285 (1993)("There appears to be no continuing need for the order's restrictions on the duration and scope of the joint venture, and continuing the restrictions in the context of the changed conditions may hinder the ability of the joint venture to respond to consumer demand.").

53. National Cooperative Research and Production Act of 1993, 15 U.S.C. § 4301-05 (2000).

54. Addamax v. Open Software Found., 888 F. Supp. 274, 281-83 (D. Mass. 1995), aff'd, 152 F.3d 48 (1st Cir. 1998)(case also involved issues related to standardization).

55. Robert Pitofsky, Proposals for Revised United States Merger Enforcement in a Global Economy, 81 Geo. L.J. 195, 197-98 (1992).

56. FTC Global Competition Staff Report, supra n.16, at 25.

57. Timothy J. Muris, The Government and Merger Efficiencies: Still Hostile After All These Years, 7 Geo. Mason L. Rev. 729, 734 (1999).

58. Id.

59. Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977).

60. Business Elecs. Corp. v. Sharp Elecs. Corp., 485 U.S. 717 (1988).

61. State Oil Co. v. Kahn, 522 U.S. 3 (1997).

62. See discussion supra pp. 7-9.

63. Greg Werden has gone further and argued that, contrary to common perception, most merger efficiencies are vertical in nature. SeeGregory J. Werden, An Economic Perspective on the Analysis of Merger Efficiencies, Antitrust, Summer 1997.

64. Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2 (1984).

65. See Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451, 462 (1992)(citing Jefferson Parish).

66. See, e.g., discussion of authorities in ABA Section of Antitrust Law, Antitrust Law Developments (Fifth) 186-91 (2002).

67. See, id. at 207-09.

68. See Richard M. Steuer, Clarity and Confusion in Vertical Restraints, 58 Antitrust L.J. 421, 430 (1989).

69. Dr. Miles Medical Co. v. John D. Park & Sons Co., 220 U.S. 373 (1911).

70. But see Continental T.V., Inc. v. GTE Sylvania Inc., supra n.22, at 58 n.29 ("We are unable to perceive significant social gain from channeling transactions into one form or another.").

71. See discussion infra pp. 41-43.

72. Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263, 275 (2d Cir. 1979).

73. See Brown Shoe Co. v. United States, supra n.8, at 344.

74. See AmeriSource Health Corp., FTC File No. 011-0122 (Aug. 24, 2001)(Statement of the Commission), available at/os/2001/08/amerisourcestatement.pdf

75. See Synopsis Inc., FTC File No. 021-0049 (July 26, 2002)(Statements of Commissioners Anthony, Thompson and Leary), available at /os/2002/07/advantanthonystmnt.htm/os/2002/07/advantthompsonstmnt.htm/os/2002/07/advantlearystmt.htm.

76. Malcolm B. Coate, Merger Enforcement at the Federal Trade Commission in Three Presidential Administrations, 45 Antitrust Bull. 323 (2000).

77. Most "monopolization" cases under Section 2 of the Sherman Act also involve predictions.

78. FTC v. Staples, Inc., supra n.41.

79. An example close to home is my own rationale for supporting the complaint in FTC v. H.J. Heinz, Co., supra n.40; Thomas B. Leary,An Inside Look at the Heinz Case, Antitrust, Spring 2002, at 32.

80. Vertical transactions can be even more complicated.

81. For instance, when estimating unilateral competitive effects, there are competing models of non-cooperative oligopoly pricing. In the Cournot model, firms seek to maximize profit by setting output, taking their rivals' outputs as given. In the Bertrand model, firms seek to maximize profit by setting price. The assumption regarding the nature of the competitive process will affect the predicted effects of a transaction.

82. Dennis W. Carlton, Presentation at Federal Trade Commission Empirical Industrial Organization Roundtable, 10 (Sept. 11, 2001)available at /be/empiricalioroundtabletranscript.pdf

83. Anyone who has worked in the corporate world or studied it closely also knows that a lot of internal estimates are not objective, either. Even (or especially) those presented to the Board are likely to be biased in favor of whatever the current CEO wants to do.

84. See, e.g., John Kelly & Colin Cook, Synergies: A Business Guide, KPMG (2001); Norm Augustine, Corporate Marriage: Bliss or Blight? A Monograph on Post-Merger Integration, A.T. Kearney (Apr. 1999); Gerry Adolph et al., Merger Integration: Delivering on the Promise, Research Summary, Booz-Allen & Hamilton (2001); Dorian Swerdlow et al., Managing Procurement Through a Merger: Capturing the Value of the Deal, Booz-Allen & Hamilton (2001).

85. See, e.g., Donald Shay et al., Speed Makes the Difference: A Survey of Mergers & Acquisitions, Price Waterhouse Coopers (2000).

86. Competitors can be a valuable source of industry information, although it is always necessary to discount their arguments (like the arguments of the parties) for obvious self-interest. They may really be most afraid of the most pro-competitive transactions.

87. See ABA Section of Antitrust Law, Perspectives on the Concepts of Time, Change, & Materiality in Antitrust Enforcement 320-22 (2001). There is, however, some survey literature on factors that may make some deals work better than others - - a very significant one seems to be speed of implementation. See, e.g., Shay supra n.84; Dean McCauley, Executing the Successful Merger: Smart Play in a High Risk Game, CSC Index Genesis (1997).

88. Janusz A. Ordover, Presentation at FTC Empirical Industrial Organization Roundtable, supra n.81, at 53. (regarding predictions of some econometric models, "however much I adore these models, I do remain skeptical about the ability to draw profound conclusions from such pinpointed estimates, with extremely tight standard errors around these estimates").

89. The speaker was Don Baker, but I do not remember the occasion.

90. The courts have also focused on whether efficiencies are likely to be passed on to consumers. See FTC v. University Health, Inc., 938 F.2d 1206, 1223 (11th Cir. 1991); FTC v. Butterworth Health Corporation, 946 F. Supp. 1285, 1301 (W.D. Mich. 1996).

91. Compare Wealth Transfers as the Original and Primary Concern of Antitrust: The Efficiency Interpretation Challenged, 34 Hastings L.J. 65, 68 (1982), with ABA Section of Antitrust Law, Mergers and Acquisitions, Understanding the Antitrust Laws 145 (Robert S. Schlossberg & Clifford H. Aronson eds. 2000).

92. Commissioner of Competition v. Superior Propane Inc., 2000 Comp. Trib. 15, File No. T1998002; The Commissioner of Competition v. Superior Propane and IOG Propane Inc., 199 D.L.R. (4th) 130 (2001).

93. See also, Leary, supra n.78.

94. Dennis A. Yao and Thomas N. Dahdouh, Information Problems in Merger Decision Making and Their Impact on Development of an Efficiencies Defense, 62 Antitrust L.J. 2, 41-43 (1993). In fact, a rigorous pass-on requirement could effectively vitiate an efficiencies "defense" only in situations where it might be needed in the first place. See Pitofsky, supra n.54, at 207-08. Another provocative article argues that a pass-on requirement should be rejected for an entirely different reason. See Paul L. Yde and Michael G. Vita, Merger Efficiencies: Reconsidering the "Passing-on" Requirement, 64 Antitrust L.J. 735, 740 (1996)(a prediction of economic theory that "the more competitive the relevant market, the less likely it is that merger-specific efficiencies will be reflected in the post-merger market price.")(emphasis in original).

95. In the standard schematic diagrams showing the effects of changes in supply and demand, cost savings and wealth transfers show up as changes in the dimensions of rectangles and allocative effects show up as changes in the dimensions of triangles. See Terry Calvani, Rectangles & Triangles: A Response to Mr. Lande, 58 Antitrust L.J. 657, 659 (1989)("[t]he debate over rectangles and triangles is trivial").

96. Report to the President and the Attorney General of the Nat'l Comm. for the Review of Antitrust Laws and Procedures (Jan. 22, 1979).

97. Id. at 141-42.

98. Id. at 357-58 (Separate views of Senator Hatch).

99. Id. at 361 (Hatch dissent). The athletic analogy is not at all farfetched. Business competition is also a team sport.

100. Some critics have since argued that persistent leadership may result, not from efficiency, but from so-called "network effects." The idea is that if the benefits to a consumer of a particular product or service increase in proportion to the number of other consumers who buy the same thing, even an inefficient supplier with a first-mover advantage can enjoy prolonged dominance. Although the theory sounds plausible, it suffers from a lack of empirical support. Timothy J. Muris, The Federal Trade Commission and the Law of Monopolization, 67 Antitrust L.J. 693, 718 (2000).

101. Thomas B. Leary, Do the Proposals Make Any Sense From a Business Standpoint?, 49 Antitrust L.J. 1281, 1285-88 (1980).

102. They were Cadillac, Buick, Oldsmobile, Pontiac and Chevrolet.

103. This superior efficiency was apparently sufficient to overcome whatever "network effects" Ford may have enjoyed through a presumably more extensive distribution network.

104. In later years, it was evident that some competitors had become even more efficient.

105. In fact, however, courts have tended to go further and simply assume the efficiencies in cases involving non-price vertical restraints.See, e.g., Richard M. Steuer, supra n.67.

106. See, e.g., William J. Baer, Testimony before the House Committee on the Judiciary Concerning the Effects of Consolidation on the State of the Competition in the Financial Services Industry (June 3, 1998) available at /os/1998/06/finanser.tes.htm ("Designing divestitures in retail markets can be particularly difficult. It is often critical to require a divestiture of a sufficient set of retail locations to a single buyer. Divestiture to a single buyer is often preferable so that a firm can acquire the full range of distributional and advertising efficiencies.").

107. In fact, I thought it would be ludicrous to explore the managerial competence of a prospective buyer of divested assets and, at the same time, to entirely ignore the managerial competence of the buyer in the principal transaction.

108. 1997 Revision to Sec. 4 ("The Agency will only consider these efficiencies likely to be accomplished with the proposed merger and unlikely to be accomplished in the absence of either the proposed merger or another means having comparable anticompetitive effects.").

109. 15 U.S.C. § 4301-05 (2000). See also discussion supra p. 17.

110. See H. Conf. Rep. No. 98-1044, 87th Cong., 2d Sess. 8 (1984), reprinted in 1984 U.S.C.C.A.N. 3131, 3132.

111. Thomas B. Leary, Antitrust Problems in International Trade - The Congressional Response, Ch. 1, in Private Investors Abroad - Problems and Solutions in International Business in 1985, The Southwestern Legal Foundation (Janice R. Moss, ed. 1985).

112. General Electric/Honeywell, Case No. Comp/M. 2220 (2001).

113. See, e.g., the contrasting comments of Francisco-Enrique Gonzalez-Diaz and Carl Shapiro, Roundtable Discussion [of Transatlantic Antitrust: Convergence or Divergence], Antitrust, Fall 2001, at 7.

114. These can be perfectly legal if openly disclosed to customers.

115. Carl Shapiro argues that it doesn't make any difference whether changed incentives are called efficiencies or not because, presumably, both have a moderating effect on price. See Shapiro, supra n.112, at 10. It does make a difference. For one thing, they may change in different ways over time.

116. See Leary, supra n.78, at 32, 33.

117. See discussion supra pp. 29-30.

118. See, e.g., Adolph, supra n.84; Max M. Habeck et al., After the Merger: Seven Rules for Post-Merger Integration, A.T. Kearny (1999).

119. Some scholars have suggested that it is possible to evaluate these risks in an organized way. But, the ultimate conclusion still depends on the data that are plugged into the scheme, and it still will be subjective to the extent the data are controversial.

120. Sylvania, supra n.22.

121. Id. at 42 n.19.

122. Sharp, supra n.59, at 725.

123. See discussion of authorities in ABA Section of Antitrust Law, Antitrust Law Developments (Fifth) 154-58 (2002).

124. See 96 Cong. Rec. H5663, 1 Cong. Index (CCH)(June 30, 1992) at 23,052 (bill to amend the Sherman Act regarding retail competition. For: 175; Against: 225). But see, e.g., Warren S. Grimes, Market Definition in Franchise Antitrust Claims: Relational Market Power and the Franchisor's Conflict of Interest, 67 Antitrust L.J. 243 (1999).

125. See Senator Howard A. Metzenbaum's comments on proposed bill S. 430, Retail Competition Enforcement Act, 134 Cong. Rec. S.8296-04 (June 1988). ("In the Sharp decision, the Supreme Court found that the agreement between the high priced store and the manufacturing company to cutoff a distributor because it is charging low prices is not automatically anticompetitive. I have difficulty in understanding that, I might say. It is hard to imagine a more anticompetitive agreement. The case has already hurt discounters.")

126. There are also a number of other issues, like the applicability of the "state action" doctrine and the commerce clause.

127. Federal Trade Commission Workshop on Possible Anticompetitive Efforts to Restrict Competition on the Internet (Oct. 8-10, 2002).

128. The applicable guidelines for evaluating these ventures acknowledge that efficiencies may arise from two kinds of integration: so-called "financial" integration, which creates common incentives for efficient operation, and "clinical" integration, which may achieve results more directly. MedSouth, Inc., opinion letter, available at /opa/2002/02/medsouth.htm. See U.S. Dep't of Justice and Fed'l Trade Comm'n, Statements of Antitrust Enforcement Policy in Health Care (1996) reprinted in 4 Trade Reg. Rep. (CCH) ¶ 13,153.

129. System Health Providers, File No. 0110196, available at /opa/2002/08/shp.htm

130. "Title" is still important in a resale pricing case. See Dr. Miles Medical Co. v. John D. Park & Sons Co, 220 U.S. 373 (1911).

131. See discussion of the so-called "Campbell bill" in Warren S. Grimes, The Sherman Act's Unintended Bias Against Lilliputans: Small Players' Collective Action As a Counter to Relational Market Power, 69 Antitrust L.J. 195, 212-216 (2001). Grimes proposes an alternative "safe harbor" of his own for collective action by "small players," but it is not clear to me that the potential improvement on accuracy would be worth the substantial decline in predictability.

132. For an extended discussion of these issues, see Thomas B. Leary, The Antitrust Implications of Clinical Integration: An Analysis of FTC Staff's Advisory Opinion to MedSouth, St. Louis U. L.J. (forthcoming).

133. See Leary, The Significance of Variety in Antitrust Analysis, supra n.32.

134. Similar difficulties of proposed prediction arise in the context of an advisory opinion about a venture like MedSouth, discussedsupra pp. 48-49.

135. However, the Federal Trade Commission has recently made a conscious effort to volunteer explanations for its internal decisions.See Synopsys, supra n.74; Royal Carribean Cruises, Ltd., FTC File No. 0210041 (Oct. 4, 2002)(Statement of the Commission), available at /os/2002/10/cruisestatement.htm and (Dissenting Statements of Commissioners Anthony and Thompson) available at/os/2002/10/cruisedissent.htm.

136. Thomas B. Leary, The Dialogue Between Students of Business & Students of Antitrust, supra n.32.