This paper addresses three issues regarding the antitrust analysis of joint ventures: 1) applying the per se rule and the rule of reason; 2) identifying and evaluating "classic" market power; and 3) identifying and evaluating "exclusionary market power." I also discuss the possibility of joint FTC/DOJ competitor collaboration guidelines and include some suggested guideline language as an Appendix.

I. INTRODUCTION

One of the most challenging problems in articulating enforcement policy regarding joint ventures is the breadth of the subject. At one time -- and to some extent, still today -- practitioners and firm managers use the term "joint venture" to refer to a separate corporation owned by independent parents.(2) A broader definition includes all cases where firms collaborate in carrying on some activity that each firm might otherwise perform alone.(3) Sometimes the term has been used to refer to virtually any collaboration by competitors, short of merger.(4) This definition sweeps in a vast range of joint activity, from a highly integrated production joint venture, to a loosely integrated marketing network, to a set of ethical rules regarding advertising.(5)

A threshold question, then, is how broadly to view joint venture enforcement policy. The term "joint venture" should be limited to cases where firms are cooperating to perform some activity, rather than simply agreeing about how to perform individual activity. However, there are frequently ancillary agreements associated with the joint activity, which restrict individual firm conduct outside of the joint venture. The evaluation of these restraints then becomes part of "joint venture policy," too.

One could attempt to distinguish types of cooperation based on the form of the collaboration, e.g., whether competitors jointly own a separate firm. Usually, however, the same results can be achieved with a wide variety of organizational forms. Consequently, such distinctions risk being overly-formalistic. Many of the same policy issues arise whether competitors choose to collaborate by contract, by unenforceable agreements, by taking equity ownership in a third entity, or by taking an equity interest in each other.(6)

In fact, the need to elevate substance over form seems particularly important in this area. Consequently, it is appropriate to conceive of "joint venture enforcement policy" -- and guidelines articulating such a policy -- as broad enough to encompass all these forms of collaboration. I suggest that a statement of the antitrust agencies' enforcement policies in this area be called "Competitor Collaboration Guidelines" to reflect this broad conception.(7)

Such guidelines would be useful for several reasons. First, it is quite clear that collaboration among competitors is increasing in frequency and significance. The importance of innovation means this trend will almost certainly continue. Second, the courts have not developed a clear framework for analyzing collaboration. Consequently, there is some uncertainty in the legal and business communities about how the courts and agencies will analyze competitor collaboration. Third, the brief chapter on joint ventures in the 1988 Guidelines for International Operations(8) was viewed by many as a helpful statement of policy. No doubt it is still being relied on by some practitioners. The 1995 International Guidelines(9) were issued without a statement about joint ventures. The Intellectual Property Guidelines does contain a short section on joint ventures,(10) but there is no general discussion that has the endorsement of the agencies. Finally, there will be some value in publishing guidelines for comment, even if they are never issued. Joint venture hearings, while useful, are inherently somewhat unfocused because of the breadth of the topic. Comments on draft policies will certainly focus attention on particular policy approaches.

II. THE KEY ISSUES

The central issues to be addressed by any policy statements are the following:

1. What is the appropriate analytical approach for evaluating various forms of competitor collaboration? Under what circumstances should the per se rule be applied? Under what circumstances is some "truncated" or "limited" rule of reason appropriate?

2. When "classic" market power is relevant, how should it be identified and evaluated? What different considerations arise in evaluating the market power of different forms of joint ventures? What weight should be given to the possibility of continued competition by the collaborators outside the scope of the joint undertaking?

3. When "exclusionary" market power is relevant, how should it be identified and evaluated? How should the agencies evaluation situations when collaborators exclude outsiders? Are there particular types of exclusions, e.g., standard-setting activity, that warrant a special approach?

4. What standards should be applied in evaluating "ancillary restraints," i.e., agreements among competitors that accompany the underlying agreement to collaborate? For example, should the agencies require the "least restrictive alternative," require only that the restraints be no greater than "reasonably necessary," or apply some other test?

5. How should the agencies identify and evaluate "spillover effects," i.e., harm to competition outside the scope of the collaboration, including competition in other markets? Which kinds of collaboration have the greatest potential for spillover effects? Under what circumstances will the agencies give weight to procedures adopted by competitors to limit these effects?

A set of guidelines addressing each of these questions, with accompanying hypothetical cases, would address the great majority of issues that antitrust counsel face in advising clients about collaborating with competitors. I discuss the first three of these areas in this paper.

III. THE APPROPRIATE ANTITRUST ANALYSIS

Even at this late date, there is not a clear consensus on the application of the per se rule and the rule of reason. The Supreme Court has stated various formulations of both.(11) The Commission commented extensively (and inconsistently) about the role and application of the per se rule in Mass. Board of Optometry ("Mass. Board")(12) and California Dental Association ("CDA").(13) The Antitrust Division proposed a framework recently.(14) The 1988 International Guidelines described an analytical approach.(15) The Intellectual Property Guidelines include a framework for applying the rule of reason.(16) The Health Care Policy Statements set out a framework as well.(17) These various policy statements are not entirely consistent. In addition, they leave certain important questions unanswered. In this part of the paper, I discuss whether guidelines could make a contribution in clarifying the agencies' approach to the per se rule and the rule of reason.

Obviously, any agency statements must conform to the requirements set out in the NCRPA as well as any binding principles established by the courts.(18) However, within those constraints, the agencies have considerable discretion in deciding how to analyze these arrangements. In considering what statement the agencies should make in this area (if any), it is useful to separate differences in substantive analysis from differences in terminology. Substantive differences clearly lead to real differences in outcome for particular cases. Differences in terminology may not, if properly understood, reflect a difference in substance. However, they can create enough uncertainty and confusion that they lead to differences in result as well.

A. Categorizing Cases

One way of analyzing the application of the per se rule and the rule of reason is to identify distinct categories of cases, which at various time, have been analyzed differently by the agencies and the courts. Below I suggest a spectrum of five categories of cases. In general, as cases move along this spectrum, they increase in complexity and difficulty of analysis. The key policy issues have to do with how these categories should be "labeled," how they should be distinguished from each other, and how the analysis of each category should proceed.

Category I. Traditional, "Blatant" Per Se Violations

This category includes agreements that have long been identified as per se violations, including seller and buyer price-fixing cartels, output restrictions, certain types of boycotts,(19) and market allocation schemes. These restraints have no plausible procompetitive justifications.

Category II. Agreements Substantively Equivalent to Traditional Per Se Violations

This category includes agreements that differ only in form from traditional violations included in Category 1. This category includes barely disguised cartels as well as restrictions on price and output that arise from loosely integrated competitor collaboration. For example, an agreement not to engage in competitive bidding and an agreement not to offer credit-free transactions are forms of price-fixing.(20) Like Category I, these restraints have no plausible procompetitive justifications, though the parties may offer some non-competitive justifications.

Category III. Other Restraints with No Procompetitive Justifications

This category includes significant, obvious restraints on competition that have not traditionally been viewed as per se violations but could be identified as "new additions" to the per se category. The advertising restrictions found to be per se unlawful in CDA and the licensing restrictions recently challenged by the Antitrust Division(21) arguably fit into this category. Like restraints in the first two categories, these restraints have no plausible procompetitive justifications though the parties may offer non-competitive justifications.

Category IV. Significant Restraints with Plausible, though not Substantial, Procompetitive Justifications

This category includes significant, obvious restraints on competition, which have some plausible procompetitive justifications. While there are plausible justifications for these restrictions, they are insubstantial in light of the significance of the restraints. Therefore, it is (at least arguably ) possible to conclude that these are unlawful without regard to market power.

Category V. Significant Restraints with More Substantial Justifications

This category includes significant restraints on competition, which have legitimate, perhaps substantial, procompetitive justifications. The overall effects of the restraints are difficult to determine.

There is no particular disagreement -- among the agencies and commentators at least -- about the substantive treatment of Categories I and II. Both agencies and most commentators would find restraints in these categories unlawful without an assessment of market power or actual competitive effects.(22)

In addition, there is probably no serious disagreement about cases in Category V. The effects of the restraints in these cases are difficult to determine based on a limited review of the transaction. Therefore, an extensive rule of reason analysis must be employed before the agencies (and generally courts) will find them unlawful. However, there are significant disagreements about substantive analysis for Categories III and IV, and there are disagreements about terminology even for Categories I and II.

B. The Initial Characterization

All analytical approaches begin with an initial characterization of an agreement. Such a characterization is based on a summary review of the nature and purpose of the agreement. If an agreement obviously fits into a category of restraints "which because of their pernicious effect on competition and lack of any redeeming virtue are conclusively presumed to be unreasonable,"(23) the restriction is unlawful without an examination of market power. These are Category I cases.

There is no disagreement among the agencies and commentators that the "literal" nature of the arrangement is not dispositive in making this initial characterization.(24) Cases in Category II -- for example, a bare arrangement to fix prices that is structured in form as a joint venture -- can also be held unlawful without an examination of market power. At the same time, the parties to a price-fixing agreement may be able to show plausible justifications.(25) Thus, there has been a general rejection of the formalistic approach of Topco, which labels a restraint as per se unlawful based on the existence of certain characteristics, such as territorial allocation, without regard to any other factors.(26) The Supreme Court's position is not as clear as it might be. For example, in Palmer, the Court cited Topco with approval.(27) Palmer should not necessarily be read as endorsement of Topco formalism.(28) Nevertheless, Topco and a handful of other early Supreme Court cases continue to cast a cloud of uncertainty over whether some agreements are prohibited under the per se rule based on their form alone.(29)

The Commission in Mass. Board rejected use of the per se/rule of reason distinction in favor of an initial determination whether an agreement is "inherently suspect."(30) "Inherently suspect" restraints were defined to be restraints that "appear likely, absent an efficiency justification, to 'restrict competition and decrease output.'"(31) If the restraints were not "inherently suspect," the analysis proceeded to an extensive rule of reason, including an examination of market power. If the restraints were deemed inherently suspect, and there were no plausible procompetitive justifications, they were deemed unlawful. If these justifications were "plausible,"(32) the Commission was to determine if they were "valid." If they were not "valid," the restraint was deemed unlawful without an assessment of market power. If the justifications were "valid," then the Commission was to proceed to a "full balancing test of the rule of reason."(33)

In contrast to Mass. Board, the Commission opinion in CDA endorsed the per se/rule of reason distinction. The Commission stated that the Supreme Court had used the per se/rule of reason distinction to reflect the difference between an "abbreviated, categorical approach" and an "individualized, contextual examination."(34) The Commission majority then stated an intention to use the per se/rule of reason terminology to differentiate the Commission's own analyses.(35)

On the facts before it, the majority in CDA went on to conclude that a dental association's restrictions on price advertising were per se unlawful. The opinion noted that this type of restraint had not clearly been classified as per se unlawful in previous cases.(36) On the other hand, the Commission concluded that restraints on non-price advertising were not per se unlawful. This conclusion was based on the facts that the effect on competition was not as obvious, the courts have had less experience in evaluating this type of restriction, and the Supreme Court has expressed reluctance in applying the per se rule to professional associations.(37) Thus, these non-price restraints should be evaluated under the "default, rule-of-reason analysis."(38) However, the Commission concluded that it could find these restrictions unlawful without a "detailed" examination of market power because "the anticompetitive effects ...are sufficiently clear and the claimed efficiencies are sufficiently tenuous."(39)

C. Consideration of Procompetitive Justifications

How different are Mass. Board and CDA? One possible difference concerns the opportunities for the parties to offer procompetitive justifications. Mass. Board invites the parties to offer procompetitive justifications for any restraints, including those that are "inherently suspect."(40) Mass. Board rejects prior Supreme Court categorization of cases as per se unlawful. Instead, a new category of "inherently suspect" restraints must be identified over time. The majority in Mass. Board considered this lack of rigidity as an advantage, primarily because it avoided sweeping cases into the per se net that did not belong.(41) On the other hand, the pliable nature of this "inherently suspect" categorization also lends itself to misuse. To the extent there is a temptation to apply this label too loosely, parties could routinely find themselves having to offer procompetitive justifications or lose the case at the initial characterization stage. Even the overuse of the "inherently suspect" categorization might not lead to mistakes in enforcement decisions, however, if the parties can easily meet their burden of presenting plausible procompetitive justifications. Mass. Board assumed that this burden would indeed be modest -- requiring only that the efficiency argument "cannot be rejected without extensive factual inquiry."(42)

Under Mass. Board, the lack of a defined category of per se unlawful restraints also has implications for future cases involving restraints similar to ones the Commission found unlawful. For example, presumably any future litigant could offer procompetitive justifications for advertising restrictions even if they were virtually identical to the ones condemned in Mass. Board. Similarly, if the Commission's approach in CDA always allows parties to offer procompetitive justifications, which could, in effect, "remove" the per se categorization, then the initial characterization is not dispositive. The per se label simply serves to distinguish those cases where procompetitive justifications turn out to be plausible from those cases where they are not. If there are plausible procompetitive justifications, the case then "moves" to the next stage of the process. Under that interpretation Mass. Board and CDA would differ (at least on this point), in terminology only. In contrast, if the Commission's "categorical, per se approach" in CDA precludes parties from offering procompetitive justifications for restraints identified to be per se unlawful, then procompetitive justifications in future cases are irrelevant. The irrelevance of procompetitive justifications in future cases, in fact, might be the whole point of a "categorical" approach.

Despite their different language, it is not clear if the two cases actually differ in allowing the parties to offer procompetitive justifications. In CDA, the Commission said it was willing to consider possible procompetitive justifications for the price advertising restrictions but the parties had offered none.(43) However, will future parties with the same restraints be able to offer procompetitive justifications? Some language in the opinion suggests that they will not if the conduct is "equivalent."(44) On the other hand, would the Commission in a future case instruct its administrative law judges to exclude as irrelevant claims that a ban on price advertising are procompetitive? Presumably, such arguments would not be given much weight and the parties' attempts to distinguish the restraints condemned in CDA and their own particular case would be rejected as well. However, if the Commission is at least going to listen, then there truly is no substantive difference between Mass. Board and CDA at this stage of the analysis. Under both cases, the Commission will consider possible procompetitive justifications and determine whether they are "plausible." If they are not, restrictions with such a high potential for competitive harm are unlawful. If parties offer plausible procompetitive justifications, then the transaction must be evaluated under some form of rule of reason.

While CDA appears ambiguous on this point, the Antitrust Division has endorsed the position that a per se characterization means that procompetitive arguments are irrelevant. In a recent case, the Division alleged that certain restrictions in licensing agreements between General Electric and hospitals are per se illegal.(45) The Division stated that this kind of restraint is, in substance, the same as the territorial allocation scheme struck down in Palmer. It acknowledged that this precise scheme had not previously been branded per se unlawful.(46) However, according to the Division, possible procompetitive justifications for arrangements determined to be in the per se category are simply not considered.(47) The sole question is whether the agreement is "ancillary" to some other arrangement.(48)

The comments by the agencies suggest that the Antitrust Division, and perhaps the Commission, have retained the traditional categorical approach that has occasionally plagued courts in the past -- identifying a class of restraints as unlawful and closing off procompetitive arguments for any restraints within the class. In effect, the per se categorization precedes the decision to consider procompetitive justifications, rather than the consideration of justifications preceding the categorization. As a result, courts can fail to consider procompetitive arguments that might justify further inquiry.(49) In short, too much depends on initial characterization and not enough depends on actual effects on competition.(50)

D. Applying a Limited Rule of Reason

A limited rule of reason lies somewhere between per se condemnation and a full-blown rule of reason.(51) There is broad support for a limited rule of reason, but there is significant disagreement about how and when it should be applied. The cases and commentary suggest three versions of a limited rule of reason, ranging from a version that puts the greatest burden on the party challenging a restraint to the version that puts the least burden on that party. For want of better terminology, I will refer to these as the "weak," "medium," and "strong" versions. The weak version is the one that puts the greatest burden on the party challenging the restraint. Under the weak version, it is appropriate to dispense with a market power inquiry only when an agreement substantially restrains competition and there are no plausible procompetitive justifications. In that case, the government or private plaintiff need not present any evidence of market power or anticompetitive effects. However, if there are plausible procompetitive justifications, then the party challenging the restraint must make a detailed showing of market power. Following the categorization of cases suggested above, we can say that the weak version of a limited rule of reason has its impact in the analysis of cases in Category III (and perhaps in some cases in Category II). Restraints in these cases, though perhaps not per se unlawful, have no plausible procompetitive justifications. Consequently, there is no need to determine market power because there is a substantial risk of competitive harm and no prospect of competitive benefit. To use traditional (but ambiguous) language, we can say that the weak version of a limited rule of reason serves to dispense with the market power inquiry only in the cases of "naked restraints." Arguably, that position is consistent with the both NCAA and Indiana Federation of Dentists.(52)

A medium version of a limited rule of reason also serves to dispense entirely with a market power analysis when there are no plausible procompetitive justifications. However, the medium version has some effect when there are at least "plausible" procompetitive justifications. In the event that these justifications are plausible, but upon examination turn out to be insubstantial, the medium version serves to "truncate" the market power inquiry. The party challenging the restraint need only present some evidence that an agreement has effects in the marketplace, but need not make a full-blown demonstration of market definition, market share, barriers to entry and so on. Following the categorization suggested above, we can say that the medium version serves to dispense with the market power inquiry altogether for restraints in Category III and to "truncate" the market power inquiry for cases in Category IV.

Finally, a strong version of a limited rule of reason goes even further to lessen the burden on the government. The strong version serves to dispense altogether with a market power inquiry, not only in cases where there are no plausible procompetitive justifications, but also when procompetitive justifications are simply insubstantial. For example, a strong version can operate to place the burden on the parties to an arrangement to present substantial procompetitive justifications. If the parties fail to make such a showing, the arrangement fails the strong version without any proof of market power or marketplace effects in the individual case. Following the categorization suggested above, we could say that the strong version of a limited rule of reason eliminates the need for any market power inquiry for Cases in Category III and Category IV as well.

There is widespread support for at least the weak version of a limited rule of reason.(53) Moreover, the Supreme Court appears to have endorsed dispensing with the market power inquiry at least in cases where there are no plausible procompetitive justifications.(54) The other two versions are more controversial. The decision about which of these versions to apply has great significance for the parties because it fundamentally changes the nature of the antitrust inquiry. Not only does it determine what kind of proof of market power is required (if any), it may shift the burden of proof from the agencies to the parties on some aspects of the case. For example, the strong version, and even the medium version, can become critical to parties who find it difficult to demonstrate significant efficiency arguments for their collaboration. If they have a small market share, their arrangement will survive the weak version. However, it will not survive the strong version. The counseling and enforcement implications are quite significant.

Although the question is not free from doubt, Mass. Board appeared to endorse the weak version of a limited rule of reason. Under Mass. Board, a "full" rule of reason is necessary if the procompetitive justifications are "valid." In this context, "valid" apparently meant only that a procompetitive justification is entitled to some weight, not that it is significant.(55) Although the opinion itself did not explain how to determine if a restraint is "valid," it emphasized that the Supreme Court in NCAA and IFD dispensed with the need to show market power only after it rejected entirely the procompetitive justifications offered by the parties.(56)Thus, Mass. Board suggests that a market power inquiry is required whenever a procompetitive justification cannot be completely rejected after a factual inquiry.

In contrast to Mass. Board, the opinion in CDA, though somewhat ambiguous, appears to endorse the medium version. In analyzing restrictions on non-price advertising in CDA, the majority commented that the Commission could dispense with a "detailed" market power inquiry because "the anticompetitive effects of CDA's advertising restrictions are sufficiently clear, and the claimed efficiencies are sufficiently tenuous..."(57) At another point, the opinion commented that a "more detailed" examination of market power is necessary "[w]here the consequences of a restraint are ambiguous, or where substantial efficiencies flow from a restraint..."(58) In the case before it, the majority concluded that more limited evidence of market power was sufficient.(59) In short, CDA suggests that the market power inquiry can be truncated, though not eliminated, if significant restraints are not accompanied by substantial procompetitive justifications.

Finally, the Antitrust Division appears to endorse the strong version of a limited rule of reason. Parties to a horizontal agreement, which is not per se illegal, but which "directly limits competition on price or output" must demonstrate procompetitive justifications. Only if the parties demonstrate "significant procompetitive benefits," does the Division proceed to consider anticompetitive effects.(60) Thus, the Department does not conduct even a truncated inquiry about market power unless the parties have met this burden.

The differences in these three approaches -- the Commission's in Mass. Board, the Commission's in CDA, and the Antitrust Division's -- cannot be stated with precision. CDA and Mass. Board reached the same results on very similar facts. Moreover, words such as "plausible," "valid," and "substantial," are inherently imprecise. Nevertheless, it does seem clear that there are real differences in analysis reflected in these various agency statements.

There are certainly enforcement advantages in applying the medium or strong versions of a limited rule of reason. First, determining market definition and market shares can be a formidable undertaking. If these assessments have to made in every case, the time and expense (for the government and the parties) can be substantial. By communicating reasonably clear rules, the agencies discourage firms from coming "close to the line" and entering into agreements based on the (perhaps mistaken) assumption that they do not have market power. A limited rule of reason, if it is carefully applied, would find restraints unlawful only where the anticompetitive potential is substantial, the procompetitive justifications deserve little weight, and there is at least some evidence of harmful marketplace effects.

On the other hand, there are clearly some dangers in both the medium and strong versions. First, it is difficult to predict when they apply. The Antitrust Division's standard -- a restraint that "directly or indirectly limits competition on price or output between two or more parties" is obviously subject to dispute.(61) Just as the category of "inherently suspect" restraints could be expanded too easily, this categorization is subject to similar misuse. The procompetitive justifications in these cases will lie somewhere between "plausible" and "substantial." They must at least be plausible; otherwise, the case would be resolved at the per se stage. On the other hand, the justifications cannot be substantial, because that would push the case into a full rule of reason inquiry, which requires an examination of market power.(62) Practitioners could understandably complain about the vague scope of such a rule.

Second, both the medium and strong versions can operate to shift the burden of proof to the parties. The strong version has the more direct impact on the nature of the case since the parties must demonstrate significant procompetitive justifications before there is any showing of competitive harm. The effect of the medium version is less drastic since the medium version requires at least a "truncated" analysis of market power. In either case, there is some temptation to characterize cases in a way that invokes a limited rule of reason since it lightens the government's burden considerably. In the extreme case, the agencies could demand proof of procompetitive justifications simply because agency staff do not understand the transaction.(63) An even more troubling possibility is that private plaintiffs could shift the burden to the defendants without a showing of market power.(64)

E. Summary and Recommendations

I suggest three modifications, or at least clarifications, in current agency policy. First, the agencies should be cautious in identifying any "new" per se violation since it is easy to state the "new per se violation" too broadly.(65) In turn, there can be considerable confusion on the part of practitioners about the agencies' position. As a practical matter, most significant restraints unaccompanied by any plausible procompetitive justifications amount to a form of a restriction on price or output or a market allocation scheme, restraints which are traditional per se violations. Thus, there may not be much point in insisting that these are "new" per se violations.(66) On the other hand, there is value in communicating to practitioners and management that certain restraints are likely to be unlawful without regard to the market share of the firms involved even if they have not previously been identified as per se unlawful.

Second, the Agencies should avoid creating some arbitrary category of cases for which procompetitive justifications are simply irrelevant. After all, why would procompetitive arguments ever be "irrelevant" when the key question in each case is how competition is affected? Perhaps it is possible to read too much into the statements in CDA or in Joel Klein's recent speech, that the Commission and the Division will not consider procompetitive justifications for certain classes of restraints.(67) However, to the extent that either agency takes the position that it will not consider procompetitive justifications, this position needs to be revised. At best it leads to confusion and, at worst, it leads to mistakes in enforcement decisions by cutting off any inquiry about competitive justifications.(68) In my view, there are no enforcement or analytical costs in saying, as a general principle, that the agencies will consider procompetitive justifications in all cases of competitor collaboration.(69) Non-competitive justifications, such as the desire to discourage unionization, would simply be rejected as inconsistent with antitrust policy.(70) "Procompetitive arguments" that make no sense from an economic view would simply be rejected as implausible.(71) Flatly ruling out procompetitive arguments risks enforcement mistakes and makes the agencies' position more vulnerable in court.(72) Using the per se label sparingly and keeping the door open to procompetitive arguments strikes the best balance. It communicates the idea to practitioners that they can run afoul of the Sherman Act even if firms have a small market share. At the same time, it minimizes agency mistakes in sweeping the per se net too broadly.

Finally, the agencies need to be as precise as possible about when and what type of a limited rule of reason is appropriate. If a limited rule of reason is appropriate only when the procompetitive justifications, though plausible, turn out to deserve no weight, then the agencies have adopted the weak version described above. The "limited rule of reason" describes the inquiry about the plausibility of procompetitive justifications. That is probably a fair statement of the Court's analysis in NCAA, IFD and Professional Engineers. In all those cases, the Court examined a series of justifications (both competitive and non-competitive) and concluded that none of them were entitled to any weight. If the agencies entirely dispense with proof of market power even in cases where the procompetitive justifications do deserve some weight, they are applying the strong version. Finally, if the absence of substantial justifications serves to truncate, but not to eliminate the market power inquiry, the agencies have adopted the medium version. In my own view, the medium version best accomplishes the goals of simplifying the analysis in more clear-cut cases and minimizing the risk of enforcement mistakes. It eliminates the need for an elaborate market power assessment when the likelihood of overall competitive harm is high. At the same time, it insures that no arrangement offering some procompetitive benefit is challenged unless there is some evidence of competitive effects.

The Appendix suggests some guideline language that incorporates these suggestions.

IV. THE IDENTIFICATION AND EVALUATION OF MARKET POWER

One of the most significant contributions of the Merger Guidelines is to provide a coherent general approach to analyzing market power. Many of the same analytical principles apply to joint ventures, including the evaluation of market definition and entry barriers. However, there are also significant differences. The guidelines could make a contribution by setting out some basic approaches to evaluating market power in the joint venture context and by comparing merger and joint venture market power analysis where they differ. In addition, there could be some hypothetical cases in which market power is evaluated in different contexts. In this part of the paper, I discuss some particular issues in assessing market power, including some ways in which analyzing the market power of joint ventures differs from merger analysis. (73)

A. Coordinating Price and Output Decisions

Evaluating market power created by an agreement of competitors requires taking into account the degree to which the agreement will lead to joint, or coordinated, price and output decisions. In the case of a merger, the question of future coordination is easy since it is assumed that the combined firms will operate as a single rational actor. However, a joint venture that includes most or all the market does not create market power unless it also enables collaborators to coordinate price and output decisions.(74)

Production Joint Ventures

Production joint ventures generally present the most straightforward market power analysis, and the one that most resembles mergers. Since the collaborators must jointly determine the output of the venture, the question is how these output decisions can affect the overall market. This determination must take into account the share of the market represented by the venture itself and the degree to which competition will continue outside the scope of the venture. The concern is that the venture will create power to raise prices over competitive levels -- either on the part of the venture itself, or the venture in coordination with others. If there are no ancillary agreements that restrict competition outside the scope of the venture, the assessment is based, at least as a starting point, on the capacity or sales of the venture itself. However, under some circumstances, the capacity of the venture alone gives a misleading indication of the overall effect on market performance. To a large extent, the question becomes the degree to which the collaborators have the incentive to compete against each other and against the venture. Guidelines could be helpful in setting out factors that they agencies will consider in making this determination.(75)

Marketing and Sales Joint Ventures

Assessing the market power of a sales and marketing joint ventures can also be straightforward. A bare joint sales agreement will lead to coordination of price and output as effectively as a price-fixing cartel. However, there are a vast array of collaborations that fall into this category, ranging from barely disguised price-fixing cartels to loosely integrated collaborations that are unlikely to facilitate coordination, including information sharing programs and cooperative image advertising. While market concentration and market shares are always relevant in identifying the existence of market power, the underlying question is whether the particular type of collaboration has the potential to raise price and restrict output. Each arrangement inevitably must be evaluated based on the particular facts it presents.

Traditionally, a crucial step in this analysis has been characterizing the restraint as per se unlawful or subject to a rule of reason. Since most marketing and sales joint ventures do not have a large market share, much depends on this characterization. This determination has been based on the nature and degree of integration.(76) A significant question for the agencies is whether this integration requirement has been applied too narrowly. To the extent that particular types of integration, e.g., risk-sharing or some integration of operations, mark the lower boundary of efficiency-enhancing collaboration, these factors serve as a good proxy for "plausible" procompetitive justifications. However, the experience with the Health Care Statements suggest that there are difficulties in setting a rigid lower boundary on integration as a device for per se classifications. As I discussed in Part II, one way of mitigating this danger is to avoiding shutting the door to procompetitive justifications for any category of restraints. Another way is to base rule of reason treatment, not only on the precise degree of integration, but also on a determination of whether the collaboration has brought a new product or new capacity to the market. I discuss this possibility below.

Research and Development Joint Ventures

In the case of an R&D joint ventures, the primary concern is that the participants have eliminated the incentive to compete in innovation. If these disincentives are sufficiently strong, it is conceivable that the decreased competitive pressure to innovate will outweigh whatever economies of scale or other efficiencies are likely to be generated by the collaboration. Unlike mergers, however, there is no generally accepted relationship between concentration in a research market and unilateral or coordinated restrictions in expenditures on research.(77) The 1988 Guidelines made the simplifying assumption that there would not be a significant loss in competition in R & D if there were at least four other comparable efforts underway or there was "substantial potential for such efforts by firms or groups of firms included in the market."(78) This "five-effort" standard functioned as a type of safe harbor under the 1988 Guidelines. This standard, which is similar to the market concentration thresholds for merger enforcement, is almost certainly too demanding.(79) Coordinating research expenditures or strategies is more difficult than coordinating prices or output levels. Thus, new guidelines should suggest a new and more permissive safe harbor. In addition, there may be a tendency for courts to overstate the degree to which collaboration has eliminated equivalent efforts by individual companies. To the extent that a competitive assessment overstates the loss of potential competition, the competitive harm of the collaboration is overstated as well. I discuss this issue in the next section of the paper.

B. Identifying a Loss of Potential Competition

Potential competition doctrines are applied sparingly in the merger context. After an enthusiastic reception by the Supreme Court in the 1960's, courts have been more reluctant to conclude that an acquisition by a non-competitor harms competition by reducing "actual" or "perceived" potential competition. In the case of actual potential competition cases, in particular, the courts have tended to insist on a strong showing that merger has precluded de novo entry.(80)

Evaluating joint ventures often presents the same problem of predicting what the collaborators would do in the absence of the venture. Moreover, the predictive exercise may be even more speculative since the agency or the court has to compare actual joint action with hypothetical individual action. Nevertheless, the courts have sometimes been willing to speculate about what would happen in the absence of the collaboration. Penn-Olin itself is a good example.(81) To the extent that courts are willing to consider potential competition, the focus becomes "what the parents would have done but for the joint venture."(82)

If courts assume that collaborators will engage in the same efforts individually in the absence of the collaboration, the baseline for assessing competitive effects of the venture is competition among collaborators. However, concluding that the individual firms would make similar efforts if forced to "go it alone" places too little weight on the risk-avoiding goal of collaborative activity. The "expected value" of the collaboration is the expected gain minus the expected costs. Costs to each collaborator are reduced through economies of scale, while the expected gain is increased by the synergies of collaboration. Thus, a single firm is frequently unwilling to make the investment necessary to enter a market or solve a technological problem on its own, but it will invest a smaller amount in a collective effort with a better chance of success. If courts incorrectly assume that individual collaborators will engage in the same effort on their own, the result is to discourage collaborative investment on the mistaken theory that there will be even more investment if firms act independently.

A related problem is failing to characterize a collaborative effort as a new contribution to the market. In the case of mergers, it is usually clear that a merger represents a combination of existing capacity. Since all the assets of the firms are combined, the primary character of the transaction is clearly a consolidation of existing capacity, rather than a creation of new capacity. A joint venture that integrates existing capacity can also be treated essentially as a merger.(83) However, because joint ventures involve a "slice" of the overall production and sale process, the dividing line between existing and new capacity is ambiguous. For example, a joint R & D venture might be characterized as a new research effort, or simply a combination of existing research programs. A joint agreement to "package" copyright licenses can be viewed as fixing the prices of existing products or bringing an entirely new product to the market.(84) An insurance plan organized by a medical society that ensures enrollees that they will incur no out of pocket expense can be viewed as a new product or a horizontal agreement to fix prices.(85)

The decision to characterize a venture as bringing new capacity to the market, versus combining existing capacity, bears directly on the market power analysis. An addition to market capacity or the contribution of a new product to the market should almost always be viewed as procompetitive, even if the collaborators are coordinating price and output. Ordinarily, it is possible to determine whether the collaborators are creating new products or capacity or simply combining to sell existing ones. In the case of most production joint ventures, a court can focus on the actual change in capacity.(86) A joint venture which unambiguously creates new capacity should be presumed to be procompetitive unless the record shows clearly that the venture precluded independent entry by the collaborators. (87) A venture that creates a new product line but uses primarily existing capacity is more ambiguous.(88) In any event, the immediate effect on output or capacity is a good starting point.

Focusing only on the particular degree of integration between collaborators can result in condemning even new products or capacity under the per se rule. The Supreme Court made this mistake in Maricopa. By focusing too much on the precise degree of integration, the Court failed to give adequate weight to the fact that the medical association had introduced a new insurance plan to the Phoenix area.(89) When collaborators are bringing a new product or new capacity to the market, the rule of reason should be applied, without regard to some particular integration threshold. In the health care area, a physician network should qualify for rule of reason treatment if it is non-exclusive and can fairly be characterized as offering a new form of health insurance plan to the market. Similarly, a joint sales venture should qualify for rule of reason treatment if the parties are introducing a new product or expanding capacity. Per se treatment should be reserved for those cases when the history of the collaboration clearly shows that it is intended to eliminate independent entry.(90)

Another related concern is that expansion of joint ventures over time may be characterized as a combination of existing capacity, rather than internal expansion. The permanent loss of competition between merged firms is assumed. Post-merger growth of the merged firm, through internal expansion, is viewed unequivocally as a healthy expansion of market capacity. However, similar internal expansion of a joint venture may be viewed as harmful. As a result, a court may conclude, long after its initial formation, that the collaboration is limiting competition between the collaborators.(91) The correct treatments should depend upon whether the venture grows by investment or by adding new members. If the internal expansion characterization is more accurate, the venture should be evaluated under Section 2 standards. Otherwise, there are substantial disincentives to procompetitive efforts by the venture to gain market share.

V. EXCLUSIONARY MARKET POWER

The courts have not set out a consistent framework for evaluating when collaborators may exclude competitors from their joint activity. As a result, collaborators may find themselves caught between competing concerns. For example, in forming an R & D joint venture, management may wish to limit participation to ensure that key rivals do not benefit from any technological gains. Antitrust counsel, on the other hand, may encourage management to open participation to others in order to avoid the risk of litigation.(92) One counseling option is to find some middle ground, e.g., limiting participation but offering to license technology at reasonable rates.(93) However, the threat of litigation over excluding outsiders may undermine the incentive to collaborate in the first place. Moreover, it encourages firms not to join a risky research effort at the outset, but to wait to determine if it is successful.(94) Finally, admitting the outsider may prevent rival groups from forming similar ventures, which could result in beneficial competition in R & D.

It is likely, in fact, that the caselaw has introduced a bias toward open participation in collaborative efforts. Most private litigation involves challenges by outsiders who are seeking to participate in a successful joint venture, rather than challenges by customers or suppliers alleging that a collaborative effort is too inclusive.(95) While private litigation points toward openness, the agencies have been more sympathetic to exclusion. For example, the 1988 International Guidelines emphasized the benefit of exclusion as a means of encouraging the growth of competing collaborations.(96) To the extent a bias in the caselaw toward open participation exists, it would be useful for the agencies to clarify the circumstances under which exclusion poses actual competitive risks.

A. Exclusionary Market Power

Exclusion from collaborative activity can arise in many different ways, only some of which can fairly be characterized as an example of exclusionary market power. Competitors can exclude firms in order to police a cartel. For example, a production joint venture may threaten to exclude participants if they compete with the venture or buy inputs from outside the venture. Exclusionary policies in this context are really devices for facilitating the exercise of classic market power. The concern is that the participants already have the power to restrict output based on their current participation, not that excluding outsiders creates market power. To the extent these kinds of practices can be detected, they can be analyzed as a form of ancillary restraint.(97)

Exclusionary policies can also be used by collaborators in a vertical relationship(98) For example, parties in a vertical relationship may agree that either one or both cannot deal with outsiders. In this case, the exclusionary policies can be analyzed as a vertical exclusive agreement, which creates barriers to entry. Again, the concern is about classic market power in one or both of the markets in which the collaborators operate.

The term "exclusionary market power," on the other hand, refers to market power that is created or maintained as a result of a group's exclusion of a competitor. However, the term is somewhat ambiguous. Exclusionary market power includes cases when collaborators maintain or achieve classic market power through reducing the output of competitors by excluding them from collaborative activity.(99) A somewhat broader conception of exclusionary market power includes situations when a group without classic market power prevents more efficient outsiders from forcing prices down through participating in the collaborative activity. This latter case is significant because it includes situations when there appears to be active competition among the collaborators but the exclusion prevents the outsider from pushing prices down even further.(100) While this situation could hurt consumers, too, it is more difficult to identify. Thus, it may be appropriate to restrict enforcement to more clear-cut cases.

Anticompetitive effects do not necessarily follow from the fact that a group of collaborators has the "power" to exclude outsiders, either by excluding them from a market entirely or, in the less severe sense, by making the outsider a less effective competitor. There may be sufficient competition within the group so that admitting an outsider makes no difference. Similarly, the fact that a group has "classic market power" does not mean that an exclusion harms competition. In fact, admitting the outsider might only increase the power of the group by ensuring that the outsider now can participate in coordinated activity. In short, neither the fact that a group has the "power" to exclude an outsider from a market, nor the fact that a group has market power, means the exclusion itself harms competition. Excluding competitors is harmful only when output is reduced as a result of the exclusion. The challenge is to articulate an administrable standard for identifying those situations.

B. Standards for Assessing Exclusion

There is a good basis for the claim that the courts have not set out a clear framework for evaluating exclusion. At various times, the courts have articulated a virtual per se rule, some version of an essential facility doctrine, a qualified per se rule, and a highly generalized rule of reason. The FTC Competition Policy staff report identified three general approaches: 1) the essential facility doctrine; 2) an evaluation of the purpose of the exclusion; and 3) a balancing of the overall competitive effects of exclusion.(101) Guidelines could be helpful in stating the circumstances under which one or more of these standards is appropriate, and in articulating how the agencies will apply them.

The lower courts today frequently rely on the Supreme Court's comments in Northwest Stationers.(102) There the Court found that a buying cooperative organized by retailers should be analyzed under a rule of reason when it excluded a wholesaler. The Court obviously wished to make the point that dictum in earlier cases, which suggested that all boycotts are per se unlawful, is no longer the law.(103) However, the Court's comments about when, if ever, a per se rule would be appropriate, were ambiguous. Some language in the opinion suggests that an exclusion is per se unlawful if the cooperative possesses market power or the outsider is denied access to an essential input.(104) Other language suggests that either one of these factors is a necessary, but not sufficient, element of the plaintiff's case.(105) As a result of this ambiguity, the lower courts have applied Northwest Stationers differently. Some courts have said that either element is sufficient to establish a per se violation.(106) Others have said that at least one element is necessary, but not sufficient to make out a violation.(107) As a practical matter, neither requirement is usually met.(108)

Finding exclusion to be per se illegal simply because the group constitutes a dominant share of the market is almost surely wrong. If the excluding group already has market power, requiring it to admit an outsider may only increase market power, rather than dissipate it. The question should be whether the exclusion operates so as to preserve or achieve market power, not simply whether the group has it. The essentiality prong of Northwest Stationers presents its own analytical difficulties. First, determining "essentiality" has proven to be difficult whether the defendant is a single actor or a group.(109) An "essential" input lies along a spectrum from one that is absolutely required to one that reduces the costs to the outsider compared to alternatives. Second, even if a facility is viewed as essential, excluding an outsider may not harm competition if there is sufficient competition among the collaborators or other competitors have found an adequate substitute for the resource controlled by the collaborators.(110) Finally, even if these other requirements are met, the justifications for the exclusion might outweigh the competitive benefit of requiring the group to provide access to the outsider.

Alternatives to the essential facility doctrine also present difficulties. Finding a violation based on the purpose of the exclusion can present significant proof problems. As in predatory pricing cases, an intemperate internal memo can suggest some intent to harm competition when a restraint on participation has no real effect in the market. Conversely, an exclusion may preserve market power when no anticompetitive purpose exists (or at least when it cannot be proved). In short, purpose alone should be probative but not sufficient to show an exclusion harms competition. A standard based on overall balancing of effects, most directly states the goal of Section 1. However, balancing effects, particularly in close cases, is often exceedingly difficult. In addition, such a general standard provides the least guidance to practitioners. Thus, there is a need for a more specific framework.

A Framework for Evaluating Exclusions

The basic inquiry is whether an exclusion has a significant effect by creating or maintaining market power. Typically, the excluding group prevents an outsider from obtaining access to an input.(111) The outsider claims that the exclusion prevents it from competing in the output market. For example, a railroad may claim that it is denied use of a terminal,(112) a newspaper may claim that is denied access to news reports,(113) or a seller of consumer products may claim that is denied a certification of seal of approval.(114) Thus, as a general matter, an exclusion can be analyzed much like a vertical exclusive dealing agreement, which may create barriers to entry. Instead of the exclusivity running between two firms, it runs between the same group of firms operating in two different markets. This distinction does raise different considerations in analysis. However, the competitive effects are similar.

It follows from this perspective that a plaintiff challenging exclusion from a joint venture should first show that the collaborators have market power in the input market. If the venture does not have market power in the input market, the outsider has nothing to complain about. Substitutes are available at competitive prices. One way of describing an "essential facility," then, is to say that it simply identifies an input market in which one or more firms have market power.(115)

Second, the outsider should have to establish that the group has market power in the output market. If the output market is competitive, consumers are not hurt by the lack of participation by the outsider. Usually, the outsider should have to show that the group has a dominant share of the market, or that the market is sufficiently concentrated that competition between the group and other firms or groups in the market is inadequate.(116) If there is vigorous competition within the group, or between the group and others, the exclusion is not harmful. (117)

Third, the outsider should have to show that the exclusion contributes to the creation or maintenance of the market power (or, to put it another way, that if the outsider is allowed to participate, market power would be dissipated). To some extent, this inquiry overlaps with the first two. If the outsider has access at reasonable cost to a substitute for the resource controlled by the group, the exclusion is not contributing to the exercise of market power in the input market. If allowing the outsider to have access to the input will simply facilitate the exercise of classic market power in the input market, the exclusion is operating to promote competition, not to inhibit it.(118) Similarly, allowing the outsider to have access to the input will only promote competition if there is an assurance that the collaborators will compete in the output market.(119) The question is -- what does the outsider bring to the output market if it is granted access to the input? Determining whether admitting the outsider dissipates market power or reinforces it depends on the nature of the collaboration and the nature of the outsider.(120)

Finally, the group should be able to offer procompetitive justifications for the exclusion. This inquiry resembles the analysis of justifications for vertical restraints. Are there efficiency justifications for the collaborators restricting access to the input? If granting access to outsiders undermines the point of the collaboration, then the justifications are very strong. At the time the case is brought, eliminating the exclusion may not appear to undermine the efficiency of the collaboration. Granting access to the input may not increase costs or harm the collaborators other than in the (desirable) sense that they will face more competition in the output market.

However, often the justification is for the exclusion is based on the original risk-taking investment by the collaborators. If the group has continued to admit newcomers who are not a threat to their market power, then this history undercuts the argument the exclusion undermines the incentives to collaborate.(121) For example, the group may benefit from expanding the membership due to economies of scale on both the supply side and the demand side.(122) In that case, the exclusion is more likely to be aimed at keeping a maverick out of the market.(123)

When all these requirements are met and the justifications for exclusion are insubstantial, the agencies can safely conclude that the exclusion harms competition. When some or all of these requirements are not clearly met, there is still an argument that exclusion harms competition, but the case is much harder to evaluate. For example, even though the collaborators have some degree of market power over the input, the outsider may still be able to have access to a substitute at somewhat higher costs. Alternatively, the collaborators may compete among themselves, but the outsider claims that it can use the input more efficiently and force prices down even further.(124) In most cases, these claims are simply too speculative to evaluate with any confidence.

Special Cases: Standard Setting

There are special cases when these requirements could be relaxed. For example, in the standard setting context, the claim is often that the standard setting-body has arbitrarily denied the outsider a certification or "seal of approval," or perhaps, has denied consideration of a revised product standard proposed by the outsider. In this case, the input market is the standardizing and certifying process itself. The outsider may be able to show that the certifying group has power in this market, but it may not be able to show that eliminating the exclusion will promote competition in the output market. For example, the output market may be unconcentrated and apparently competitive. However, standard setting presents some special considerations that warrants a presumption in certain cases that an exclusion does harm competition. First, if the outsider is excluded because it is proposing a revised standard, it can play the role of the maverick, disrupting established competitive patterns and spurring members of the group to innovate themselves.(125) If the outsider is arbitrarily excluded from a certification, the group is subverting its own professed goal of certifying all qualified firms.(126) Second, the justifications for exclusion from a standard-setting or certifying collaborative effort are less persuasive than in the case of other collaborations. For example, there is less fear that the outsider can free ride on risky investments taken by the group. Thus, the outsider should only have to show that the group has market power in the certifying and standard-setting market, and that the group has arbitrarily or discriminatorily excluded the outsider.(127)

APPENDIX

1. Professor of Law, Northeastern University School of Law; currently, Scholar in Residence, Federal Trade Commission

2. See Joseph F. Brodley, Joint Ventures and Antitrust Policy, 95 Harv. L. Rev. 1521, 1524 (1982).

3. See Hovenkamp, FEDERAL ANTITRUST POLICY 185-86 (1994).

4. See Brodley, supra note 1, at 1524-25; Robert Pitofsky, Joint Ventures Under the Antitrust Laws: Some Reflections on the Significance of Penn-Olin, 82 Harv. L. Rev. 1007, 1007 (1969); Phillip Areeda & Louis Kaplow, ANTITRUST ANALYSIS 270 (4th ed. 1988).

5. Even firms in a vertical relationship can be said to have created a joint venture when they work closely in developing specifications for products to be sold by one to the other. See Statement of Joseph Kattan Before FTC 2-3 (June 5, 1997).

6. If the parties take a sufficient equity interest in each other, they would be viewed as having become a single economic actor. See Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752 (1984) (parent and wholly owned subsidiary should be treated as single actor). Even if the participants to a joint venture remain independent companies, some courts treat joint ventures as single actors if they appear to have organized themselves so as to pursue common, rather than divergent, interests. See, e.g., Siegel Transfer, Inc. v. Carrier Express, Inc., 54 F.3d 1125, 1127 (3d Cir. 1995); City of Mt. Pleasant v. Associated Electric Cooperative, 838 F.2d 268, 276-77 (8th Cir. 1988).

7. The phrase is not original with me. Several persons testifying before the Commission suggested this terminology.

8. U.S. Dep't of Justice, Antitrust Enforcement Guidelines for International Operations, 4 Trade Rep. (CCH) para. 13,109 (1988) [hereinafter 1988 International Guidelines].

9. U.S. Dep't of Justice and FTC, Antitrust Enforcement Guidelines for International Operations (1995) [hereinafter 1995 International Guidelines].

10. See U.S. Dep't of Justice and FTC, Antitrust Guidelines for the Licensing of Intellectual Property Part 3.4 [hereinafter Intellectual Property Guidelines], 6 Trade Reg. Rep. (CCH) para. 13,132 (April 6, 1995).

11. Recent cases with extensive discussions of the relationship between the per se rule and the rule of reason include National Society of Professional Engineers v. United States, 435 U.S. 679 (1978); Broadcast Music, Inc. v. Columbia Broadcasting Sys.; Inc., 441 U.S. 1 (1979); NCAA v. Board of Regents, 468 U.S. 85 (1984); and FTC v. Indiana Federation of Dentists, 476 U.S. 447 (1986).

12. Massachusetts Board of Registrations in Optometry, 110 F.T.C. 549 (1988); 1988 FTC Lexis 34 [hereinafter Mass. Board].

13. FTC v. California Dental Association, 1996 FTC Lexis 88; aff'd, California Dental Assoc. v. FTC, 1997-2 Trade Cas. (CCH) para. 71,954; 1997 U.S. App. Lexis 28882 (Oct. 22, 1997) [hereinafter CDA].

14. Joel I. Klein, A Stepwise Approach to Antitrust Review of Horizontal Agreements, Address Before the ABA Antitrust Section (Nov. 7, 1996) [hereinafter, DOJ's Stepwise Approach].

15. Part 3.41.

16. Part 3.4.

17. U.S. Dep't of Justice & FTC, Statements of Antitrust Enforcement Policy in Health Care (1996) [hereinafter Health Care Statements], reprinted in 4 Trade Reg. Rep. para. 13,153. The Health Care Statements are intended to treat health care joint ventures no more strictly or more leniently than joint ventures in other industries. Id. at 20,800.

18. Congress itself mandated a form of rule of reason for certain joint ventures in the National Cooperative Research and Production Act ("NCRPA"). 15 U.S.C. secs. 4301-4305.

19. The per se rule on boycotts now appears limited to boycotts of customers or suppliers, intended to drive competitors of the boycotters out of the market See FTC v. Indiana Federation of Dentists, 476 U.S. 447, 458 (1986). Fashion Originators' Guild of America, Inc. v. F.T.C., 312 U.S. 447 (1941), and Eastern States Retail Lumber Dealers' Ass'n v. United States, 234 U.S. 600 (1914), are examples.

20. See National Society of Professional Engineers v. United States, 435 U.S. 679 (1978) (agreement among engineers not to engage in competitive bidding); Catalano, Inc. v. Target Sales, Inc., 446 U.S. 643 (1980) (agreement among wholesalers not to offer credit-free sales).

21. See United States v. General Electric Co., No. CV96-121-M (D. Mont. filed Oct. 28, 1996).

22. There might well be differences as to which restraints are appropriate for criminal prosecution, but that issue is outside the scope of this discussion.

23. The quoted language is from Northern Pac. Ry. Co. v. United States, 356 U.S. 1, 5 (1958).

24. The Court made the point that the literal form of the agreement does not prevent a finding of per se illegality in Timken Roller Bearing Co. v. United States, 378 U.S. 158, 170 (1964).

25. The importance of not applying the per se rule based on the "literal" form of the agreement appear was made fifteen years after Timken in Broadcast Music, Inc. v. Columbia Broadcasting Sys., Inc., 441 U.S. 1 (1979).

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

27. See Jay Palmer v. BRG of Georgia, 498 U.S. 46 , 49(1990).

28. The territorial allocation scheme in Palmer was unrelated to any plausible goal of promoting interbrand competition. Two bar reviews, which had vigorously competed in Georgia, agreed that one would not operate in the state and the other would not operate outside the state. As a result, bar review prices rose from $150 to over $400. See id. at 46-49.

29. Market allocations schemes are the most obvious example. In addition to Topco, see United States v. Sealy, Inc., 388 U.S. 350 (1967). Some might put the price-fixing arrangement struck down in Arizona v. Maricopa County Medical Society, 457 U.S. 332 (1982), in this category as well.

30. 1988 FTC Lexis 34 at 10-13.

31. Id. at 13. The opinion mentioned price-fixing and market divisions as examples. Id.

32. A "plausible" justification was defined to mean that one that "cannot be rejected without extensive factual inquiry." Id.

33. See id. at 13-14.

34. 1996 FTC Lexis 88 at 26. The Commission read Professional Engineers to say that the per se rule applies to "agreements whose nature and necessary effect are so plainly anticompetitive that no elaborate study of the industry is needed to establish their illegality" and that the rule of reason applies to "agreements whose competitive effect can only be evaluated by analyzing the facts peculiar to the business, the history of the restraint, and the reasons why it was imposed." Id. (quoting Professional Engineers, 435 U.S. at 691-92).

35. "We use the term 'rule of reason' when speaking about the individualized analysis, in contradistinction to the categorical per se approach." Id.

36. 1996 FTC Lexis 88 at 29 ("We ... believe it to be well grounded in ... experience and precedent to strip CDA's price advertising restrictions of their professional garb and declare them per se unlawful as naked restraints on price competition.").

37. Id. at 31-32.

38. Id.

39. Id. at 32-33. The Commission, nevertheless, concluded that the association did have market power. Id.

40. 1988 FTC Lexis 34 at 13.

41. See, for example, id. at 11, where the Commission discussed Blalock v. Ladies Professional Golf Ass'n, 359 F. Supp. 1260 (N.D. Ga. 1973). In that case, a district court found that the suspension of a golfer by a professional association for moving her ball was per se illegal as a group boycott!

42. 1988 FTC Lexis at 13.

43. See id. at 30 n.13 ("[W]e have been open to arguments that might carry weight under Broadcast Music, but CDA has simply failed to assert the requisite competitive benefits that might save it from per se condemnation.").

44. See id. ("The view that the Commission's reasoning foreshadows summary condemnation for a vast array of future cases...overstates our conclusion. Only cases involving equivalent conduct will be accorded similar treatment in the future.").

45. See United States v. General Electric Co., No. CV96-121-M (D. Mont. filed Oct. 28, 1996). GE licensed diagnostic software to hospitals that purchased GE's medical equipment and used the software to service the equipment. The hospitals competed with GE in servicing other hospitals' equipment. As a condition of the licensing agreement, GE insisted that the hospitals could not compete with it in servicing any other equipment regardless of brand. The Division took the position that the agreement should be seen as a "naked agreement not to compete in the medical services market.". The case is discussed in DOJ's Stepwise Approach, supra note ___ .

46. Id. Thus, under both agencies' approach, there will continue to be an evolution of the per se rule, expanding to cover certain arrangements (and perhaps contracting as well).

47. See id. ("[I]n the next step of our analysis, if we conclude that a horizontal agreement that directly limits competition on price or output between competitors is not per se illegal, we then inquire whether there's a procompetitive justification for the agreement.")

48. Id. ("This [per se] category appears to be reasonably well-defined and usually the sole question we face in deciding if a particular agreement fits within it is whether, despite the effort to make it seem like the agreement is ancillary to some arrangement, it in fact isn't.")

49. The Supreme Court has sometimes said that procompetitive justifications are irrelevant in the case of per se offenses. See, for example, Justice Stevens' comments in Arizona v. Maricopa County Medical Society, 457 U.S. 332, 351 (1982) ("The anticompetitive potential inherent in all price-fixing agreements justifies their facial invalidation even if procompetitive justifications are offered for some.") (citing United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 226 n.59 (1940)) However, this is the precise proposition that was rejected in Broadcast Music, Inc. v. Columbia Broadcasting Sys., Inc., 441 U.S. 1 (1979).

50. Ironically, perhaps, the Antitrust Division has stressed that it wishes to avoid placing too much significance on characterization. See id. ("[W]e reject the notion that there should be two methods of analysis -- per se or full-blown rule of reason market analysis. As a matter of both sound and efficient antitrust analysis, we think this dichotomy is too stark and, frankly, that it leads to far too much of a front-end emphasis on which approach to apply, a choice that can sometimes be outcome determinative.")

51. I use the term "limited rule of reason" because the suggested terminology varies widely and there is no consensus about what any of the terms mean. Other descriptions of a limited rule of reason include "presumptive illegality," "strict antitrust scrutiny," "hard-boiled rule of reason," "truncated rule of reason," and "quick look." See, e.g., Richard Steuer, Indiana Federation of Dentists: The Per Se-Rule of Reason Continuum, 8 Cardozo L.Rev. 1101 (1987); Donald L. Beschle, "What Never? Well, Hardly Ever": Strict Antitrust Scrutiny As An Alternative to Per Se Illegality, 38 Hastings L.J. 471 (1987); Timothy Muris, The New Rule of Reason, 57 Antitrust L.J. 859, 864 (1988); Phillip Areeda, The Changing Contours of the Per Se Rule, 54 Antitrust L.J. 27 (1985).

52. In both NCAA v. Board of Regents, 468 U.S. 85, 109 (1984), and FTC v. Indiana Federation of Dentists, 476 U.S. 447, 460 (1986), the Court commented that the absence of market power does not justify a "naked restriction on price or output."

53. Several commentators advocate dispensing with the market power inquiry for at least cases in Categories II and III. See Muris, supra note ___, at 864; Areeda, supra note ___; Statement of Ernest Gellhorn Before the FTC 17 (June 20, 1997) ("At this preliminary state, the issue is narrowly limited to whether there is a legitimate basis for concluding that the collaboration is likely to produce market benefits -- not whether its benefits outweigh its costs, whether the collaboration constrains entry or output, or whether it magnifies market concentration and imperils competition.").

54. In both NCAA v. Board of Regents, 468 U.S. 85 (1984), and FTC v. Indiana Federation of Dentists, 476 U.S. 447 (1986), the Court said that "naked" restraints can be found unlawful in a rule of reason case without proof of market power. However, both opinions then went on to discuss marketplace effects. Thus, both statements are dicta.

55. Mass. Board distinguished between "plausible" and "valid" efficiency justifications. A "plausible" efficiency justification was simply one that is rational on its face, that is, it "cannot be rejected without extensive factual inquiry." 1988 FTC Lexis 34 at 12. The meaning of a "valid" justification is less clear. The opinion concluded that there was not a "plausible" procompetitive justification for a restriction on price advertising and that the justifications offered for a ban on certain forms of non-price advertising were not "legally plausible." Id. at 21-22. Thus, the opinion did not explain how a restraint would be determined to be "valid."

56. Id. at 9-10.

57. 1996 FTC Lexis 88 at 32-33.

58. Id. at 40.

59. See id. at 40. The majority relied on the facts that the association could police and enforce their prohibition and that the market did not force the dentists to advertise even consumers desired additional information. Id. at 40-42. In addition, the Commission cited evidence of entry barriers into dentistry in California and the fact that, even though some dentists wished to advertise, they preferred to remain in the organization. Id. at 48-49. See also the majority's discussion of how CDA's "quick look" approach would have led to the same results in Mass. Board by focusing on some indicators of marketplace effects. Id. at 61-62.

60. See DOJ's Stepwise Approach, supra note ___. ("[I]f we conclude that a horizontal agreement that directly limits competition on price or output is not per se illegal, we then inquire whether there's a procompetitive justification for the agreement....[W]e expect a response that doesn't merely speculate about the existence of efficiencies, but rather comes forward with real-world evidence -- factual evidence, expert economic evidence, and preferably both...And if we find that the proffered procompetitive justifications are unsubstantiated, we conclude that the agreement should be struck down. On the other hand, if we find that there are significant procompetitive benefits..., we then...seek to determine whether its likely anticompetitive effects outweigh its procompetitive benefits.").

61. An example cited by the Department is a joint selling arrangement between two radio stations, under which one of the stations agreed to pay the other station a fixed price in return for which the second station could sell all of the first station's advertising and keep the revenue. The Department concluded that there might be procompetitive justifications, but the parties were not able to offer any. See DOJ's Stepwise Approach, supra note __.

62. The Intellectual Property Guidelines use different, but perhaps equally ambiguous, language. See Part 3.4 ("[The rule of reason] inquiry may be truncated in certain circumstances. If the Agencies conclude that a restraint has no likely anticompetitive effects, they will treat it as reasonable, without an elaborate analysis of market power or the justifications for the restraint. Similarly, if a restraint facially appears to be of a kind that would always or almost always tend to reduce output or increase prices and the restraint is not reasonably related to efficiencies, the Agencies will challenge the restraint without an elaborate analysis of particular industry circumstances.") (emphasis added) (footnotes omitted).

63. See Gellhorn, supra note ___, at 3-4 (arguing against a requirement that there must be financial or operational integration to avoid application of the per se rule because integration may be a response to other legal requirements or exogenous effects, such as tax consequences).

64. In that case, however, a court might be able to dismiss the case for failure to show antitrust injury.

65. The Supreme Court, of course, has a long tradition of stating overly broad per se rules, that then must be narrowed in subsequent cases. The per se rule against price-fixing described in United States v. Socony-Vacuum Oil Co., 310 U.S. 150 (1940); the per se rule against boycotts described in United States v. General Motors Corp., 384 U.S. 127 (1966); and the per se rule against market allocations described in United States v. Topco Associates, Inc., 405 U.S. 596 (1972), are good examples.

66. Courts are hesitant to accept new "per se" violations. See, e.g., California Dental Assoc. v. FTC, 1997-2 Trade Cas. (CCH) para. 71,954; 1997 U.S. App. Lexis 28882, at 17-18 (Oct. 22, 1997) (rejecting the Commission's conclusion that price advertising restrictions are per se unlawful).

67. See supra notes __ and accompanying text.

68. The Department's treatment of the licensing restriction discussed above may be an example. See supra note ___.

69. The fact that parties are "allowed" to offer a justification does not preclude criminal enforcement where the arguments are spurious. As long as a jury concludes that there was a purpose to restrict competition, the mens rea requirements for criminal conviction are met. See United States v. United States Gypsum Co., 438 U.S. 422 (1978).

70. For example, a desire to discourage unionization does not change the conclusion that a restriction on hours of operation is per se unlawful See Detroit Auto Dealers Association, 111 F.T.C. 417 (1989), aff'd in part, 955 F.2d 457 (6th Cir. 1992), cert. denied, 113 S. Ct. 461 (1992).

71. One might put in this category, for example, the claimed justifications by the beer wholesalers in Catalano. There, the argument was that an agreement not to provide interest-free credit to customers promoted competition by encouraging entry. The court of appeals was persuaded by this argument, though the Supreme Court was not. See Catalano, Inc. v. Target Sales, Inc., 446 U.S. 643, 648-50 (1980).

72. In Detroit Auto Dealers Association, 111 F.T.C. 417 (1989), aff'd in part, 955 F.2d 457 (6th Cir. 1992), cert. denied, 113 S. Ct. 461 (1992), the majority was troubled by what it concluded was a "per se" approach by the Commission in finding an agreement to restrict hours of operations to violate the FTC Act. See 955 F.2d at 471. Nevertheless, it affirmed the Commission's finding of an unreasonable restraint of trade, in part because the Commission was able to show harmful competitive effects. Id. 49-50 and 49 n.15. Another member of the panel, who would have affirmed the Commission's findings in their entirety, stressed that the Commission was not applying the per se rule. ("[T]he FTC ...did not use a per se analysis. Under a per se analysis, the agreement would have been invalid without any consideration of its procompetitive effects. The FTC, however, did consider the efficiency justifications offered by respondents....") Id. at 475. It seems likely all three members of the panel would have been troubled if the Commission had flatly declared procompetitive justifications to be irrelevant. Similarly, the Ninth Circuit affirmed the Commission's decision in CDA only after rejecting the Commission's per se characterization and relying on the Commission's finding of anticompetitive effects. See California Dental Assoc. v. FTC, 1997-2 Trade Cas. (CCH) para. 71,954; 1997 U.S. App. Lexis 28882 at 26-27 (Oct. 22, 1997).

73. In this section, I discuss the power to price over competitive levels obtained through tacit or express collusion, rather than the power obtained by excluding competitors or raising their costs. The issue of exclusionary market power is discussed in the next part of the paper.

74. See 1988 International Guidelines, Part. 3.3 ("[A joint venture] may be unlikely to facilitate the exercise of market power in circumstances [where a] complete merger would be likely to do so.").

75. There are several possible considerations. For example, special provisions to limit information exchange may reduce the likelihood of spillover effects; the incentive to compete may increase as the number of collaborators increase, see Pitofsky, supra note __, at 1012; the incentive to compete may decrease if the venture is large compared to individual capacity, see Edmund W. Kitch, The Antitrust Economics of Joint Ventures, 54 Antitrust L.J. 957, 962 (1987); and the likelihood of competition may decrease to the extent that parents continue to have control over the venture's price and output decision, Timothy F. Bresnahan and Steven C. Salop, Quantifying the Competitive Effects of Production Joint Ventures, 4 Intl. J. Ind. Org. 155, 156 (1986). These issues are discussed generally in Michael S. McFalls, The Role of Classical Market Power in Joint Venture Analysis 31-33 (FTC Staff Discussion Draft) (1997).

76. See American Bar Assoc., Antitrust Law Developments (Fourth) 395 (1997) ("Ordinarily, the threshold issue with respect to a joint venture involving actual or potential competitors is whether it involves a significant integration of the economic resources of the parties to escape condemnation as a per se unlawful cartel....").

77. Some literature suggests that a monopolist will not underinvest in research. See 1988 International Guidelines 53 n.236.

78. Id. at 53.

79. A five effort rule, in which all of the ventures are the same size, would lead to an HHI market concentration of 2,000, close to the 1800 threshold for highly concentrated markets. See Merger Guidelines, Part 1.51.

80. See, e.g., BOC Int'l Ltd. (British Oxygen) v. FTC, 557 F.2d 24 (2d Cir. 1977) (government must prove that entry is sufficiently probable and imminent); United States v. Siemens Corp., 621 F.2d 299 (2d. Cir. 1980); FTC v. Atlantic Richfield Co., 549 F.2d 289 (4th Cir. 1980).

81. See United States v. Penn-Olin Chem. Co., 378 U.S. 158 (1964). In Penn-Olin, two firms formed a joint venture to establish a new plant. The new venture increased the capacity in a highly concentrated market by about 50%. The district court had found that there would be a violation of Section 7 only if the government could show that both firms would have entered the market independently. The Court remanded for a determination whether one of the collaborators would have entered the market on its own, while the other collaborator remained "at the edge of the market," exerting a health procompetitive influence on market behavior. Id. at 173.

82. Pitofsky, A Framework for Antitrust Analysis of Joint Ventures, 54 Antitrust L.J. 893, 897 (1985).

83. See, e.g., United States v. Ivaco, 704 F. Supp. 1409, 1414 (W.D. Mich. 1989).

84. See Broadcast Music, Inc. v. Columbia Broadcasting Sys., Inc., 441 U.S. 1 (1979), where the Court viewed package licenses as new products.

85. See Arizona v. Maricopa County Medical Society, 457 U.S. 332 (1982), where the Court viewed a physician-organized health insurance plan as simply price-fixing.

86. See, e.g., United States v. Columbia Picture Industries, Inc., 507 F. Supp. 412, 430-31 (S.D.N.Y. 1980), aff'd mem., 659 F.2d 1063 (2d. Cir. 1981), where the court focused on the fact that a joint venture among producers to sell movies involved existing movies not new ones.

87. Penn-Olin involved a clear and substantial addition to capacity. See supra note __. On remand, the district court found that neither firm was reasonably likely to enter the market, so eventually the right result was reached. See United States v. Penn-Olin Chemical Co., 246 F. Supp. 917, 928 (D. Del. 1965), aff'd per curiam, 389 U.S. 308 (1967). Nevertheless, the prospect that such a significant addition to market capacity could be challenged can give pause to parties planning similar efforts.

88. The finding of a loss of potential competition in Yamaha Motor Co., Ltd. v. FTC, 675 F.2d 971 (8th Cir. 1981), cert. denied, 456 U.S. 915 (1982), was more defensible than Penn-Olin. Brunswick, a domestic manufacturer of outboard motors entered into a joint venture with Yamaha, a Japanese manufacturer of outboard motors. Brunswick contributed capital and technology and the venture created a motor designed especially to compete in the U.S. See 1979 FTC Lexis 107 at 208. However, the venture used primarily existing capacity owned by Yamaha. Id. at 980 Moreover, there was substantial evidence that the venture precluded independent entry by Yamaha into the United States market. Id.

89. See 457 U.S. at 353-53 (rejecting argument that the plan represented a new addition to the market).

90. Jay Palmer v. BRG of Georgia, 498 U.S. 46 (1990), is an example. Two bar reviews agreed that one would no longer compete within the state, while the other agreed that it would not enter other states.

91. One suggestion is that, after an initial antitrust review at formation, joint ventures should generally been judged as single actors. Gellhorn, supra note __, at 11-12.

92. See Statement of Robert Skitol Before the FTC 2-5 (June 5, 1997).

93. Id.

94. See Hovenkamp, FEDERAL ANTITRUST POLICY 207 (1994).

95. There are several Supreme Court cases dealing with a claim by excluded outsiders. See, e.g., United States v. Terminal Railroad Ass'n, 224 U.S. 383 (1912); Associated Press v. United States, 326 U.S. 1 (1945); Northwest Wholesale Stationers v. Pacific Stationery & Printing Co., 472 U.S. 284 (1985); Radiant Burners, Inc. v. People's Gas Light & Coke Co., 364 U.S. 656 (1961); Allied Tube & Conduit Corp. v. Indian Head, Inc., 468 U.S. 492 (1988). See also Statement of Steven C. Bomse Before the FTC 22 (1997).

96. Part 3.42 ("[T]he Department in general is concerned about the anticompetitive effects of a joint venture when it is overinclusive -- that is, when it restricts competition among competitors that account for a large portion of sales (or capacity) in the market, thereby creating, enhancing or facilitating the exercise of market power.") A recent FTC Staff Report is somewhat more ambivalent, at least with regard to network joint ventures. See FTC Staff Report, Competition Policy in the New High-Tech, Global Marketplace 9-29 (1996) [hereinafter Competition Policy] ("[W]e find that a denial of membership in a network joint venture or of access to a standard may enhance the incentives to establish a competing venture or to develop a new standard, while maximizing the reward to incumbents for creating and developing the existing venture or standard...On the other hand, we find that mandating access to membership or to a standard may increase competition within the joint venture or among firms who make use of the established standard.").

97. In these cases, the role of the exclusionary policy (e.g., a group's "by-law" or contractual provision) is likely to be facially ambiguous. The purpose and effect of the policy must be determined from its historical use.

98. See, e.g., United States v. MCI Communications, Corp., 1994-2 Trade Cas. para. 70,730 (D.D.C. 1994), where the Antitrust Division challenged a joint venture involving MCI and British Telecommunications.

99. See Dennis W. Carlton and Steven C. Salop, You Keep on Knocking but You Can't Come in: Evaluating Restrictions on Access to Joint Ventures, 9 Harv. J. of Law & Technology 319, 330 (1996) ( "A joint venture access rule can harm competition by raising the costs or otherwise disadvantaging rivals of its members in the output market. This can reduce competition and lead to higher prices than would otherwise occur.").

100. See id.

101.Competition Policy, supra note ___, at 9-6 to 9-7.

102. Northwest Wholesale Stationers v. Pacific Stationery & Printing Co., 472 U.S. 284 (1985).

103. See id. at 294.

104. Id. at 296 ("Unless the cooperative possesses market power or exclusive access to an element essential to effective competition, the conclusion that expulsion is virtually always likely to have an anticompetitive effect is not warranted... Absent such a showing with respect to a cooperative buying arrangement, courts should apply a rule of reason analysis.")

105. See id. at 298 ("A plaintiff seeking application of the per se rule must present a threshold case that the challenged activity falls into a category likely to have a predominant anticompetitive effect. The mere allegation of a concerted refusal to deal does not suffice because not all concerted refusals to deal are predominately anticompetitive. When the plaintiff challenges expulsion from a joint buying cooperative, some showing must be made that the cooperate possesses market power or unique access to a business element necessary for effective competition.") In FTC v. Indiana Federation of Dentists, 476 U.S. 447, 458 (1986), the Court characterized the breadth of the per se rule stated in Northwest Stationers as a narrow one. ("As we observed [in Northwest Stationers], the category of restraints classed as group boycotts is not to be expanded indiscriminately and the per se approach has generally been limited to cases in which firms with market power boycott suppliers or customers in order to discourage them from doing business with a competitor...").

106. See, e.g., Wigod v. Chicago Mercantile Exch., 981 F.2d 1510, 1517 (7th Cir. 1992); Bascom Food Prods. Corp. v. Reese Finer Foods, Inc., 715 F. Supp. 616, 633-34 (D. N.J. 1989).

107. See, e.g., Hahn v. Oregon Physicians' Serv., 868 F. 2d 1022, 1030 n.9 (9th Cir. 1988); Collins v. Associated Pathologists, 844 F.2d 473 (7th Cir 1988), cert. denied, 488 U.S. 852 (1988) ("[B]oycotts are illegal per se only if used to enforce agreements that are themselves illegal per se -- for example, price-fixing agreements."); Carleton v. Vermont Dairy Herd Improvement Ass'n, 782 F. Supp. 926, 933 (D. Vt. 1991).

108. See American Bar Assoc., Antitrust Law Developments (Fourth) 426 (1997).

109. In the context of denial of a facility by a monopolist, an influential formulation of the doctrine stated that it applies only if: 1) there is control of the facility by the monopolist; 2) the competitor is unable to reasonably duplicate the facility; 3) the competitor is denied use of the facility; and 4) it is feasible to provide access to the facility. See MCI Communications Corp. v. AT&T, 708 F.2d 1081, 1132-33 (7th Cir. 1983), cert. denied, 464 U.S. 891 (1983). The doctrine has been frequently criticized. See Hovenkamp, supra note ___, at 273 ("The so-called 'essential facility doctrine' is one of the most troublesome, incoherent and unmanageable of bases for Sherman section 2 liability. The antitrust world would almost certainly be a better place if it were jettisoned...."); Phillip Areeda, Essential Facilities: An Epithet in Need of Limiting Principles, 58 Antitrust L.J. 841 (1990).

110. In that case, the particular facility controlled by one group is essential to the plaintiff, but not to others.

111. See Carlton and Salop, supra note ___, at 328-29. This discussion borrow from Carlton and Salop but the framework and terminology suggested here are somewhat different.

112. United States v. Terminal Railroad Ass'n, 224 U.S. 383 (1912).

113. Associated Press v. United States, 326 U.S. 1 (1945).

114. Radiant Burners, Inc. v. People's Gas Light & Coke Co., 364 U.S. 656 (1961).

115. See Hovenkamp, supra note __, at 274.

116. However, market power in the output market may be based on the fact that individual members have market power by virtue of their participation in the collaboration in the input market. In Associated Press, the market power in the output market, sale of newspapers, was held by individual newspapers. 326 U.S. at 17-18.

117. There may be special cases in markets for differentiated products, where the outsider a offers close substitute for the products produced by the group. Even though the excluding group does not have large market share, preventing the outsider from competing may allow the group to raise prices to some extent. See Merger Guidelines, Part 2.21; Hovenkamp, supra note __, at 212-13.

118.For example, excluding the outsider may simply mean that the outsider must enter the input market on its own or collaborate with other outsiders to enter.

119.In both Terminal Railroad and Associated Press, the Court assumed that admitting outsiders would promote competition. In Terminal Railroad, the Court required that the joint venture of railroads, which owned a terminal and transfer facilities, make them available to any other railroads on non-discriminatory terms, even if these new users did not elect to become owners. The Court assumed that the large number of railroads that could use the facility, including non-owners of the terminal, would result in competition. In Associated Press, the Court required a joint venture of newspapers, which gathered news and distributed it to all members, had to admit outsiders. By preventing a member newspaper from excluding a competitor in its own city, the Court assumed that competition would be promoted in cities with more than one newspaper. See Hovenkamp, supra note __, at 208-09.

120. Under some circumstances, the "disruptiveness" of the outsider will be relevant. For example, excluding a discount broker from a listing service suggests that the broker, if admitted, will promote competition within the group. Under these circumstances, there may be particular concerns about excluding these maverick firms from the market, just as there are particular concerns about their being acquired. See Merger Guidelines, Part 2.12.

121. In Associated Press, the efficiencies of the joint newsgathering activities could still be achieved if a newspaper was not allowed to bar another newspaper in its city from participating. In particular, the participants in the venture did not appear to be concerned that latecomers would take a free ride on earlier risk-taking. See Hovenkamp, supra note __, at 209.

122.Demand side economies can occur, for example, if the product becomes more desirable the larger the number of collaborators. See Competition Policy, supra note ___, at 9-1.

123. See, e.g., United States v. Realty Multi-List, Inc., 629 F.2d 1351, 1372-74 (5th Cir. 1980). There, a group of real estate brokers, who would ordinarily benefit from sharing the listings of additional brokers, attempted to exclude brokers who could promote competition by charging discount fees.

124. See, e.g., Carlton and Salop, supra note __, at 330.

125.The possibility that the outsider would be particularly disruptive played a role in Radiant Burners, Inc. v. People's Gas Light & Coke Co., 364 U.S. 656 (1961) (plaintiff stated a cause of action by alleging that the standards applied by the standard-setting organization were arbitrary and capricious); and Allied Tube & Conduit Corp. v. Indian Head, Inc., 486 U.S. 492 (1988) (plaintiff proposed new product standard, which was rejected by established competitors).

126.This is not to say that the fairness of the procedures should become an important aspect of determining whether the outsider is "arbitrarily excluded." The Court rejected a general requirement of procedural fairness in Northwest Stationers, 472 U.S. at 293. In Allied Tube, the unfair procedures used by the association ("packing" the meeting to vote against a proposed new standard) amounted to a discriminatory application of a standard by economically interested competitors. See 486 U.S. at 509. In addition, the control over standard-setting effectively excluded the outsider because the group's standards were routinely adopted by public and private building codes. Id. at 495-96.

127. There still remains the problem of determining whether exclusion is "arbitrary" without involving courts in the analysis of the objective validity of a certification standard. To this extent, it may be necessary to examine the motive of the organization despite the difficulties of doing so. The alternative -- a court's attempting to assess the legitimacy of a professional or technical standard -- is even less workable. See, e.g., Wilk v. American Medical Ass'n, 719 F.2d 207, 213 (7th Cir. 1983), cert. denied, 467 U.S. 1210 (1984), where the Seventh Circuit put the burden on the association to show good faith in excluding a chiropractor access to medical facilities. See also Hovenkamp, supra note __, at 213.