V. How Can "Insider Competition" Change the Ability and Incentive to Exercise Market Power?

Single-share analysis implicitly assumes that the venture and its parents act as a single firm in all markets related to the joint venture. Under this assumption, the focus of single-share analysis is on whether other competitors are willing and able to prevent the exercise of market power by the joint venture. Steven Salop has termed this "outsider competition," referring to competition from firms that do not belong to the joint venture.(93)

Joint ventures, however, may preserve three additional sources of competition that could prevent the venture or the collaborators from exercising market power. First, the parents may continue to compete against each other outside of the venture in markets affected by their collaboration. Second, the parents may compete against each other within the joint venture by independently marketing joint venture output. Third, the venture might permit its members to compete against it as part of other collaborations. Steven Salop has labeled these forms of rivalry "insider competition," referring to competition from firms that participate in the venture.

If outsider competition is significant in markets affected by the joint venture, the joint venture (and its partners) will not be able to raise prices above levels that existed before the formation of the joint venture.(94) If, on the other hand, "outsider" competition may not be sufficient to defeat the exercise of market power, then courts and agencies should examine the ability and incentive of the collaborators to compete against each other and their joint venture.

The first part of this section explores how "insider competition" can prevent the collective exercise of market power by the joint venture. The second part describes the factors relevant to whether insider competition is likely to occur. The third part examines how courts and agencies have assessed insider competition. The final part addresses how the potential for insider competition should be weighed in market power analysis.

A. How Does Insider Competition Limit the Potential Market Power of Competitor Collaborations?

When collaborators have the ability and incentive to continue competing against each other and their venture, it is less likely that the formation and operation of the venture will allow the partners or the venture to exercise market power in any relevant market. This section describes how three types of insider competition can operate to prevent or reduce the exercise of market power.

1. The Ability and Incentive of Collaborators to Compete Against Each Other Outside of the Joint Venture, through Separate, Independent Operations in the Joint Venture and Other Markets

If collaborators possess the ability and incentive to continue competing against each other outside of the venture, their coordinated activities will more likely be limited to the joint venture itself. When this sort of insider competition is likely to continue after the formation of the venture, it may be improper to assign the venture and its parents a single market share in determining whether the venture facilitates the collective exercise of market power. Instead, the venture should be assigned a market share based on its actual revenues (or capacity), and the parents should be assigned separate market shares based on the size of their independent operations.(95) In these cases, it is less likely that the creation of the venture will lead to the exercise of unilateral market power (because the scale of the venture's operations is smaller than complete collaboration between the parents) or tacit collusion (because the continuing independence of the parents makes the market less concentrated than it would be otherwise).

In the case of joint ventures that operate upstream or downstream from their parents, however, courts and agencies assess competitive effects in two markets: the markets in which the parents compete, and the market in which the joint venture will operate. Although insider competition between the parents may reduce the likelihood of anticompetitive effects in markets in which they compete, it will not prevent the venture from exercising market power in its own market. Nor will this form of insider competition prevent members from realizing supracompetitive profits through the joint venture entity.(96) Thus, courts and agencies should be careful to examine other forms of insider competition.

2. The Ability and Incentive of Collaborators to Compete Against Each Other Within the Joint Venture

Another potential source of "insider competition" is price and output rivalry within the joint venture. Even when partners contribute all of their significant productive assets to the collaboration, they may still compete against each other by pricing and marketing the venture's output independently. The nature of some ventures will even permit individual collaborators to increase joint venture output above levels that would exist if the parties jointly controlled venture output.(97) If the partners have the ability and incentive to market the venture's production independently, it is less likely that the firms will use the joint venture to exercise collective market power even if they do not have separate, independent production capacity outside of the joint venture. This form of insider competition is especially significant in markets which require cooperation between the firms to make their product or service available. Because collaborators in these markets are unable to conduct independent business operations outside of the venture, rivalry between the firms within the venture may be the only source of insider competition that can prevent the exercise of market power.

Even when partners have the ability and incentive to market joint venture production independently, the collaboration may still exercise market power by setting high transfer prices for the sale of production to the collaborators. Although the partners appear to compete, they realize supracompetitive profits through the joint venture entity. This has led some commentators to observe that competition within the venture will be most effective when the joint venture transfers production to its members at marginal cost.(98) But the same commentators have also cautioned that courts and agencies should avoid requiring transfers at marginal cost when the venture must recoup high fixed costs in order to function effectively.(99)

3. The Ability and Incentive of Collaborators to Compete Against the Joint Venture As Part of Another Collaboration

Even if partners continue to compete against each other outside and inside the collaboration, the joint venture may be able to exercise market power by denying its members the right to compete as part of another collaboration in the joint venture market. Thus, in some circumstances, courts and agencies will examine whether the venture prevents its members from competing against it as part of another collaboration. If the collaborators have the ability and incentive to compete against their joint venture, it is less likely that the venture will be able to exercise market power.

B. What Factors Make Insider Competition More or Less Likely?

When effective, insider competition may reduce the ability of collaborators to exercise market power through the creation of a joint venture. Often, however, courts and agencies must assess the likelihood and effectiveness of insider competition when many fundamental facts remain unknown. When the parties have agreed to one or more forms of ongoing insider competition or have not expressly prohibited it, courts and agencies conduct a more detailed inquiry into (1) whether insider competition is likely to occur or is occurring, and (2) whether it is sufficient to prevent the creation or exercise of market power.(100)

This section examines five factors that make insider competition more or less likely: (1) the ability of the partners to continue to conduct independent business operations outside of the joint venture; (2) the effect of ownership and control on the incentive to compete; (3) the duration of the collaboration; (4) the potential to exercise market power by eliminating insider competition; and (5) the exchange of information through the joint venture.

1. Ability of Partners to Compete

Joint venture activities may have a substantial impact on the ability of the participants to continue to compete against each other and the venture. Therefore, it may be desirable to examine how the formation and operation of the joint venture affects the cost and capacity of the parent's independent business operations. Although the specific economic impact of joint venture formation will vary from case to case, certain types of ventures are more likely than others to reduce the ability of collaborators to continue to compete against each other and the venture.

Collective buying arrangements, for example, are not likely to reduce the ability of members to compete against each other or the venture. Because purchasing is largely a contracting function, collective purchasing arrangements rarely require the members to make significant financial contributions. The formation of the joint venture does not divert assets from individual members that are necessary to continue independent buying. Members remain able to purchase the same amount of inputs that they procured before the formation of the joint venture.

The effect of marketing joint ventures on the ability to conduct independent operations is also fairly easy to assess. Like purchasing joint ventures, joint marketing arrangements do not require members to contribute or divert hard assets to the collaboration. As long as members do not commit a specific percentage of their output to the collaboration, marketing joint ventures are not likely to alter the ability of members to remain effective independent competitors against their partners or the venture, although such ventures may well affect the incentives of the members to remain effective independent competitors.

The effect of vertically integrated joint ventures on the ability of members to compete is more difficult to evaluate. In the case of input production joint ventures, members may be able to buy inputs from suppliers that compete with the joint venture. If the venture attempts to restrict the quantity of inputs, in some circumstances members may have the fairly costless ability to purchase cheaper inputs from rival suppliers.(101) In general, members will likely be unable to turn to cheaper sources of supply only if the input joint venture has created market power in the input market.

In the case of vertically integrated output production joint ventures, members may retain the ability to compete against their partners if they do not commit all of their input production to the collaboration. If members are not required to commit any of their input production to the collaboration, they remain able to produce and market the same independent level of output that existed before the formation of the venture. In those circumstances, the formation of the venture would be less likely to affect competition in the input market or to deprive downstream rivals of the joint venture of the inputs they need to compete effectively.

Output production joint ventures also raise complex issues. A key question is whether the joint venture will produce its output from new capacity in the market. If the collaboration is limited to production at new facilities, the formation of the venture may not reduce the ability of the partners to manufacture the same level of independent output that existed before the joint venture.(102) If, by contrast, the partners contribute some of their productive assets to the joint venture, the collaboration likely will reduce the ability of the partners to manufacture and market the same level of independent output that existed before the joint venture, assuming that parents were operating at full capacity before the formation of the venture In this scenario, productive capacity that had been used competitively will now be used cooperatively. Thus, even if the competitors retain the incentive to compete against each other and the venture, they will have lost part of their ability to do so.

Even the creation of a new product may reduce the ability of collaborators to compete against each other and their venture. When collaborators combine their independent productive assets to produce and market a new product that competes with their older product lines, they are increasing joint control over previously competing sources of output.(103) Even when the new joint venture product does not compete with independent product lines, the creation of the new product might reduce the ability of members to continue competing in the sale of goods produced with assets that are used to manufacture the joint venture product.

2. Ownership, Control, and Incentives to Compete

Variations in the ownership and control of the joint venture may significantly affect the incentives of collaborators to continue competing against each other and the joint venture. The greater the parent's stake in the joint venture, the less likely that the parent will continue to compete against the venture in the relevant market.(104) Commentators have observed that the size of the joint venture in relation to the independent activities of the parents may have a significant impact on the incentive of the members to continue competing in the relevant market.(105) For example, a fifty percent stake in a venture that is five times larger than the parent's independent operations in the relevant market will reduce the incentive of the parent to compete against its joint venture. Similarly, the parent's ability to coordinate its independent business decisions with those of the joint venture increases as its control over the joint venture increases. A controlling parent may attempt to ensure that the venture does not cannibalize the parent's independent sales in the joint venture market.(106)

Similar disincentives to compete may arise when the joint venture takes the form of interlocking equity investments among the partners, an increasingly popular form of strategic alliance.(107) When the alliance is horizontal, partners may have less of an incentive to take competitive actions that would reduce the profitability of the investment in their competitor. When the alliance is vertical, partners may have less of an incentive to deal with competitors of their partner.

How can courts and agencies determine the probable impact of ownership and control on the incentives of collaborators to compete against each other and their venture? One alternative is to modify single-share analysis to incorporate alternative ownership and control arrangements. Steven C. Salop and Timothy F. Bresnahan have constructed a modified HHI test to examine the impact of ownership and control on the competitive effects of output production joint ventures.(108)

Bresnahan and Salop first examine the effects of ownership and control without factoring in the specific market shares of the parents or the specific level of their respective ownership interests in the venture. Using their modified HHI analysis, they conclude that, all else being equal, output production joint ventures are least likely to result in anticompetitive effects if the venture independently sets its price and output decisions to maximize its own profits, and if the parents continue to compete against each other in the relevant market.(109) Output production joint ventures are most likely to result in significant anticompetitive effects when the parents use the joint venture to maximize the profits of the venture and its partners, and the parents stop competing against each other in the joint venture market. Bresnahan and Salop label this structure "partial merger."(110)

According to Bresnahan and Salop, ownership and control mechanisms that fall between these extremes result in more ambiguous effects. When a single parent controls the venture's price and output decisions, Bresnahan and Salop demonstrate that it will result in fewer anticompetitive effects than a partial merger between the parents.(111) Similarly, when the venture sets price and output levels to maximize the profits of itself and the parents, it will still result in fewer anticompetitive effects than a partial merger if the parents continue to compete against each other. Bresnahan and Salop label this structure "limited joint control."(112)

After Bresnahan and Salop incorporate the specific market shares and equity stakes of the parents into the modified HHI, they arrive at two additional conclusions. First, they demonstrate that control of joint venture price and output by one firm is more competitive than limited joint control only if the controlling firm's equity share in the joint venture is larger than its share of total parental output.(113) For example, control by Firm A will be more competitive than limited joint control if Firm A owns 60% of the venture, but controls only 40% of the total output manufactured by the partners outside of the venture. Second, they show that control by Firm A is more competitive than control by Firm B only under the same circumstances that control by Firm A is more competitive than limited joint control.(114)

Salop and Bresnahan have constructed a useful framework for assessing the impact of alternative control and ownership arrangements on the competitive effects of output production joint ventures.(115) Unfortunately, it is not clear whether unambiguous results are possible beyond the specific assumptions of their model, or whether their model is applicable to joint ventures that are involved in other types of activities.

3. Duration of the Collaboration

The duration of the venture will affect the likelihood that participants in nonexclusive collaborations will continue to compete against each other in markets affected by the venture. When the duration of the joint venture is limited, "the parents will continue to compete in the relevant market knowing that their collaboration's end is on the near horizon."(116) Indeed, some commentators have suggested that courts and agencies limit the duration of joint ventures where greater collaborative activity between the partners could result in anticompetitive effects.(117) Duration, however, is a factor whose significance will vary considerably depending on the nature of the joint venture and its size in relation to the independent business activities of the parents in the relevant market. Duration is also a factor that remains within the control of the parties after the transaction; the parties may extend the duration of their collaboration after their transaction has been reviewed. Finally, duration may be unknown at the time of joint venture formation. All of these considerations limit the utility of duration as an independent factor in determining the likelihood of insider competition.

4. The Potential to Exercise Market Power

The potential for exercising collective market power through the joint venture will also have an impact on the likelihood of insider competition. For example, assume that five competing purchasers form a collective buying arrangement that permits members to continue purchasing outside of the venture. If the venture is able to use market power to secure lower prices, it is less likely that the members will continue independent purchases in the relevant market. Conversely, if the purchasing venture were unable to secure lower prices through the exercise of market power, it is more likely that members will continue to compete against each other and the venture in buying inputs. Thus, courts and agencies should consider the potential market power of the venture in determining whether insider competition is likely to continue after the formation of the venture.

5. Exchanges of Information

Insider competition is more likely to occur when the parties create procedural safeguards that limit the exchange of competitively sensitive information about the joint venture market. The absence of such protections increases the ability of joint venture members to coordinate their activities with those of the joint venture and the independent operations of other members in the joint venture market.(118) "Spillover" may facilitate coordinated activities even when the joint venture agreement allows the parties to compete with each other and the joint venture. The parties may create mechanisms to ensure that information exchanges are limited to topics that are essential to the effective operation of the venture. The specific circumstances of each venture will determine the nature and extent of safeguards that are necessary.

C. How Have Courts and Agencies Analyzed Insider Competition?

Courts and agencies have analyzed insider competition primarily in two contexts. First, they have looked at insider competition in determining whether the joint venture (or its members) will be able to exercise market power. In these cases, insider competition is part of a larger inquiry into the actual or probable effects of the collaboration. Second, courts and agencies have assessed whether the elimination of insider competition may be an independent antitrust violation. Thus, insider competition is itself the focus of the inquiry. Both types of cases are discussed below.

1. Ability and Incentive of Collaborators to Compete Against Each Other Outside of the Joint Venture, through Separate, Independent Operations in the Joint Venture and Other Markets

In many cases, the most significant form of "insider competition" is the ability and incentive of actual competitors to continue competing against each other and the venture in the joint venture market. This can occur when the collaboration is designed to manufacture or sell products that compete with products sold by the members. In addition, joint venture members may or may not continue to compete in other (sometimes related) markets. Antitrust enforcers must therefore evaluate the impact of the collaboration not only on the joint venture's market, but also on other markets in which the parents may currently compete.

The ability and incentive of joint venture members to continue to compete against each other is a significant factor in determining whether the collaboration will enable the venture or its members to exercise market power in the joint venture market or other markets. In Broadcast Music, Inc. v. Columbia Broadcasting System, Inc.,(119) for example, CBS sued two organizations (ASCAP and BMI) that issued blanket licenses for the use of copyrighted music. CBS alleged that the collective negotiation of fees for blanket licenses represented price-fixing that was per se illegal under § 1 of the Sherman Act. The Supreme Court disagreed, finding that the blanket licenses reduced the immense transaction costs of individual negotiations for copyright licenses and perhaps created a new product.(120)

Instead of eliminating competition between composers, ASCAP and BMI offered a new product with unique characteristics that no individual composer could offer. As the Court explained, "[t]o the extent the blanket license is a different product, ASCAP is not really a joint sales agency offering the individual goods of many sellers, but is a separate seller offering its blanket license, of which the individual compositions are raw material. ASCAP, in short, made a market in which individual composers are inherently unable to compete fully effectively."(121) The joint pricing of the blanket license was therefore not plainly anticompetitive: "[t]he individual composers and others have neither agreed not to sell individually in any other market nor use the blanket license to mask price fixing in such other markets."(122) In addition, consent decrees required ASCAP and BMI to permit individual composers to continue to sell their individual rights, allowing consumers to obtain individual compositions without purchasing blanket licenses.(123)

Why did the court rely in part upon the consent decrees in deciding not to apply the per se rule? It was not because the consent decrees allowed individual artists to provide meaningful competition against the joint venture. The Court's opinion implies that buyers were not interested in individual licenses as competitive alternatives to the blanket license. If individual composers were unable to compete against their copyright collectives, the consent orders could not have prevented BMI and ASCAP from exercising market power in the sale of blanket licenses.(124) Rather, the real significance of the consent orders was in preventing the copyright collectives from eliminating competition in the sale of individual licenses.

Perhaps what the Court was saying in BMI was that the blanket license created a new product market without reducing the level of potential output in the older one. Not only did producers remain free to market their individual products outside of the venture, but they remained able to produce the same level of output that existed before the formation of the joint venture.(125) Thus, the creation and operation of BMI and ASCAP could be viewed as procompetitive because consumers did not lose the benefits of any competition that existed before the creation of the blanket license.(126) When viewed as an output production joint venture, BMI appears to be a collaboration that created a new product from new capacity.(127) In short, the Court was correct in focusing on the impact of the venture on the ability and incentive of individual composers to continue to compete against each other outside of the collectives.

Shortly after the Supreme Court decided BMI, similar issues arose in a government antitrust challenge to a collaboration in the burgeoning cable television programming market. In United States v. Columbia Picture Industries, Inc.,(128) the district court granted a preliminary injunction against the formation of a joint venture that would have allowed four of the top six motion picture production companies to form a premium pay television movie channel. The joint venture agreement provided that each producer was allowed to contribute as many films as it liked to the joint venture. If a producer chose to sell its film to pay television channels, the joint venture would enjoy exclusive rights to the film for nine months. Contributors would be reimbursed under a complex allocation formula that included a flat payment of $1 million per film and a percentage of revenues that the film earned at the box office.(129) Net revenues from the pay channel would be repaid to participants on the basis of their contributions to the venture. Any residual profits would be distributed equally to all participants. The government alleged that the joint venture was a price-fixing and group boycott agreement.

The defendants attempted to invoke the Court's holding in BMI, arguing that the joint venture restraints were part of a larger effort to create a new product, and should therefore be assessed under the rule of reason.(130) The court found that BMI was inapplicable to the facts of the case. Unlike the copyright collectives, the motion picture producers were not combining to create a new product market.(131) Instead, they were jointly setting prices for products that they had previously sold in competition with each other.(132) Although the court did not believe that the price formula was per se illegal, it agreed with the government that the formula would result in anticompetitive effects: "[T]he defendants have, in effect, substituted a profit sharing formula for the competitive negotiations over the value of individual films in the pay television market in which the movie company venturers used to engage."(133) The court believed that anticompetitive effects would flow from the defendants' possession of over half of the market for motion pictures sold to cable television programmers.(134) The court also noted that the joint venture would enable other motion picture producers to raise their licensing fees in sales to competing pay channels.(135)

Columbia demonstrates the importance of evaluating the likelihood that collaborators will continue competing against each other in markets other than the joint venture market. The formation of the joint venture could have created unilateral market power if the nine-month exclusivity window allowed the venture to raise prices in the pay movie channel market. This was more likely to occur because the venture prohibited the producers from continuing to compete in sales to channels that would have competed with the venture.(136) Thus, the elimination of competition between the parents outside of the venture would have facilitated the exercise of market power in the production and pay channel markets.

The impact of insider competition between the parents is easier to assess when the venture operates in the same market as the parents. If the parents are willing and able to continue competing against each other and the joint venture in the joint venture market, assigning a single market share to the venture and its participants will not provide an accurate assessment of whether the venture will result in anticompetitive effects in that market. Under these circumstances, a limited joint venture is less likely to result in anticompetitive effects than a merger between the collaborators in the joint venture market. In General Motors Corp.,(137) for example, the Commission approved a consent decree permitting a production joint venture between two of the world's largest automobile manufacturers. Under the terms of their joint venture agreement, General Motors and Toyota would each own 50% of a production joint venture that would manufacture subcompact cars at a mothballed GM plant in Fremont, California. Under the terms of the consent decree, the two companies agreed to limit annual production at the Fremont facility to 250,000 cars.(138) Each party would continue to compete in the manufacture and sale of other automobiles, and the venture was scheduled to terminate after twelve years. The consent decree also established procedural safeguards against the anticompetitive exchange of sensitive information.(139) Absent these procedural safeguards, there would have been a greater likelihood that the parties could coordinate their independent activities outside of the joint venture, perhaps making insider competition between the parents illusory. Because the parents remained willing and able to compete against each other and their joint venture, the Commission did not treat the joint venture as a merger.

As BMI, Columbia, and General Motors demonstrate, the presence or absence of insider competition between partners outside of the collaboration will often be a significant factor when courts and agencies analyze the market power of joint ventures. Although courts and agencies may not be certain that parents will continue to compete outside of their collaboration,(140) certain factors make this form of insider competition more or less likely. In BMI and General Motors, the formation of the joint venture did not reduce the ability or incentive of the parents to continue independent production and marketing -- in BMI, in a separate market, and in General Motors, in the joint venture market. Single-share analysis would have been inappropriate for determining the impact of the joint venture in markets in which members competed. In Columbia, however, the venture would have eliminated competition between the collaborators. The court could therefore apply single-share analysis to determine that the joint venture would result in anticompetitive effects.

2. Ability and Incentive of Collaborators to Compete Against Each Other Within the Joint Venture

Insider competition may also prevent the exercise of market power when collaborators have the ability and incentive to compete against each other within the venture, i.e., to compete in sales of the product of the joint venture. This is especially important in markets where some element of cooperation is required to make goods and services available. Because members are unable to offer such goods and services outside of the venture, they will not provide any independent constraint on the collective exercise of market power by the joint venture. Even when cooperation is not essential, it may be so efficient that collaborative activity becomes economically desirable. As in markets requiring cooperation, competition within these sorts of joint ventures may reduce concerns that partners will be able to exercise market power through the formation and operation of the venture.

In NCAA, the Supreme Court addressed the significance of competition within joint ventures. The NCAA had pooled the college football telecast rights of its members for sale to television networks. Like BMI, NCAA involved the joint sale of intellectual property rights. But unlike BMI and ASCAP, the NCAA did not permit its individual members to sell rights outside of the joint venture. The NCAA's sale of broadcast rights was thus a direct restriction on the output of college football telecasts.(141) Although direct output restrictions are usually per se illegal, the Court was reluctant to condemn the joint sale because some element of cooperation was essential to produce college football.(142) Nevertheless, the Court quickly determined that the restraint was so likely to result in anticompetitive effects that the defendants had to offer some procompetitive justification to avoid summary condemnation. The defendants failed.(143)

Why was the NCAA's joint sale of television rights any different from the blanket licenses in BMI? The Court identified two distinctions. First, the NCAA eliminated all competition between its members in the sale of telecast rights: "[in BMI] there was no limit of any kind placed on the volume that might be sold in the entire market and each individual remained free to sell his own music without restraint. Here, production has been limited, not enhanced."(144) Second, the NCAA's joint sale of telecast rights did not create a new product: "Unlike Broadcast Music's blanket license covering broadcast rights to a large amount of individual compositions, here the same rights are still sold on an individual basis, only in a noncompetitive market."(145) By preventing schools from marketing their telecast rights independently, the NCAA not only eliminated competition between its members, but also prevented schools from competing against their collaboration. The restrictions therefore reduced output below the level that would have existed if the schools were able to compete.(146)

Competition within joint ventures may also be important even when collaboration is not necessary to manufacture or market a product. In SCFC ILC, Inc. v. Visa USA, Inc.,(147) for example, competition within Visa's national network of credit card issuers had a decisive impact on the court's assessment of Visa's market power. Visa had passed a rule forbidding membership to any firm offering credit cards that Visa's board deemed "competitive." When Sears, the owner of Discover Card, sought to issue a Visa "Prime Option" card, the Visa board rejected its application. Sears sued Visa, alleging that Visa's new rule was a group boycott that reduced competition in the issuance of general purpose credit cards. The district court affirmed a jury verdict against Visa, finding that Visa's members possessed over 70% of the general purpose credit card market.(148) The Tenth Circuit reversed, holding that the district court improperly aggregated the market shares of Visa's members in determining that Visa wielded market power in the card issuing market.(149) Because Visa set no limitations on the interest rates, terms, or number of cards that its members offered, Visa's issuers were able to engage in competition within the joint venture. Visa's rule was therefore less likely to result in anticompetitive effects in a market that the 10th Circuit described as "atomistic."(150)

Competition within the venture will be more effective if members have greater independent control over the amount of output available through the joint venture. In United States v. Alcan Aluminum,(151) for example, the court approved a consent decree that permitted either member of a production joint venture to expand production at their jointly owned facility. If one member was not operating its portion of the factory at full capacity, the other member had the option to expand production using the other member's idle capacity by assuming the variable cost of production. Moreover, each member could unilaterally expand capacity through capital investments. By preventing the venture from limiting the output of either partner at the facility, the consent decree increased the ability of the members to compete against each other within the joint venture.

As the Supreme Court observed in NCAA, "[e]nsuring that individual members of a joint venture are free to increase output has been viewed as central in evaluating the competitive character of joint ventures."(152) Competition within the venture may be an especially significant factor in preventing the exercise of market power by joint ventures operating in markets where cooperation is necessary. In those markets, competition between the parents may only occur within the venture. Moreover, even if parents are also willing and able to compete against each other outside of the venture, competition within the collaboration may provide an additional safeguard against the exercise of market power.

3. Ability and Incentive of Collaborators to Compete Against the Joint Venture As Part of Another Collaboration

Insider competition may also have significant procompetitive effects when members have the ability and incentive to compete against the joint venture. This is especially important when the venture operates in a market in which its parents do not or cannot compete independently. In markets requiring some degree of cooperation, partners may not only compete against each other within the venture, but may also compete against the venture as part of another collaboration. This form of insider competition enhances the likelihood that outsider competition -- in the form of alternative networks -- will prevent the venture from exercising market power.

The First Circuit addressed the significance of this form of insider competition in U.S. Healthcare, Inc. v. Healthsource, Inc.(153) Healthsource was a physician-owned HMO in New Hampshire that offered higher compensation to doctors who would not deal with competing networks.(154) A competing HMO brought suit under § 1, alleging that the provisions raised entry barriers in the health care financing market. The court observed that the challenged exclusivity agreements presented "one common danger for competition: an exclusive arrangement may 'foreclose' so much of the available supply or outlet capacity that existing competitors or new entrants may be limited or excluded, and, under some circumstances this may reinforce market power and raise prices for consumers."(155) But the court found that the plaintiff had failed to offer sufficient evidence that Healthsource's restraints prevented health plans from competing effectively in the health insurance market.(156) Because physicians could terminate their participation in Healthsource's incentive plan upon 30 days' notice, competing health plans could continue to obtain access to Healthsource physicians. The court also pointed out that Healthsource had secured exclusive participation from only 25% of the primary care physicians in the relevant market.(157) Competing health plans thus retained access to the remaining 75% of physicians, providing some constraint on Healthsource's ability to exercise market power through its incentive plan. In the absence of evidence that foreclosure had occurred, the court was unable to conclude that the restraints were unreasonable.

ATM networks have also become embroiled in antitrust controversies over provisions that discourage or prohibit members from participating in other networks. In The Treasurer, Inc. v. Philadelphia National Bank,(158) an ATM network sought to enjoin Philadelphia National Bank's ("PNB") acquisition of Mellon Bank's CashStream ATM network. The acquisition agreement provided that PNB's Money Access Service ("MAC"), the largest regional ATM network, would bring Mellon and its branches into its network.(159) MAC did not permit its members to participate in competing networks. In addition, Mellon would solicit other CashStream members to become participants in the MAC network. According to the plaintiff, the acquisition would simultaneously increase MAC's market power in regional ATM markets and deprive the Treasurer of potential participants in its own network.

The district court held that the acquisition would not result in anticompetitive effects. First, the court found that the acquisition gave the Treasurer an opportunity to solicit CashStream members to become part of its network. Second, the court believed that MAC's exclusivity requirements were not significant because members could terminate their membership upon 180 days' notice. The court expressed doubts about the plaintiff's narrow market definitions, contending that plaintiff had not adequately accounted for potential entry or bank ATMs that were not presently affiliated with any network. The court also noted that MAC had adopted its exclusivity requirement at the time of its formation, making it less likely that the restriction was anticompetitive: "The restriction is merely part and parcel of an obviously successful, comprehensive marketing strategy."(160) Finally, the court found that Treasurer would only benefit if exclusivity permitted MAC to raise ATM fees.

One may question the court's analysis of MAC's exclusivity requirement. Although exclusivity may be procompetitive when a firm has no market power, it may result in anticompetitive effects over time if the venture acquires a significant portion of the relevant market. If alternative networks are unable to gain access to the dominant network's members, the firm may be able to exercise market power. It also was not clear that CashStream's acquisition actually expanded opportunities for the Treasurer. Because CashStream had previously been nonexclusive, its members could also have participated in the Treasurer network. Now presented with an all-or-nothing choice by an ATM network with significant market share (which is significant in network industries driven by economies of scope and scale), former CashStream members were forced either to join MAC or to continue competing as members in smaller networks. If they chose to join MAC, there would be fewer constraints on MAC's ability to exercise market power in regional ATM markets. If they remained outside of MAC, they faced the risk of participating in networks that may have been too small to compete against MAC.

As the holding in Treasurer demonstrates, caution is warranted in assessing the ability and incentive of collaborators to compete against their joint venture as part of other collaborations. If the joint venture possesses a significant percentage of the network market, the ability of members to participate in other networks may be essential to prevent the joint venture from exercising market power. But even when members are able to terminate their membership in an exclusive network, they may lack the incentive to join networks that are too small to compete effectively in the relevant market. Permitting other networks to obtain access to members of a leading network may enhance intersystem competition if members have the ability and incentive to participate in other collaborations. But on the other hand, requiring networks to allow their members to join other ventures could reduce intersystem competition by aligning the incentives of members in competing networks.(161)

D. How Should the Potential For Insider Competition Be Weighed in the Analysis of Market Power?

Insider competition may reduce the ability and incentive of collaborators to exercise market power through the formation of a joint venture. If insider competition is likely to occur, assigning a single market share to the venture and its members may tend to overstate the actual competitive impact of the collaboration.(162) That has led some commentators to conclude that joint ventures should be approved if a merger between the collaborators would be permissible in markets affected by the partnership.(163) According to these observers, if a merger would be permissible, courts and agencies do not need to evaluate the likelihood or impact of insider competition.(164) Some commentators have gone further, contending that the potential for insider competition makes joint ventures inherently less anticompetitive than complete integration between the partners. According to these observers, courts and agencies should require more significant evidence of market power than in merger cases before concluding that a joint venture is likely to result in anticompetitive effects.(165)

Neither of these approaches describes how courts and agencies should weigh the potential for insider competition in the market power inquiry. The first approach makes insider competition irrelevant if a merger between the parties would be permissible in the joint venture market. The second approach makes insider competition important, but only in the abstract. Rather than explore the probability and magnitude of insider competition, the second approach would raise the permissible market power threshold for all joint ventures based on the potential for insider competition. As the cases discussed above demonstrate, however, there can be no easy assumptions about whether insider competition is likely to occur, or whether it will be sufficient to make a transaction less anticompetitive than a merger between the same parties.

So how should insider competition be incorporated into the market power inquiry?(166) It is improbable that the factors which make insider competition more or less likely can be incorporated into market power analysis in a quantitative manner.(167) The likelihood of insider competition depends upon a number of factors that may be unknown at the time of antitrust review. Indeed, even clear contractual provisions remain subject to change after courts and agencies approve certain transactions. Similarly, the effectiveness of insider competition will depend upon variables whose impact may be uncertain. Even if courts and agencies can predict when insider competition is likely to occur, they may not be able to determine the magnitude of its impact in preventing collaborators from exercising market power through the joint venture.(168) Courts and agencies should therefore continue to conduct the detailed factual inquiry that has led to correct results in most of the cases described above.

But there are ways to structure the analysis of insider competition that are likely to improve its accuracy and reduce its complexity. This section of the paper has articulated the most significant factors that make insider competition more likely to occur and more likely to be effective in reducing the potential anticompetitive effects of joint ventures. The questions that agencies and courts might ask include:

• Will the parties be able to exercise collective market power if they stop competing against each other and do not compete against their venture?

• What are the competitive concerns revealed by single-share analysis, and which forms of insider competition could make a difference?
• If members contend that they will compete against each other outside of the venture through separate, independent business operations, how does the joint venture affect the ability of the members to continue to compete against each other or their venture in any affected markets? Has the venture required the members to divert valuable assets from their independent operations?
• If members contend that they will compete against each other within the joint venture, how will the venture set the transfer price of production to the members? Will individual members be able to expand the venture's output without consulting its partners?
• If members contend that they will compete against each other and the venture through membership in other collaborations, will they have the ability and incentive to join competing networks? Will competing networks be able to obtain sufficient scope and scale to compete against the venture?
• How many competitive variables will members continue to control independently (e.g., output, price, advertising, investments, research, development)?
• How does the member's financial interest in the venture affect its incentive to compete against the venture or to deal with rivals that compete with the venture?
• Does the venture have the ability to enforce restrictions on insider competition by penalizing members that compete against the joint venture or other members?
• Will the length of the collaboration make it more likely that the member will continue competing against its partners and the venture?
• Have the parties created procedural safeguards that will reduce the risk of anticompetitive exchanges of information about markets in which they compete?

Although these factors do not lend themselves to precise mathematical balancing, they are essential to making qualitative judgments about market power in the joint venture context.

93. Salop, supra note 49, at 225.

94. Salop points out that "[a]s long as outsiders have the ability and willingness to compete using the old technology, the venture cannot price above the initial level because the venture would be undersold by the outsiders." Id. at 225-26.

95. As Salop has observed, when "membership in the cooperative venture does not eliminate the ability of insiders to continue to compete using the old technology, then the potential for competition by insiders using that old technology also will prevent price from rising above [the initial level]." Id. at 226.

96. In the context of input production joint ventures, competition between members in the downstream market may not prevent the partners from charging themselves high transfer prices for inputs and realizing supracompetitive profits through the joint venture. See Dennis W. Carlton and Steven C. Salop, You Keep On Knocking But You Can't Come In: Evaluating Restrictions on Access to Input Joint Ventures, 9 Harv. J. L. & Tech. 319, 334 & n.34 (1996); Carl Shapiro and Robert D. Willig, On the Antitrust Treatment of Production Joint Ventures, 4 J. Econ. Perspectives 113, 116 (1990).

97. If there are few physical and financial obstacles to increasing joint venture production, collaborators may be able to exercise greater independent control over the level of output that they can secure from the joint venture. This is especially true when the marginal cost of joint venture production approaches zero, as is often the case with intellectual property. In manufacturing ventures, by contrast, a facility's capacity places a ceiling on the level of output that venture members can obtain through the collaboration.

98. See Salop, supra note 49, at 237 n.10; Shapiro and Willig, supra note 96, at 117.

99. Salop, supra note 49, at 231 (venturers may need to avoid competing within the venture "especially where cooperative activities are characterized by severely increasing returns to scale in the sense of high fixed costs and low marginal costs."); Shapiro and Willing, supra note 96, at 117 (If "intra-venture competition does not permit parents to recoup the fixed costs of the venture, requiring it may cause the venture to collapse.").

100. The agencies have recognized that insider competition may not occur even when partners have the contractual right to compete against each other and their venture. See Statements of Antitrust Enforcement Policy in Health Care Issued by U.S. Department of Justice and Federal Trade Commission (August 1996) (hereinafter, the "Health Care Statements"). In the context of physician network joint ventures, the agencies look for evidence that physicians continue to compete outside of the venture, either individually or as part of other collaborations. Among the factors that the agencies examine are whether (1) viable competing networks exist with sufficient physician participation; (2) member physicians are actually willing to contract with other plans or networks; (3) member physicians earn significant revenue through contracting with other plans or networks; (4) the formation of the network has led or will lead to significant "departicipation" in other plans or networks; and (5) members have coordinated pricing through their participation in other plans or networks. See Health Care Statement 8.A.3.

101. Members could also attempt to manufacture inputs, but that could entail significantly higher costs than purchasing inputs from existing competitors of the joint venture.

102. This does not mean that the venture necessarily would be procompetitive. If, for example, the parents would have created new capacity independently, the formation of the venture may reduce the level of independent output that would have existed if the joint venture had not been formed. Similar competitive concerns arise when the venture imposes collective control over new capacity that one of the parents had created before the formation of the venture. See, e.g., Yamaha, 657 F.2d at 980-81.

103. If the parties only combine unutilized capacity, members may remain able to manufacture the same level of independent output that existed before the venture, although their incentive to do so will be affected by the extent of competition between the new and old product lines.

104. See Pitofsky, supra note 8, at 1012 (the incentive to compete against the venture increases as the number of joint venture owners increases); Thomas A. Piraino, Jr., Beyond Per Se, Rule of Reason or Merger Analysis: A New Antitrust Standard for Joint Ventures, 76 Minn. L. Rev. 1, 65 (1991) ("Regardless of whether they have entered into an express non-competition agreement, the partners to a joint venture will have a natural inclination to avoid competing with the joint venture. Direct competition is contrary to the partners' interest because it reduces the profits of their own affiliate.").

105. See Edmund W. Kitch, The Antitrust Economics of Joint Ventures, 54 Antitrust L.J. 957, 962 (1987).

106. See Timothy F. Bresnahan and Steven C. Salop, Quantifying the Competitive Effects of Production Joint Ventures, 4. Intl. J. Ind. Org. 155, 156 (1986). According to Salop and Bresnahan, joint venture partners have "a private incentive to structure the joint venture agreement in a way that removes the discretion of the joint venture management over these competitive instruments and replaces discretion with either control by the parents or with a formula that determines prices and outputs."

107. See Robert J. Reynolds and Bruce R. Snapp, The Competitive Effects of Partial Equity Interests and Joint Ventures, 4 Intl. J. Ind. Org. 141 (1986).

108. Bresnahan and Salop, supra note 106.

109. Id. at 166. In this scenario, the participation of the parents is limited to a silent financial interest in the venture.

110. Id. at 161.

111. Id. at 167. Salop and Bresnahan refer to this structure as "control by one parent."

112. Id. at 160.

113. Id. at 167.

114. Id.

115. Bresnahan and Salop also modify HHI analysis to reflect "competitor-based" transfer pricing arrangements within the joint venture, but get less precise results. Id. at 161-66. As Bresnahan and Salop explain, "[c]omparing schemes with competitor-based escalator clauses depends on the price details of the escalators." Id. at 168.

116. Piraino, supra note 104, at 37 (advocating greater permissibility of joint ventures that are limited in scope and duration).

117. See, e.g., Joseph F. Brodley, Joint Ventures and Antitrust Policy, 95 Harv. L. Rev. 1523, 1547 (1982) (suggesting that limitations upon the duration and scope of the venture will ensure that the parents retain the incentive to compete in the relevant market),

118. See Kattan, supra note 64, at 949 ("Joint ventures therefore may serve as conduits for coordinating the participants' market behavior or for exchanging competitively sensitive information. The potential for such adverse spillover effects may be limited, however, if venture participants lack sufficient market shares in markets in which they compete to create, enhance or facilitate the exercise of market power."); Brodley, supra note 117, at 1561 (discussing spillover in input joint ventures); Pitofsky, supra note 8, at 1033-34.

119. 441 U.S. 1 (1979) (hereinafter, BMI).

120. Id. at 21-24

121. Id. at 22-23.

122. Id. at 23-24.

123. Id.

124. But see Salop, supra note 49, at 234 (In BMI, "insider competition thus could constrain the power of the collective and prevent prices from rising above the level that would be achieved in the absence of cooperation.") (emph. supplied).

125. It is important to remember, however, that other forms of insider competition may be relevant in assessing a venture's market power. See Salop, supra note 49, at 234 ("The Court did not discuss the degree of competition between the two major collectives, ASCAP and BMI, nor how the intensity of competition might increase if there were a larger number of smaller collectives. Had they analyzed this issue, the Court would have found it necessary to balance the likely increase in competition if there were more collectives against the potential loss of efficiency benefits if the market structure were altered in this way.").

126. Indeed, this is precisely what the Second Circuit concluded when BMI was remanded: "[I]f the opportunity [to purchase individual licenses] is fully available, and if copyright owners retain unimpaired independence to set a competitive price for individual licenses to a licensee willing to deal with them, the blanket license in not a restraint of trade.") Columbia Broadcast System, Inc. v. Am. Soc. of Composers, Authors and Publishers, 620 F.2d 930, 936 (2d Cir. 1980), cert. denied, 450 U.S. 970 (1981); see also Buffalo Broadcasting, Inc. v. Am. Soc. of Composers, Authors and Publishers, 744 F.2d 917, 933 (2d Cir. 1984), cert. denied, 469 U.S. 1211 (1985).

127. See Robert Pitofsky, A Framework for Antitrust Analysis of Joint Ventures, 74 Geo. L.J. 1605, 1619 (1985) (BMI effectively involved the creation of a new facility by contract).

128. 507 F. Supp. 412 (S.D.N.Y. 1980), aff'd mem., 659 F.2d 1063 (2d Cir. 1981).

129. Id. at 420.

130. Id. at 429-30.

131. The court reasoned that "the circumstances in the instant case include already existing markets - a market for the product that Premiere proposes to sell and a highly competitive market for the product that the movie company venturers propose to reserve solely to Premiere." Id. at 430.

132. The court observed that "[i]t is true that, as to the defendants and pay television program distribution, Premiere is a new venture. The heart of that venture, however, is the fifty or so new, never-before-shown-on-television, motion pictures that the defendants annually distribute and would, otherwise, be selling to the existing market." Id.

133. Id. at 431.

134. Id.

135. Id. at 432 n.49.

136. Even when vertically integrated joint ventures permit members to buy or sell to its rivals, members may have the incentive to cease competing against each other outside of it. See United States v. Primestar Partners, 1994-1 Trade Cas. ¶ 70,652 (S.D.N.Y. 1994) (consent decree prohibiting vertically integrated cable operators from using direct broadcast satellite [DBS] joint venture to deny programming to competing DBS providers).

137. 103 F.T.C. 374 (1984).

138. Id. at 383.

139. Id. at 384-85.

140. When courts and agencies are evaluating conduct that has occurred in the past, they will often have direct evidence of whether the parents have continued to compete outside of the venture. But even when parents have continued to compete outside of the venture, it still may be appropriate to assign the venture and the collaborators a single market share. See Addamax Corp. v. Open Software Foundation, Inc., 888 F. Supp. 274 (D. Mass. 1995) (although aggregating market shares is usually improper when collaborators continue competing against each other and their venture, specific circumstances raised genuine issue of material fact as to whether single market share was appropriate).

141. 468 U.S. at 108.

142. Id. at 101.

143. Id. at 111-15.

144. Id. at 114-15.

145. Id. at 113-14.

146. See also Chicago Prof. Sports Ltd. Partnership v. National Basketball Ass'n, 961 F.2d 667 (7th Cir.) (affirming preliminary injunction against NBA rule reducing the number of games that teams could independently market to "superstations"), cert. denied, 506 U.S. 954 (1992).

147. 36 F.3d 958 (10th Cir. 1994), cert. denied, 115 S. Ct 2600 (1995) (hereinafter, Visa II).

148. See SCFC ILC, Inc. v. VISA U.S.A., Inc., 819 F. Supp. 956 (D. Utah 1991) (hereinafter, Visa I).

149. 36 F.3d at 967-68.

150. Thus, in assessing Visa's market power, the Tenth Circuit focused on how "intra-joint venture" competition limited any market power that Visa might have had. A separate question would be how the challenged rule affected "inter-joint venture" competition -- that is, competition between credit card systems. Such issues are discussed in the next section.

151. 605 F. Supp. 619 (W.D. Ky. 1985).

152. 468 U.S. at 114 n.54.

153. 986 F.2d 589 (1st Cir. 1993) (hereinafter, Healthsource).

154. Id. at 593.

155. Id. at 595.

156. Id. at 595-597.

157. Id. at 596.

158. 1988-1 Trade Cas. ¶ 67,943 (D.N.J.), aff'd mem., 853 F.2d 921 (3d Cir. 1988).

159. Id. at 57,795.

160. Id. at 57,803.

161. See State of New York v. Visa U.S.A., Inc., 1990-1 Trade Cas. ¶ 69,106 (S.D.N.Y. 1990) (prohibiting overlapping membership in Visa and MasterCard debit card networks to facilitate intersystem competition); Worthen Bank & Trust v. National Bank Americard, Inc., 485 F.2d 119 (8th Cir. 1973) (holding that restrictions on membership in competing joint ventures were not per se illegal; restrictions could increase intersystem competition and prevent a de facto merger between credit card networks), cert. denied, 415 U.S. 918 (1974); Visa I, 819 F. Supp. at 997 (acknowledging that admission of Discover Card to Visa would reduce competition at the network level, but "not so substantially as to rise to the level of a Section 7 violation."); see also Nat'l Bank of Canada v. Interbank Card Ass'n, 507 F. Supp. 1113, 1123 (S.D.N.Y.) (upholding prohibition on overlapping membership that protected network's investments in the creation of a new product and lasted only eight years), aff'd on other grounds, 666 F.2d 6 (2d Cir. 1981).

162. The agencies have recognized that conventional merger analysis, while appropriate for some ventures, should not be applied to collaborations if insider competition is likely to occur. As Joseph Kattan explains:

[A]lthough the antitrust enforcement agencies analyze joint ventures as mergers, they have been more lenient toward joint ventures and have permitted some joint ventures to proceed in circumstances in which they had or would have challenged a merger of the same parties. This more permissive approach is grounded in the belief that restrictions on the scope and duration of joint ventures tend to limit the effects of joint ventures vis-a-vis mergers of the same firms. Unlike mergers, joint ventures may often maintain the participants' status as independent competitors outside the framework of the collaborative effort. Kattan, supra note 64, at 947.

163. The Department of Justice embraced this assumption for its analysis of anticompetitive effects in the joint venture market in its 1988 Antitrust Enforcement Guidelines for International Operations (the "1988 International Guidelines"). In analyzing the potential anticompetitive effects of joint ventures, DOJ would first examine "with respect to each market the level of and increase in concentration that would result if the parties merged. If, based on market concentration, the Department would not challenge a merger of the joint venture participants in a relevant market, then the Department concludes without detailed examination of other factors that the joint venture and its individual restraints would not likely have any anticompetitive effects in that market. . . . " Id. at § 3.42. The Guidelines also considered competitive effects outside of the joint venture market. Id. at § 3.43. The 1988 International Guidelines were withdrawn in 1995.

164. See Shapiro and Willig, supra note 96, at 127-28 ("Any genuine joint venture should almost surely be allowed if the participants would be permitted to merge, and this would almost surely be the outcome of a rule of reason analysis."); Richard Schmalensee, Agreements Between Competitors, in ANTITRUST, INNOVATION, AND COOPERATION (T. Jorde & D. Teece eds.) at 112 (1992) (regardless of intent or efficiencies, if a merger would be legal, the joint venture must be legal); Martin B. Louis, Restraints Ancillary to Joint Ventures and Licensing Agreements: Do Sealy and Topco Logically Survive Sylvania and Broadcast Music?, 66 Va. L. Rev. 879, 905 (1980) (""A court, therefore, should not deny to a lawful joint selling agency a useful restraint that an equivalent merger among the competitors could lawfully impose within the merged entity.").

165. See, e.g., Schmalensee, supra note 164, at 113 ("Since partial integration by contract is generally reversible, it would be appropriate to require plaintiff to prove more substantial market power than would be necessary in a merger case."); cf. Kitch, supra note 105, at 961 n.10 ("As a matter of theory it is not clear that the 'cut off' -- necessarily a rough number -- should be the same for mergers and joint ventures . . . . I would argue that for all joint ventures except those relating to the current marketing and sale of products, the 'cut off' should be higher because the threat to competition is less.").

166. In Health Care Statement 8, the agencies incorporate insider competition into safety zones for integrated physician network joint ventures. Physician networks enjoy a safety zone if they employ fewer than 30% of the physicians in each relevant specialty and if physicians continue to compete against the venture (either individually, or as part of another venture). If, on the other hand, physicians do not continue to compete outside of the network, the market share safety zone decreases to 20%.

167. In Visa I, the district court rejected Visa's argument that HHI analysis proved that Discover's compulsory admission to Visa would result in anticompetitive effects. Because Visa and Discover would continue to compete outside of the venture, Discover's admission "would not constitute a complete merger of the two entities. At best, it could be called a partial or hybrid merger. Under such circumstances, the court finds that the HHI analysis is not particularly useful in determining the probability of anticompetitive effects." 819 F. Supp. at 994 n.47.

168. The 1988 International Guidelines seem to acknowledge that the likelihood and impact of insider competition will depend on the facts. Although the Guidelines note that insider competition could significantly reduce the potential for anticompetitive effects, they do not describe how it would be evaluated in specific cases.