EXCLUSIVITY RULES IN NETWORK JOINT VENTURES
An important characteristic of a "network" is that, over a significant range, per transaction costs decline as the number of transactions or the number of network participants increases.(2) Networks thus offer the potential of significant efficiencies that increase as the network grows.(3) For that reason, antitrust regards joint ventures that are necessary to create networks hospitably. But the fact that network joint ventures can produce significant efficiencies as a general matter does not mean that every joint venture is efficient or that every agreement or policy made by such ventures is efficient; nor does it mean that the venturers have the same incentives that a single firm occupying the same market niche would have.
Setting aside the more obvious anticompetitive threats of price fixing or product or geographic market exclusion, this paper focuses on anticompetitive effects that may result when network joint venture agreements prevent members from competing outside the venture. Joint ventures produce clear opportunities for innovation by giving firms profit incentives to engage in it jointly. But venturers can also capture equivalent gains by retarding innovation by individual members or outsiders, or curtailing particularly aggressive competitors. Agreements forbidding competition outside the venture can limit innovation or aggressive competitiveness that would otherwise occur.
To illustrate, suppose an industry containing 100 firms makes Widgets. The cost to any individual firm of making and selling a Widget is $5.00, but a new process makes it cheaper to market Widgets jointly. In fact, if two firms market Widgets jointly, per unit costs decline to $4.99; if three do, per unit costs go to $4.98; and so on until costs drop to about $4.00 if all 100 firms join.
Now suppose that firm A in this market develops the Gidget, which has precisely the same production costs and distribution as the Widget and could be produced within the venture, but consumers find the Gidget to be superior.(4) The Gidget presents the individual Widget makers with a significant threat to established market shares. A decisive number of board members responds with a rule that prohibits any firm that wishes to participate in the Widget network venture from selling Gidgets.
The effect of the rule is to present firm A with a choice: it can either give up its venture membership or it can abandon its plan to make Gidgets. If it decides to leave the venture, existing venture members experience a cost increase from $4.00 to $4.01 as one member leaves; but firm A experiences a cost increase from $4.00 to $5.00, as it is excluded from the venture altogether. Which choice the firm makes presents an empirical question, requiring the firm to assess the risks and anticipated profits of moving into Gidgets against the more secure profits of remaining in the Widget network. The remaining venturers profit from either decision: (1) if Firm A stays in the venture and eschews Gidgets, the other members have preserved the status quo and do not have a rival's product innovation to worry about; but (2) if firm A leaves the venture and makes Gidgets, they have raised its costs by $1.00 and to that extent protected their market share. By the same token, consumers are injured by either course of action: if the first, they lose the benefit of a new product; if the second, they obtain the new product but only at a higher price.
Stripped of all collateral facts, the story of the joint venturers' response to Firm A's innovation is anticompetitive. The venture has acted to retard rather than facilitate innovation in its market. The difficult questions for antitrust policy are Does the story describe something that is extremely rare or non-existent, or something that is relatively common? If it does occur, can antitrust tribunals recognize it without an unreasonable number of false positives? Are remedies then available that eliminate anticompetitive evils while preserving efficiencies? I do not have decisive answers to any of these questions, but rather offer a few observations.
I. Joint Venture Exclusion Generally
As a basic premise, exclusion rules by ongoing network-style joint venture must be treated differently than exclusion from traditional joint ventures -- but one must not pursue the point too far.
For example, suppose Kodak, Polaroid, and General Electric develop a joint venture to produce a disposable flash cube.(5) Berkey Photo is asked to join but decides not to. Then, three years later when most of the development has been brought to a successful conclusion, Berkey asks to join. Antitrust would not require the venturers to admit Berkey; to do so would permit Berkey to profit from the investment of others by avoiding the risky developmental period and joining only after success was ensured. Under such a rule, no one would join the venture early and the venture would not form in the first place.
As a general proposition, exclusion rules by ongoing network joint ventures are a different matter. Many of them experience costs that decline as the number of members increases. For that reason exclusion from the venture itself may be tantamount to exclusion from a sizeable portion of the market. For example, a bank not able to join the ventures facilitating electronic fund transfers or an insurance company not able to obtain risk data from Insurance Services Office(6) might be effectively put out of business. In that case, the venture's decision makers may be in a position to force those wishing to innovate outside the venture to make costly and anticompetitive tradeoffs.
But even in the case of non-network ventures one must distinguish between exclusion rules preventing new members from coming in from exclusion rules limiting the kind of outside activities engaged in by venture members. The later are inherently more threatening to competition and require closer scrutiny in both network and non-network ventures. To return to the previous non-network example, we would not want Berkey Photo to have an antitrust right to be a late joiner into the Kodak-Polaroid-GE venture to produce a new flash cube. But suppose the three firms agreed not only that they would produce a new flash attachment, but also that none of them would do any independent research or development of alternative flash devices. Or suppose, there being no such agreement at the onset, Polaroid independently launches into a project to use electronic strobe lights as repeatable flash attachments. Kodak and GE, fearing that Polaroid's strobe will compete with the venture's flash cube, on a 2-1 vote then change the venture's bylaws to prevent any venture member from developing competing products outside the venture. As a result, Polaroid is put to the choice of having to give up its participation in the flash cube venture or else give up its pursuit of the alternative strobe technology.
To be sure, free rider concerns(7) may justify the restriction on Polaroid. For example, there may be no alternative way of preventing Polaroid from misappropriating the learning of the flash cube venture for the benefit of its strobe research. But without passing on the merits of this practice at this point, it seems clear that even in the traditional, non-network joint venture antitrust is appropriately concerned about rules restricting competition outside the venture. If Kodak, Polaroid and GE dominate the market for the development of flash attachments, their agreement may be tantamount to a horizontal agreement not to develop an alternative technology. Antitrust might condemn such rules under a rule of reason even though it generally permits closed-end ventures to turn away late applicants. In sum, not all exclusion rules are alike: some are directed to new members in general; others are aimed at limiting the extent to which venture participants attempt to innovate outside the context of the venture itself.
II. Network Ventures as Single Firms
Some writers opine that network ventures are so inherently procompetitive, and so rarely anticompetitive, that they should be treated in the same way that antitrust treats single firms.(8) That view is unrealistically sanguine. Network joint ventures typically organize the production of one (or perhaps a small number) of a firm's numerous inputs; they do not organize everything, and anticompetitive incentives remain that would not apply to a single firm.
For example, while a firm ordinarily profits by maximizing total output, joint venture members maximize individually. For example, GM's Buick, Oldsmobile, and Pontiac divisions (all wholly owned by GM) manufacture automobiles. Suppose Buick develops an innovation strongly preferred by customers, but which the other divisions, given their existing technologies, cannot readily employ themselves. GM would not ordinarily force Buick to bury the innovation simply to protect the sales of Oldsmobile and Pontiac. Rather, GM would permit Buick to go ahead with the innovation, predicting that aggregate sales (and profits) across the three divisions would increase. This is ordinarily the case for two reasons: first, the Buick innovation steals sales not only from Pontiac and Oldsmobile, but also from Ford, Chrysler, Honda and other competing firms; second, the product innovation does not merely transfer market share from others, it also brings new customers into the market.
Assume that prior to the innovation Buick, Oldsmobile and Pontiac produced 100 units each, at a profit of $1 per unit. After the innovation Oldsmobile's and Pontiac's production drops to 90 units, while Buick's increases to 130 units, per unit profits remaining constant. GM would ordinarily prefer the innovation because it is profitable overall, notwithstanding that two divisions lost money while one gained.(9)
By contrast, joint venturers can be driven to a different decision. Although the Buick innovation enlarges the aggregate sales and profits of the venture, all the increases go to Buick, while Oldsmobile and Pontiac face only increased competition from a superior product. Oldsmobile and Pontiac might then respond with a bylaw preventing any firm employing the new innovation from participating in the venture.(10) Buick would then have to choose between continuing its participation in the venture or developing the new product.(11)
III. Collusion v. Exclusion: Prerequisites and Anticompetitive Effects
External Structure. Successful collusion or exclusion by joint ventures both require that the joint venture as a collective wield substantial market power. If all economies to the venture were exhausted at relatively few members and the market contained several competitively viable ventures, then ambiguities about joint venture practices should generally be resolved in favor of the venture. For example, a single venture's prohibition of competition outside the venture is presumptively harmless if ample opportunities exist for the excluded firm to develop its innovation elsewhere while taking advantage of alternative ventures with substantially the same efficiency potential. The competitiveness of the market suggests that free rider concerns rather than concerns to eliminate competition explains the venture's exclusion rule.
Internal Structure. Even if the venture dominates its market, collusion is unlikely if membership is numerous and generally open to newcomers. For example, the 5500 members of Visa could hardly collude on credit card terms, particularly in a world where such collusion is unlawful and thus could not be enforced by public sanctions. Further, Visa is open to new members, provided that they do not compete outside the venture.
But rules forbidding venture members from competing outside the venture are a different matter. Such activity is generally "structural" in nature and readily detectable. For example, a Visa member can hardly offer a competitive credit card outside the venture in secret.(12) Enforcement of such rules requires no monitoring of output or individual sales -- the kind of activity that so often bedevils cartels.
Anticompetitive Effects. The antitcompetitive effects of collusion are reduced output and higher prices for the cartelized product. By contrast, the anticompetitive effects from rules forbidding competition outside the venture are eliminated or delayed development of product alternatives. This causes welfare losses even if the venture's product is sold competitively. For example, if steel conduit makers in a joint venture that promulgates building codes conspire to exclude plastic conduit from the market,(13) competitive injury occurs notwithstanding that members of the venture continue to compete in steel conduit sales. The loss to consumers comes from the denial of plastic conduit. The incentive to the conspirators need not be the preservation of monopoly profits in steel conduit. It can as easily be the preservation of market position and prior investment in steel conduit -- something that the plastic conduit innovation threatens to take away.
Although we often speak of "balancing" as an important part of antitrust's rule of reason, the fact is that courts are very bad at balancing -- certainly if that term implies some ability to quantify market power and efficiency gains and weigh these against each other. Our measurement tools are simply not up to this task, and we should resort to it only after other, less complex forms of analysis fails. This other analysis consists of looking for and then peeling away decisive issues, in the form of Occam's razor, until hopefully only a small number of cases remain that require real balancing.(14)
One of the reasons that balancing is so difficult in these cases is that, not only is empirical measurement of power and anticompetitive effect difficult in non-obvious cases, but even framing the power and effect issue is difficult. As a first step, it is sensible to require that the venture as a collective have a substantial share of a properly defined market, or that the market be sufficiently concentrated.(15) Beyond that, however, the power-and-effects question becomes very complex and often incapable of measurement. If the only competitive risk were collusion by venture members, the welfare losses of such a cartel could be roughly estimated.(16) But in many exclusion cases the competitive threat is not collusion by members but rather suppression of innovation, or denial of scale or scope economies to those who wish to compete outside the venture, or denial of venture access to those who threaten to compete with unusual aggressiveness. Just as the welfare gains from innovation are notoriously difficult to measure, the welfare costs of lost innovation are at least equally difficult.(17)
For example, suppose a joint venture of a local telephone company and several other firms develops interchange and switching devices, protocols, billing procedures and the like for seamless local/long distance telecommunications. One of the long distance participants then begins developing a wireless alternative to the local system, which could give it the capability of offering competitively attractive or superior local/long distance telecommunications on its own. The venture's decision-making body then responds with a rule forbidding venture members from offering services outside the venture which compete with the services offered by the venture itself. This puts the innovator to the choice of going "naked" with its new development, or else standing pat with the status quo.
Assuming the innovator abandons the new product in order to keep its position in the venture, the welfare losses are the destruction of consumers' surplus that would have been realized had the innovation proceeded. Measuring this is infinitely more difficult than measuring any price/marginal cost relationship. One would need to make predictions about the costs of the innovation and the demand for it at a time when nothing has been brought to market or perhaps even fully developed. Further, the true social cost of the exclusion rule is not merely the cost of this particular innovation loss, but also the loss of other innovations that this firm or other venture participants may have developed.
So the answer in such a case is not to try to balance the gains from the exclusion policy against losses. Rather, it is to assume that significant losses are possible once it is determined that the venture has a substantial market share or a decisive cost advantage over firms not in the venture. At that point any defenses offered for the rule forbidding competition outside the venture would have to be scrutinized very closely. If free rider defenses are offered, they must be documented and the fact finder must ensure that the claimed free riding is not merely competition or product complementarity.(18) Then the fact finder must determine that no less restrictive alternatives exist that will substantially solve the free rider problem.
In rule of reason analysis of joint ventures, the existence or non-existence of exclusivity rules is often decisive. In order to succeed, the cartel must reduce output; but if individual members are permitted to make unlimited outside sales, then the cartel cannot be enforced through the explicit rules of the venture itself. For example, if we are concerned that the maximum price fixing agreement of 1750 physicians in the Maricopa case is really disguised minimum price fixing, our concerns are at least partially alleviated by the fact that each member was permitted to deliver unlimited competing services outside the venture or even participate in other health plans.(19) If the group set prices at supracompetitive levels, physicians wishing to capture the difference would offer lower prices outside the plan and, at the margin, raise output back to the competitive level. By the same token, if the venture dedicates itself to a particular technology (e.g., flash cubes), the ability of individual members to explore alternatives (e.g., electronic strobes) outside the venture keeps the market for new ideas competitive.(20)
V. Exclusivity and Free Riding
As noted previously, "exclusivity" here refers not to a general venture policy of being closed to new members, but rather to a venture's efforts to prevent firms from competing outside the venture. To the extent the venture dominates its market, such rules can effectively prevent innovations that the venture's members do not approve.
Exclusivity of this sort is the exception rather than the rule. In the typical network (and many non-network) joint ventures, members are free to sell either inside or outside the venture. For example, the copy/performance right holders in the Broadcast Music case issued BMI non-exclusive licenses, which meant that they could freely license these same rights outside of the venture process.(21) Nationwide floral delivery networks permit members to deliver flowers outside the network and even to join multiple networks.(22) In one well known case members of a joint venture involved in nationwide moving of furniture and possessions permitted members to make nonventure moves as well, but forbad them from using the venture's intellectual property or equipment.(23) Indeed, even in markets requiring significant transfers of intellectual property, nonexclusivity rules are common. For example, firms that develop software for one computer platform, such as Windows 95, are generally also free to develop it for another platform, such as OS/2 or the Apple OS. As a result, popular programs such as Wordperfect or Quicken come in versions that run on multiple systems. Long distance carriers that interconnect with local telephone networks are generally free to connect with alternative local carriers, to offer cell service, and the like.
As a general principle, network joint ventures function quite well without exclusivity rules, and in the typical case they are unnecessary. If the venture provides significant cost saving or product improvement, then outside sales do not threaten the venture. The purpose of the venture is to improve product or delivery by the venture collectively -- not to make products or services delivered outside the venture less attractive.
To be sure, member ability to sell or innovate outside the venture may sometimes permit free riding on the venture's investment. If properly defined free riding is a significant threat, and if anti-free-riding rules can be implemented without significantly threatening competition, then venturers should be able to limit it. The important requirements are that free riding (1) must be properly defined and identified; (2) must be a significant threat; and (3) must be controlled by the practicable manner that is least threatening to competition.
In the great majority of cases the threatened free riding is unauthorized use of intellectual property and the appropriate fix is enforced limitations on the intellectual property's use rather than prohibition of outside sales altogether. The Rothery case thus represents an appropriate response of a network joint venture to outside sales: permit individual members to make the sales, but forbid them from using the "Atlas" name and equipment.(24) Likewise, Visa has a legitimate interest in permitting member institutions from issuing an outside card and calling it "Visa" or using other intellectual property owned by Visa. But the rule that forbad Visa members from issuing any competitive outside card at all(25) cut much too broadly and, in the presence of aggregate market power, could place a high price on innovation outside the venture.
A. Free Riding must be Identified and Properly Defined
If a network joint venture's exclusion policy is to be justified by claimed free riding, the claim must be substantiated. This means that the defendant must be able to assert the nature of the free riding and provide evidence to support the assertion. Furthermore, claimed free riding must be distinguished from simple competition or product complementarity. The Tenth Circuit erred in its Visa opinion, apparently holding that once a free rider claim is asserted, it is entitled to be accepted notwithstanding a jury's contrary verdict without any inquiry into the nature of the free rider claim or the support for it.(26)
The Tenth Circuit's view appears to have been that exclusion rules have no significant anticompetitive potential and are more or less automatically explained by free rider concerns; as a result, free riding does not have to be established in the particular case but operates as a kind of generic, universal defense. But unless network exclusion rules are virtually never anticompetitive, free rider claims must be articulated and supported. In any event, such a burden is a small one, for any network that has devised an exclusion rule in order to combat free riding should easily be able to state the reason. For example, in Rothery the defendant could easily show that members' outside sales using equipment bearing Atlas's marks, amounted to uncompensated licensing of its intellectual property.
Second, "free riding" must be properly defined. If a Visa member institution wishes to offer its own separate credit card, such as Diners Club or Discover, that issuance is presumptively an instance of competition, not of free riding. Free riding occurs only when one firm misappropriates an investment made by another firm, in a way that reduces the victim's incentive (or the incentive's of similarly situated firms) to invest. Unlicensed use of intellectual property clearly fall into this category: there is little point in making an investment in the Visa name if anyone who wishes can make uncompensated use of it.
But merely taking advantage of economies of scale or scope are not free riding, for they reduce no one's investment incentives. For example, Citibank, who issues both Visa within the Visa venture and Diners Club as a separate proprietary card, might reduce its costs by billing owners of both cards in the same envelope. To be sure, this procedure might injure other Visa issuers who cannot take advantage of Citibank's economy of scope because they do not have a second card. For example, Citibank might pass part of its cost reduction on to Visa customers and steal market share from other Visa issuers. But none of this would be free riding because we would anticipate that Visa sales overall would increase: that is, the cost reduction benefits both Visa as a whole and Diners Club as a whole. If a single firm controlled both Visa and Diners Club it would take advantage of this economy of scope.
Likewise, free riding must be distinguished from product complementarity, which is both ubiquitous and desirable. To give a favorite example, the commercial baker sells more bread because the commercial dairy has made an investment in the production of butter. Importantly, however, the reverse is also true: the dairy sells more butter because bakers bake bread. Thus the development of applications software for Windows 95 is an instance of product complementarity, not of free riding. Although applications (such as Wordperfect or Quicken) sell more copies because they offer Windows 95 versions, Windows 95 sells more copies because a wide variety of applications software is available.(27)
To constitute "free riding," a product development must appropriate the investment of another in a way that reduces the latter's incentive to invest in the first place. As a general proposition, this does not occur when the ability to make sales outside the venture yield economies of scale or scope or enable a firm to offer venture and non-venture products as complements to each other. To be sure, in such cases individual members of the venture who do not offer the two products may experience a loss of market share; but the all important thing is that venture-wide sales of the venture product should increase, and the venture's product as a whole becomes more valuable rather than less valuable.
B. Once Properly Defined Free Riding is Found, it Must be Controlled by the Least Harmful Alternative.
The D.C. Circuit's Rothery decision illustrates a network joint venture's reasonable response to harmful free riding.(28) In that case, individual members of the Atlas system joint venture of movers wished to offer competitive moving services outside the context of the joint venture. But if they did so with Atlas equipment they would wrongfully misappropriate intellectual property and investment made by the venture itself. The chosen fix was not to prohibit members from competing outside the venture at all, but rather to prevent them from the identified misappropriations: those making non-Atlas moves had to do so through separate companies not using any Atlas intellectual property or other of its equipment.
Suppose that the Atlas venture had market power in an area and that the venture set monopoly prices for moves by firms operating under the venture name. If individual movers were then permitted to make unlimited non-venture sales under a different name, and if the individual movers are too numerous to permit secret, informal methods of cartel discipline to be effective, then the non-exclusivity operates as an important escape valve: members wishing to profit from undercutting the cartel could do so. Alternatively, if Kodak, Polaroid and GE are developing the FlashCube and worried about protecting its name, then the way to deal with Polaroid's outside innovation of the strobe is not to forbid the innovation, but rather to forbid application of the venture's name.
In some cases, of course, there is no way to prevent all wrongful misappropriation. For example, when Citibank bills for the Visa and Diners Club cards in the same envelope, Diners Club may benefit merely from its association with Visa.(29) In that case the appropriate rule is to prevent the joint billing, leaving it to Citibank to decide whether joint cards are still worthwhile. Or in appropriate circumstances, those controlling the intellectual property can assess a license fee for its use.(30)
In general, the underlying question to be asked is, assuming the venture as a collective has market power, will the venture's own members be able to force competition in price or innovation through competition outside the venture. A practicable free rider fix that permits members to compete from outside should always be preferred to one that does not.
(2)Alternatively, the value of the venture's product increases with the number of transactions or participants. For example, the value of telephone service increases as the number of others who have also purchased telephone service increases.
Declines in cost are not necessarily global, and they may be offset by increases in other costs. For example, the optimal number of teams for a professional or collegiate football league is limited by the number of games spectators are willing to pay for, the difficulty of organizing a season, ranking participants, recruiting top players, and the like.
(3)On the available efficiencies, see Thomas M. Jorde & David J. Teece, Antitrust, Innovation, and Competitiveness (1992); Carl Shapiro & Robert D. Willig, On the Antitrust Treatment of Production Joint Ventures, 4 J. Econ. Persp. 113 (1990); Gene Grossman & Carl Shapiro, Research Joint Ventures: an Antitrust Analysis, 2 J.L.Econ. & Org. 315 (1986); Janusz A. Ordover & Robert D. Willig, Antitrust for High-Technology Industries: Assessing Research Joint Ventures and Mergers, 28 J.L. & Econ. 311 (1985).
(5)The facts are loosely borrowed from Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263, 301 (2d Cir. 1979), cert. denied, 444 U.S. 1093 (1980). In the actual case Polaroid was not involved in the venture nor was Berkey Photo ever invited to join.
(9)A franchisor would ordinarily make the same decision -- for example, permitting one franchisee to deploy an output increasing innovation even if the result was reduced output at the stores of other franchisees. In general, profit-maximization for the franchisor is consistent with aggregate maximization of franchisees. See Herbert Hovenkamp, Exclusive Joint Ventures and Antitrust Policy, 1995 Col. Bus. L. Rev. 1, 61-64.
(10)It is no answer that in a Coasian world Buick, Pontiac and Oldsmobile would arrive at a joint-maximizing deal -- probably one in which Buick would employ the innovation and compensate Pontiac and Oldsmobile for their losses. If we did live in such a world, antitrust would be pointless. In the real world, joint venturers are numerous, have substantial sunk costs or other prior commitments, information is highly imperfect, and the product subject to bargaining is not sold competitively.
(11)Cf. the Visa joint venture bylaw approved by the Tenth Circuit in SCFC ILC, Inc. v. VISA U.S.A., Inc., 36 F.3d 958 (10th Cir. 1994), cert. denied, 115 S.Ct. 2600 (1995): "the corporation shall not accept for membership any applicant which is issuing, directly or indirectly, Discover Cards or American Express cards, or any other cards deemed competitive...." Under the bylaw "non-VISA members who develop a successful proprietary card would be prohibited from joining the VISA system and current VISA members would be expelled from the system if they developed such a card." 819 F.Supp. at 966. (To the extent it is relevant, the author was consulted by Dean Witter/Discover in SCFC ILC, Inc. v. VISA U.S.A., Inc., 36 F.3d 958 (10th Cir. 1994), cert. denied, 115 S.Ct. 2600 (1995)).
(12)Indeed, in the VISA case itself, Sears originally attempted to issue both Visa and Discover by acquiring a Visa-issuing bank, but Visa learned this and applied its rule. See 819 F.Supp. 956, 965 (D. Utah 1993).
(14)For various descriptions of the process, see 7 P. Areeda, Antitrust Law ¶1511 (1986); H. Hovenkamp, Federal Antitrust Policy: the Law of Competition and its Practice §5.6c (1994); Hovenkamp, Joint Ventures, 1995 Col. Bus. L. Rev. at 121-125.
(15)Structural standards include those developed in the 1992 Horizontal Merger Guidelines. Or see Department of Justice, Antitrust Guidelines for the Licensing and Acquisition of Intellectual Property §4.3.1.
(16)Although even here there is considerable controversy about how such welfare losses are to be measured. See e.g., Posner, The Social Costs of Monopoly and Regulation, 83 J.Pol.Econ. 807 (1975); H. Hovenkamp, Federal Antitrust Policy: the Law of Competition and its Practice §1.3 (1994). For example, one would want to know what percentage of the venture's cartel profits are being reinvested in R&D or production, and how much are consumed in cartel enforcement activities.
(17)The central reason for the measurement difficulties is that traditional "static" neoclassical analysis no longer works. By definition, in innovation markets demand or supply curves are shifting, and one must make highly uncertain predictions about the nature and magnitude of the shifts. See generally F. M. Scherer & David Ross, Industrial Market Structure and Economic Performance 662-685 (3d ed. 1990).
(19)See Maricopa, 457 U.S. 332, 339 (1750 physicians comprising about 70% of those in the County); and id. at 360 (Powell, J., dissenting, stressing importance of fact that the physicians may make unlimited sales outside the plan, including providing services to competing plans); Broadcast Music v. CBS, 441 U.S. 1, 5 (1979) (stressing importance of non-exclusivity in approving blanket license venture).
(20)See also In re Florence Multiple Listing Serv., Inc., 110 F.T.C. 493 (1988) (consent decree forbidding multiple listing service from requiring vote on membership and insisting that applicant not compete with the multiple listing service). Compare General Leaseways, Inc. v. National Truck Leasing Ass'n., 744 F.2d 588 (7th Cir. 1984) (condemning rule under which members of joint venture agreed not to enter one another's territory).
(22)See American Floral Service v. Florists' Transworld Delivery (FTD), 633 F.Supp. 201, 205 & n.5 (N.D.Il. 1986) (none of the nationwide floral delivery networks have exclusivity rules and that most florists are members of multiple networks).
(24)Rothery, 792 F.2d 210, 213. See also American Floral Service v. Florists' Transworld Delivery (FTD), 633 F.Supp. 201, 205 & n.5, 207 (N.D.Il. 1986) (network joint venture to facilitate remote delivery of flowers; rules permitted florists to join multiple services and 85% percent of members belonged to a second service; challenged rules (a) required that when a product specifically designed and promoted by FTD was sold, FTD would collect the commission; and that when a customer specifically requested that the local florist use FTD, the florist must do so).
(28)Rothery Storage & Van Co. v. Atlas Van Lines, 792 F.2d 210 (D.C.Cir. 1986), cert. denied, 479 U.S. 1033 (1987). See also American Floral Service v. Florists' Transworld Delivery (FTD), 633 F.Supp. 201, 205 & n.5, 207 (N.D.Il. 1986).
(29)Although such claims should be regarded with suspicion. The world contains numerous dealers who sell multiple brands and advertise them together -- for example, Sears may advertise high prestige Sony and low prestige Panasonic in the same flyer. Free riding of this nature is ubiquitous, has small consequences most of the time, and certainly should not be presumed to justify arrangements with obvious anticompetitive potentials.