Antitrust Policy Toward
Network Joint Ventures
Harvey N. Bock
Senior Vice President and General Counsel, Credit Services
MORGAN STANLEY, DEAN WITTER, DISCOVER & CO.
Presented at the Hearings on
the Federal Trade Commission's
Joint Venture Project
June 30, 1997
I am Harvey Bock, Senior Vice President and General Counsel for the credit services businesses of Morgan Stanley, Dean Witter, Discover & Co. Those businesses include the Discover Card, which the company launched in 1985, and the NOVUS network, which the company announced in 1995. The NOVUS network was created to serve as an alternative to the networks operated by the Visa and MasterCard joint ventures, both for other card brands issued by our company and for credit cards issued by other companies. In addition to the Discover Card, our company issues four brands of cards on the NOVUS network, with more in the wings. However, much as we would like to have them as customers, no other issuers offer cards on the NOVUS network at this time. As I will explain, that fact is a direct result of the issues that I would like to discuss this morning regarding dominant network joint ventures.
There are many examples of dominant network joint ventures in the financial services industry. In addition to Visa and MasterCard, they include, for example, ventures that operate automated teller machine networks, such as the "MOST" network in the Washington area. Dominant network joint ventures are also found in other industries -- for example, local real estate associations that operate multiple-listing services. What these joint ventures have in common is that they operate networks that facilitate transactions among most of the firms in a given market. This morning I would like to discuss some of the very significant antitrust issues that they present.
Our General Counsel, Christine Edwards, touched on these issues in her testimony before the Commission in 1995 regarding horizontal competitive issues of credit and charge card networks, and I commend that testimony to your attention.
Let me say at the outset that my remarks differ sharply in many important respects from the comments that Paul Allen, Visa's General Counsel, made to you on June 24. To cite just two examples:
First, Mr. Allen's assertion, that a dominant horizontal joint venture like Visa should be treated for most purposes no differently than a unitary firm, tosses aside a wealth of established antitrust law recognizing the risks that exist whenever competitors coordinate their activities. There are very substantial differences between the incentives of joint ventures and single firms as economic actors, and those differences have long been recognized by antitrust law. Professor Herbert Hovenkamp, who worked closely with the late Professor Areeda as co-author of their antitrust treatise, has recently written specifically about those differences.(1)
Second, Mr. Allen's remarks ignore the practical reality that many dominant network joint ventures possess immense market power, as Visa does together with MasterCard because of their virtually identical, nearly industry-wide ownership. Their share of the U.S. market for credit card network services is around 75 percent and growing steadily, and their members' collective share of the U.S. credit card issuing market is even higher. As I will explain, network joint ventures with this degree of dominance have enormous potential for anticompetitive conduct, which calls for more, not less, antitrust scrutiny.
With all due respect, I believe that Mr. Allen's position, that antitrust scrutiny of Visa and other dominant network joint ventures needs to be relaxed, has it exactly backwards. Let me turn in more detail to why this is true, and explain why network joint ventures present competitive issues that are different from those of other joint ventures; why network joint ventures are more likely than other joint ventures to become dominant, and why they then deserve special antitrust scrutiny; and what are some of the types of anticompetitive conduct by dominant network joint ventures that antitrust enforcement should focus on.
2. The Special Competitive Characteristics of Network Joint Ventures
Antitrust law has long recognized that joint ventures of actual or potential competitors have the potential for abusive conduct, but when those joint ventures are formed to operate networks they deserve special scrutiny. Joint ventures can, of course, create efficiencies that enhance competition. Network joint ventures, however, are capable of producing the special efficiencies that economists call "positive externalities," whose competitive impact is more complicated. These are the efficiencies that accrue as participation in a network grows, simply as a result of the growth of the number of participants in the network. The telephone system is a standard example: A network with one telephone customer is useless; a network with a hundred participating customers is of some value, although that value is limited because of the small number of connections that can be made by means of the network; and a network that connects millions of telephones has exponentially greater value.
These positive externalities of network joint ventures have direct implications for competition and antitrust policy. As network joint ventures grow, these positive externalities add to their incumbency advantages and help to entrench them competitively. A prospective challenger must not only offer superior price or quality; it must, in addition, compete with the advantage that the incumbent enjoys simply by virtue of the number of participants that it has, and overcome the reluctance that customers will have to use a network that has fewer participants.(2)
Let me illustrate with an example from my own industry. If a firm wants to compete as a national issuer of credit cards on the MasterCard or Visa network, it can recruit customers by offering a product that is superior in price or quality. The cards that it offers will be as useful to its first customer as they will be to its thousandth customer. No customer will turn down the firm's MasterCard credit card just because there are not many other consumers who carry it. By contrast, if the same firm wants to start a new credit card network and offer the network's services to merchants and card issuers, the barriers to its entry are much greater. Even if the new network offers lower prices and better quality, it will still be at a huge competitive disadvantage simply because of its initial size. It will face a daunting uphill battle attracting merchants, cardholders and card issuers to a network that initially has far fewer of each as participants. In short, the unique incumbency advantage that network joint ventures enjoy as a result of their positive network externalities can be extraordinarily difficult for other networks to overcome, even if the aspiring competitors have significant cost or quality advantages.
3. The Still Greater Scrutiny Due Dominant Network Joint Ventures
These network externalities give rise to a "positive feedback" effect as network joint ventures grow. As a result, it is not at all unusual for one network joint venture to become dominant in a particular market. This is true of the ATM networks in many regions. The side-by-side Visa and MasterCard networks may at first glance appear to be a counter-example, but it is more a case of "the exception that proves the rule," because Visa and MasterCard share nearly identical memberships.
When a network is operated by a joint venture of its participants, the network is likely to also become dominant in another sense: Its owners may come to include competitors who have a dominant collective position in their own markets. As you can see from the figures that I cited a few minutes ago, Visa and MasterCard, viewed collectively, are clearly a dominant network in both senses: They have an overwhelming market share in the market for network services, and their members have an even higher collective market share in the market for credit card issuing services.
A joint venture with market dominance sustained by its positive network externalities has enormous potential to abuse competition. For the reasons I discussed earlier, such a venture is relatively invulnerable to competition from other networks. In addition, its structure puts at its disposal the collective market power of its members to limit competition. It can marshall this resource either through informal persuasion or through explicit rule-making. These means can be used to the detriment of competition both in the venture's own market for network services and in its members' related markets. Let me illustrate three ways in which this can occur:
First, a dominant network joint venture can restrain member-against-member competition. For example, in response to the runaway success of the AT&T Universal Card (which was issued through a Visa member bank), Visa adopted rules to limit its members' ability to offer co-branded cards with retailers and other nonbanks. These rules directly restrained competition within Visa's membership by excluding improvements in quality. Other network joint ventures, such as Associated Press, have also have engaged in this kind of conduct.
Second, a dominant network joint venture can restrain competition by non-members against its members. Visa did this several years ago when it responded to the launch of the Discover Card by attempting to orchestrate a merchant boycott against Discover; impeding Discover Card's ability to compete had the effect of directly restraining competition against Visa's members by a lower-priced non-member. Other networks, including real estate multi-list services, have also engaged in this kind of conduct by using boycotts and other devices to exclude non-members from their members' markets.
Third, a dominant network joint venture can restrain competition in its own market, by hindering competition from other networks. In the balance of my testimony this morning, I would like to focus on this type of restraint.
Note that a single practice often will have effects on competition at more than one level at the same time. For example, when Visa refused to allow my firm to offer a Visa card because we operated our own card network, it was protecting both itself and its members from competition: itself, by deterring others from following our example and creating competing networks to Visa's; and its members, by blocking the introduction of a major new low-priced Visa card. The same was true of the merchant boycott that Visa orchestrated against Discover Card: in addition to blocking Discover Card's competition against its members, Visa was explicitly warning others against introducing competition against itself by launching cards outside the Visa and MasterCard networks.
In the limited time remaining this morning, I would like to focus on two ways that dominant network joint ventures can protect themselves from competition against themselves.
A. Refusals to Deal
The first type of restraint that I will discuss is the refusal to deal with a competing network. Let me refer you to Dennis Carlton and Steven Salop's excellent article last year on joint venture refusals to deal.(3) A dominant network joint venture has a compelling interest in ensuring that a new entrant, particularly one that is a maverick, is unsuccessful, and the joint venture can pursue this goal by means of either of two forms of refusal to deal: It can orchestrate a group boycott by its members of the competing network, or it can itself refuse to deal with the competing network. Visa and MasterCard have done both vis-a-vis my company's NOVUS network and American Express's network. The clearest example of their organizing a group boycott are the rules they have both adopted that prohibit any of their members from issuing cards on either of the smaller networks. (As you know, of course, they both permit their members to issue cards on each other's network.) Because of the dominance of the Visa and MasterCard network joint ventures, banks cannot risk giving up Visa and MasterCard's services as the price of also purchasing network services from NOVUS or American Express; so the effect of the rules is to prevent NOVUS or American Express from offering network services to banks.(4)
A dominant network joint venture can also undermine competition by smaller networks by refusing to deal with them itself. For example, Visa forbids our network and American Express's from contracting with merchants to handle Visa transactions, even though they permit MasterCard's affiliated processor do so. Visa and MasterCard have also excluded us from participation in certain fraud prevention programs. The anticompetitive purpose of these actions, which is to restrain smaller networks, is clear from the fact that in these cases, too, Visa and MasterCard do not refuse to deal with one another.
B. Venture Expansionism
The second form of potential abuse by dominant network joint ventures, to the detriment of competition, is the extension of their activities into new markets. It is striking that our antitrust enforcement mechanisms typically do not scrutinize changes in the scope of the activities of joint ventures. Even if the original formation of a joint venture was appropriate, there is no reason to assume that the network or other economies that justified its creation will also justify its encroachment into other markets. Where the joint venture has become dominant in its original area of activity, it is just as likely that its expansionism will leverage the venture's power into the new field and entrench its position in the original one. Only scrutiny of the facts of the particular case will tell, but such scrutiny generally does not occur.
For example, Visa and MasterCard have expanded into a variety of new technologies and markets since their formation almost 30 years ago. These include ATM networks (the Plus network in the case of Visa, and the Cirrus network in the case of MasterCard), electronic transaction processing for merchants, and stored-value card systems.(5) As far as I know there has been almost no review of that expansion. This is a serious oversight. Visa and MasterCard were created to enable a geographically fragmented banking industry to manage the clearing of credit card transactions in a paper-intensive world that bears little resemblance to today's emerging world of electronic commerce. In fact, I believe that in today's environment, networks owned by a much smaller number of companies would be more efficient and more competitive at providing these services, and they would certainly present less risk of abuse; but unless antitrust constraints are imposed, they simply cannot overcome the associations' ability to leverage their entrenched power into new markets.
Anticompetitive conduct by dominant network joint ventures has far-reaching consequences, not just for companies like mine that aspire to be effective and successful competitors, but for tens of millions of affected consumers in a variety of industries. I hope that these hearings will lead to more effective antitrust enforcement in this area, and we are prepared to work with the Commission and its staff in pursuit of that goal.
1. Herbert Hovenkamp, Exclusive Joint Ventures and Antitrust Policy, 1 Colum. Bus. L. Rev. 1 (1995).
2. Some of the policy implications of this feature of network joint ventures were recently discussed in William H. Pratt, James D. Sonda & Mark A. Racanelli, Refusals to Deal in the Context of Network Joint Ventures, 52 Bus. Lawyer 531 (1997).
3. Dennis W. Carlton & 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 (1996).
4. A similar practice was attacked by the Department of Justice in a non-joint-venture context in its 1994 consent decree with Electronic Payment Services, Inc. ("EPS"), in which EPS agreed to cease requiring participants in its dominant "MAC" ATM network to obtain ATM processing services from EPS. [1994-2 Transfer Binder] Trade Reg. Rep. (CCH) ¶70,796 at 73,388.
5. Similar concerns are expressed with respect to Visa and MasterCard's move into debit card processing in David A. Balto, Antitrust Curb Perverted Economics of Off-Line Debit, Am. Banker, May 9, 1997 at 9.