Panel 6:
Innovation, Technological Progress and Competition

Innovation, Competition, and
the Theory of Network Externalities

Stanley M. Besen
Charles River Associates


Writing in 1950, Harvey Leibenstein analyzed the “bandwagon effect,” by which he meant “the extent to which the demand for a commodity is increased due to the fact that others are also consuming the same commodity. It represents the desire of people to purchase a commodity in order to get into ‘the swim of things’; in order to conform with the people they wish to be associated with; in order to be fashionable or stylish; or, in order to appear to be ‘one of the boys.’”[1]

Leibenstein was not at all specific about the types of goods he had in mind other than to suggest that they were fashion goods. The bandwagon effect remained largely unexplored for another 20 years or so. At that point, economists interested in the development of telephone networks, which clearly are subject to bandwagon effects, began to explore the issue in some detail using modern game-theoretic techniques. For example, Rohlfs observed that “The utility that a subscriber derives from a communications service increases as others join the system. This is a classic case of external economies in consumption and has fundamental importance for the economic analysis of the communications industry.[2] Rohlfs then applied this insight in analyzing the origins and development of communications networks.

Except for applications to communications, the analysis of “network effects” lay largely dormant until the 1980s. At that point, economic historians such as David[3] as well as economic theorists such as Farrell and Saloner[4] and Katz and Shapiro[5] began to explore these issues in the context of the economics of standardization. This stimulated considerable interest in the topic, with the result that literally hundreds of papers devoted to network industries have been published. Moreover, this is a subject to which major contributions have been made by economic theorists, applied economists, economic historians, applied mathematicians, and engineers. Indeed, there now exists a sort of   “invisible college” in which people from a wide variety of disciplines study the subject and attend the same conferences where standards issues are discussed.

For economists, the theory of network effects, or network externalities, or standardization, has wide applicability.  Indeed, it has fundamental importance for competition policy, regulation, business strategy, intellectual property, and technical change in a wide range of industries; developments in these industries cannot be fully understood without an understanding of network effects.[6]

Sources of Network Effects

As I have already noted, network externalities exist when the value of a product to any user is greater the larger is the number of other users of the same product. There are basically two ways in which such externalities can occur. Direct network externalities exist when an increase in the size of a network increases the number of others with whom one can “communicate” directly. Indirect network externalities exist when an increase in the size of a network expands the range of complementary products available to the members of the network.

Many industries exhibit network externalities. Some examples are:

  • The Public Switched Telephone Network, where the network externalities are direct in that the value that any user places on subscribing depends on the number of others with whom he can communicate.
  • ATM networks, where the network externalities are indirect in that the larger the network the greater is the number of machines at which an ATM card can be used, and hence the greater is the value of the network to any user.
  • Networks of users of computers that use the same operating system, e.g., the Mac network, where there are direct benefits associated with more efficient file transfers and indirect benefits associated with access to a wider range of applications software as the size of the network grows.
  • Networks of users of compatible videocassette recorders, which exhibit what are probably small direct benefits from the ability to exchange cassettes and much larger indirect benefits from being able to purchase or rent a wider variety of pre-recorded cassettes that employ the same format.
  • Networks of drivers of diesel-powered automobiles, who obtain the indirect benefit of having more widely available fuel and service facilities the larger the number of other drivers of such cars.

What We Think We Know About Network Industries (Unsponsored Technologies)[7]

The study of network industries has yielded important insights, which I briefly summarize here. In this section, I discuss unsponsored technologies, where users choose among competing technologies but the suppliers of those technologies either cannot, or choose not to attempt to, influence the nature or pace of adoption, so that the focus is entirely on the behavior of users. In the next section, I discuss the case where technologies are sponsored.

  • Network effects may outweigh preferences for intrinsic product characteristics. That is, although users may prefer the characteristics of one product to those of another if they are on networks of comparable size, the effects of network size will often dominate. Undoubtedly, there are users who in some sense prefer Macs but, nonetheless, opt to join the larger Windows network because of the greater direct and indirect network externalities it provides. Even more graphically, no one would today choose a Betamax VCR even if he preferred it to the VHS technology because there is little or no pre-recorded material in the Beta format.
  • Network industries often exhibit “tippiness,” a tendency for a single technology to dominate. As one network acquires more users, this increases the value of that network to other users, inducing them to join, and so on. Unless the demand by some users for the intrinsic characteristics of the good on a smaller network is especially strong, the outcome will be “winner take all.”
  • When consumers choose sequentially, “stranding” can occur. That is, early adopters of a technology that “loses” may either obtain few network benefits or may have to incur the substantial costs of switching to the winning network.
  • Network industries may exhibit path dependence, so that the behavior of early adopters may have a disproportionate influence on the equilibrium outcome.
  • Expectations may be critical to the final equilibrium because users must often choose among technologies before those technologies have reached their ultimate network size.
  • Lock-in may occur on the “wrong” technology because if, for whatever reason, the wrong technology is chosen, it may be difficult to achieve the coordinated movement of large numbers of users required for the “right” technology to become the standard.

What We Think We Know About Network Industries (Sponsored Technologies)

I turn now to the case in which competing technologies are sponsored by firms that wish to influence the outcome of the standard-setting process, or, more generally, to determine which network “wins.” Here, the focus is on the strategies and tactics adopted by sponsors.[8] In one case, sponsors have decided to engage in a standards “battle.” Among the tactics available to such sponsors are:

  • Attempting to build an early lead -- for example, by making the product available to early adopters at low prices, in order to influence the expectations of late adopters.
  • Attracting the suppliers of complements -- for example, by providing information that facilitates the development of compatible products.
  • Preannouncing products in order to discourage users from joining rival networks.
  • Committing to low future prices in order to assure adopters that they will not be stranded on a small network.

An interesting aspect of this analysis is that the best technology may not always win the standards battle. This can occur, as noted above, because random events may give an inferior technology an insurmountable early lead. But it can also occur because the cost trajectory of the best technology may make it difficult for it to commit to low future prices. It can also occur, of course, because the sponsor of the best technology chose the wrong marketing tactics.

Finally, it is important to observe that the winner will often choose to make it difficult for others to join its network later. This will occur when the benefits to the winner from a somewhat larger network that open membership makes possible are more than offset by the increased competition to which it will then be subject.  Among the ways that the winner might deny access to its network are:

  • Enforcing intellectual property rights, thus making it illegal for rivals to produce compatible products.
  • Changing technology frequently, so that rivals are unable to respond quickly enough to offer products that users can employ on the dominant network.
  • Refusing to share information with rivals about changes in network design, which has the same effect as above.

Of course, sponsors will not always choose to play “winner take all” and compete “for” the standard. Instead, cooperation to create a standard followed by competition “within” a standard may occur. This is most likely where users’ fear of stranding is so great that few users will adopt a product without assurance of what the standard will be. When this occurs, sponsors may agree on a standard and arrange for low-cost licensing of their technologies in order to allow the benefits of standardization to be shared among competing firms. In addition, firms may agree on standards that combine aspects of the technologies of competing sponsors, so as to prevent any sponsor from being disadvantaged in future competition. Alternatively, they may agree to a form of “logrolling,” in which sponsors rotate in having their technologies chosen as the standard. These tactics are all designed to achieve the cooperation required for standardization and, thus, to avoid delays in market development as users “wait and see” which standard wins.

Finally, I want to emphasize that firms must be concerned that, after they have supported the adoption of a standard, the sponsor of the winning technology will then attempt to exploit its dominant position, in the ways described above. If that occurs, one alternative is to bring legal action. However, as we know, that is likely to be both costly and difficult, and there is no guarantee that the litigation will succeed. As a result, it may be far better for all sponsors, before the standard is established, to adopt measures that limit the scope for opportunism. These measures may include agreements to cooperate on future technology development, agreements to shift future development to third parties, and commitments to provide information about design changes to all suppliers in a timely manner. Although these are by no means perfect measures, they may be far superior to invoking the legal process after the opportunistic behavior has occurred.


[1] H. Leibenstein, “Bandwagon, Snob, and Veblen Effects in the Theory of Consumers’ Demand,” The Quarterly Journal of Economics (May 1950), reprinted in W. Breit and H.M. Hochman, Readings in Microeconomics, Second Edition (New York: Holt, Rinehart and Winston, Inc., 1971), pp. 115-116.

[2] J. Rohlfs, “A Theory of Interdependent Demand for a Communications Service,” Bell Journal of Economics and Management Science (Spring 1974), p. 16.

[3] P.A. David, “Clio and the Economics of QWERTY,” American Economic Review (May 1985).

[4] J. Farrell and G. Saloner, “Standardization, Compatibility, and Innovation,” Rand Journal of Economics (Spring 1985).

[5] M. L. Katz and C. Shapiro, “Network Externalities, Competition, and Compatibility,” American Economic Review (June 1985).

[6] Potential applications of the theory are even more widespread; the analysis of industrial location provides a good example.

[7] I have chosen this fairly modest title to signify that not everyone would agree with each of my statements and not everyone, including myself, would apply them in every case.

[8] For a fuller treatment, see S.M. Besen and J. Farrell, “Choosing How to Compete: Strategies and Tactics in Standardization,” Journal of Economic Perspectives (Spring 1994).