Besen and Farrell (1994) The horizontal competition in industries with network effects often involves choice of firms to compete between standards, rather than within a standard. That is, they often face the choice of whether to makes their product compatible with other firms. Payoffs of compatible and incompatible standard depend on two main factors in the two forms of competition. 1. Skewness of returns. 2. Sharpness of available competitive tactics. 1 Suppose there are two firms. Then there are three combinations of competition strategies: (1) Both prefer to compete to determine standard. (2) Each firm prefers its own technology as standard, but would prefer compatibility rather than go alone. (3) One firm prefers to maintain its standard while another prefer to join rival’s standard. 2 Both prefer to compete in determining standard: Tactics available are (1) Building early lead: Compete intensely in early round. (2) Attracting the suppliers of components: It is important that components (or software) of a product be generously supplied. E.g. Sony’s purchase of CBS. (3) Product Preannoucement: This can retard the growth of rival. E.g. Microsoft. It can cut both ways: Osborne effect, which is more serious for hardware than software manufactures. (4) Price commitment: A public commitment to low long term price. 3 Both prefer standard to stand-alone: (1) Competition resembles a battle-of-the-sex coordination game. (2) Tactics include commitment and concession. (3) Commitments are actions that reduce payoffs from agreeing to the outcome that is less preferred by the firm, or increase payoff to its preferred outcome. (4) Concessions are actions that makes opponent more profitable to adopt the former’s preferred technology. E.g., low-cost-licensing of Intel to AMD; commitment of JVC to share future Sony in videocassette market. 4 development drives out One prefer to compete standard while the other prefer standard: Since one firm always tries to follow the other, the tactics of the firm who does not want to be followed has at least two strategies: by asserting property rights, or by changing technology frequently. E.g., OS/2 vs. Windows; Intel’s 386, 486 chips. 5 Farrell and Saloner (1986) Suppose a durable good exhibits positive network effect, and transition to new standard is graduate, what are its implications on efficiency of adoption of new technology? Both excess inertia (in which consumers fail to efficiently adopt a new technology) and excess momentum (in which consumers inefficiently adopt a new technology) can occur. Tactics to compete include preannouncement and predatory pricing. 6 Model: Continuous inflow of new users. * * Continuous time. Before T , only technology U is available. At T , a new technology V arrives. Those who adopt U cannot switch. Only newly arrived consumers can choose. Consumers arrive at a rate n t 0 , and N t 0 n t 'dt ' . t u x : utility of consuming U when the measure of size of U is x. u' x 0 . v x is the corresponding utility of using V. * First consider two extreme cases: (1) Nobody from T on adopts V; and (2) Everybody does so. 7 A user adopting U at time T, when everyone that follows also does so, has a utility of u T u N t e r t T dt ; T where r is discount rate. If a user at T is the last one to adopt U, his utility is r t T ~ u T u N T e T u N T / r. 8 * T T If a user adopt V at when everyone that follows also does so, his utility is v T v N t N T * e r t T dt , T T * T ~ v T can be defined analogously. * A linear example: n t 1, N t t , u x a bx , v x c dx c d t T . ~ In this case u T a bT / r b / r . u T a bT / r , 2 v T c d T T * / r b / r 2 . 9 In linear case V can be superior to U in either of two ways: It can offer higher network-independent benefits network-generated benefit d b . c a , or it has higher We only have to consider the two extreme cases listed above, because of bandwagon effect. ~ * Adoption of V is SPE if v T u T . * ~ * * v T u T If , the dominant strategy for users close to T is to choose * U, and U will stay forever. 10 The linear case illustrated: 11 * ~ T* u T v If , non-adoption is SPE. * ~ T* u T v If , it is dominant strategy for early choosers and thus all that follow, to adopt V. ~ ~ * Adoption is SPE if v T u T . Nonadoption is SPE if u T v T . * * * They may hold simultaneous, so that there are multiple equilibria. If V is adopted, each user for whom v T u T gains v T u T . Two groups of losers. (a) Early adopters lose u T v T . (b) Old users. Present value of welfare change in linear case is G T * v t u t e 12 r t T * dt bT * / r 2 . * The first term is change in welfare for those arrive after T . The 2nd term is loss to the installed base. In linear case, G can be calculated to equal 2 d b 2rbT * r c a / r 3 ~ * Adoption is an equilibrium if u T u T ; that is * r c a rbT * d . 13 ~ u T ; that is, r c a rb b. Adoption is unique equilibrium if v T 14 * ~ * There can be excess inertia and excess momentum, depending on the relative importance of installed base, network independent utility, and network-dependent utility. Competitive tactics: Preannouncement and predatory pricing. * T , and Preannouncemen: Suppose new product is announced at * * [ T , T ] can pre-adopt V. consumers in 15 * T , the utility of user arriving at t T * is If all adopt V after time v ( t ) . ~ u T and v T If v T * ~ * ~ * t' e rt ' u T * t' for all t ' 0 , , then (1) Non-adoption is unique equilibrium without preannouncement.. (2) Adoption is equilibrium with pre-announcement. 16 Preannouncement can reduce welfare. Predatory pricing: By charging lower price, the manufacturer of U can prevent the potentially early users of V from adopting it, and drive out then new technology from market, even if it is efficient to adopt V. Suppose U is monopolistically supplied and V is competitively supplied. 17 * T A low price of U starting at time can induce users to adopt U (when otherwise they should adopt V). If installed base of V is built up, then V will never be adopted. 18
© Copyright 2026 Paperzz