Network economies

The phenomenon of network externalities as a subject of interest to economics emerged in the 1980s. Within a few years, several seminal contributions by two to three ¹ different groups of rather closely linked researchers were published. First were David and his group (David, 1985; David and Bunn, 1988) analyzing a number of historical cases of technology standards that may have had an inferior outcome; the approach was, however, purely qualitative. Then a group of mostly evolutionary and complex modelling economists around Arthur, David, Kaniovski and Dosi ² (Arthur et al., 1982, 1987; Arthur, 1989; Dosi and Kaniovski, 1994; Dosi et al., 1994) and a group of more neoclassical game theorists ³ (Katz and Shapiro, 1985, 1986; Farrell and Saloner, 1985, 1986) independently formalized the theory. Both groups contributed a great variety of important findings, in turn developing a rich and accurate terminology for this field of study. The term network externality which was coined by Katz and Shapiro, (Katz and Shapiro, 1985) was intended to describe a special case of Coase's externalities. ⁴ This was a failure of a theoretical neoclassical market to attain the theoretical optimum resulting in one actor losing utility (compared to the optimum) due to actions of another actor, a loss that can for some reason not be traded, i.e. reduced by compensating the other agent for not taking the harmful action. In the present case of network externalities, this reason is generated by the specific property of a good to gain value for any individual user the more users subscribe to the usage of this good (or service). In the absence of an almighty planing entity the optimum state will probably not be attained. This effect is redoubled if the possibility of switching between goods with network externalities is restricted or if there are even switching costs. David (1985) and Arthur and his collaborators (Arthur et al., 1987; Arthur, 1989) extend this argument to a technology (or an industry) being completely dominated by one standard (or product) that will continue to dominate this technology (or industry) even if the dominating standard (or product) is at some point in time not any more (or has never been) the ‘best’ one — purely by the means of an installed base of users. They call this effect a technological lock-in.