Increasing Returns and Path Dependence in the Economy

2. Competing Tecnologies, Increasing Returns, and Lock-in By Historical Small Events

CHAPTER 2 Competing Technologies, Increasing Returns, and Lock-In by Historical Small Events This paper uses the notion of technologies competing for adoption to explore the properties of allocation under increasing returns. It was already wellknown that allocation problems under increasing returns could show multiple equilibria and possible inefficiency. To these properties the paper adds the possibility of lock-in (or inflexibility) and nonergodicity (or path dependence). The paper is also concerned with the "selection problem"-how one allocation outcome comes to be "selected" over time by small, chance events when there are several possible long-run outcomes. The paper first appeared in September 1983 as Working Paper WP-83-90 at the International Institute for Applied Systems Analysis under the title "On Competing Technologies and Historical Small Events: The Dynamics of Choice under Increasing Returns." The version reproduced here is an update that appeared as Center for Economic Policy Research Publication No. 43, Stanford, 1985. The paper was published, in somewhat abbreviated form, in the Economic Journal 99 (March 1989): 116-31. The statics of markets where commodities show increasing returns or decreasing supply costs are by now becoming familiar. 1 It is well known that in such markets nonconvexities appear, so that multiple equilibria are called into being. These equilibria are typically "comer solutions" with one commodity or one firm monopolizing the market. But while information on preferences, endowments, and transformation possibilities allows us to locate and describe these various possible equilibria, it is usually insufficient to tell us which one will be "selected." There is an indeterminacy of outcome. The author would like to thank the International Institute for Applied Systems Analysis, Laxenburg, Austria for financial support in summer 1983; and Paul David, Ward Hanson, Richard Nelson, Nathan Rosenberg, Martin Shubik, Gavin Wright, and the members of the Technological Innovation Project Workshop at Stanford for useful suggestions, comments and criticisms. 1. Increasing-return studies address a much wider variety of issues than are treated here. See among others: Arrow and Hahn (chap. 7,1971), Beato (1982), Brown and Heal (1976,1979), Farrell and Saloner (1985), Flaherty (1980), Guesneries (1975), Katz and Shapiro (1983, 1985), Kehoe (1985), Krugman (1980), Scarf (1981), Schelling (1978), Spence (1981), and Weitzman (1982). The Schelling and Spence treatments are closest in spirit to the one here. 13 14 Increasing Returns and Path Dependence in the Economy Marshall (1891, p. 485) noticed this indeterminacy a century or so ago in the case of firms with long-run decreasing cost curves competing for a market. Ultimately one firm achieves a monopoly of the industry, but which firm dominates can not be deduced in advance. To proceed farther we need to examine the possible paths by which an outcome comes to be selected. We need, in other words, to examine the dynamics of allocation under increasing returns. To set ideas and a possible strategy for analysis, consider a simple example. Suppose in a certain island cars are introduced, all at more or less the same time. Drivers may choose between the right- and left-hand sides of the road. Each side possesses increasing returns: as a higher proportion of drivers chooses one side, the payoff to choosing that side rapidly rises. Casual thought tells us that we would observe a good deal of randomness to the proportions initially driving on each side, but that, if one side by chance got sufficiently ahead, other drivers would "fall in" on this side, so that eventually all cars would drive on (would allocate themselves to) the same side of the road. Of course the side that "wins"-that comes to "dominate the market"cannot be deduced in advance. The outcome is indeterminate. In such a situation the actual outcome would likely be decided by a host of "small events" outside our knowledge-drivers' reactions, dogs running into the road, the timing or positioning of traffic lights. One way then to bring allocation under increasing returns within the bounds of analysis would be to make explicit these "small events," add them to the model, and examine in detailed "slow-motion" the dynamic process by which they cumulate into an aggregate outcome. This would be difficult in our imaginary example. But as a strategy it may be possible in better-defined cases. Notice that our hypothetical example displays the familiar increasingreturns properties of potential inefficiency and nonpredictability: even though individual choices are rational, there is no guarantee that the side "selected" is, from any long-term collective viewpoint, the better of...