The dynamic implications of increasing returns: Technological change and path dependent inefficiency

This paper addresses the issue of path-dependent selection and inefficiency, counterfactual methods and the empirical evidence employed in this research trajectory. The general message is basically that the theoretical argument to prove path-dependent inefficiency is relatively straightforward, while empirical demonstrations, as well as the analysis of welfare implications are much more difficult. In the first section, we briefly state the basic argument explaining how localized learning leads to an incomplete exploration of the variety distribution of a technology. We then synthesize the most interesting historical studies, which get potential regret results, and ask whether these results suggest a large welfare loss. To try to answer this question, we start from the critical arguments of Liebowitz and Margolis to discuss the theoretical and empirical difficulties of establishing path-dependent inefficiency. (C) 1997 Elsevier Science B.V.