Potential competition, actual competition, and economic welfare

Abstract

Though the origins of anti-trust legislation, at least in the United States, undoubtedly lie in late nineteenth century Populist sentiments, its current intellectual foundations rest on modern welfare economics. These intellectual foundations provide more than just rhetoric, as courts have increasingly relied on the testimony of economists to interpret and reinterpret the inevitably vague language of the legislation. The reliance on modern welfare economics is somewhat paradoxical. For the assumptions underlying the Fundamental Theorem of Welfare Economics are peculiarly ill-suited to deal with competition in the central area to which anti-trust is applied - the industrial sectors of modern economies. These sectors are characterized by important non-convexities, which admittedly present inherent problems for competition. The competition that does take place is as likely to take the form of product and R&D competition, rather than price competition. Product competition plays no role in the conventional welfare paradigm, which assumes a complete set of commodities.’ And it is not that Arrow and Debreu simply forgot to include technical change in their formulations: the natural assumptions underlying the analysis of the production, marketing, and utilization of knowledge are basically inconsistent with those underlying the Arrow Debreu model.’ ‘Invited paper. European Economic Association meetings, Copenhagen, August, 1987. Financtal support from the National Science Foundation and the Hoover Institution is gratefully acknowledged. ‘The presence of non-convexittes in production is central in explaining why only a ftnite subset of the set of possible commodities is in fact produced, ‘Techmcal change is characterized by non-convexities in its production [Radner and Stiglitz (1984)]. If technology exhibits. say, constant returns to scale at a fixed technology, when inputs devoted to R&D are included. it exhibits non-convexities. The value or information may increuse with the scale of production. While conventional theory focuses on markets for homogeneous commodities. each piece of new knowledge must be dinerent from previously produced knowledge. Knowledge has many of the properties of public goods, it being both dificult and undesirable to exclude. Markets for knowledge are, at best, imperfect, and, as we have argued