As worked out a century ago, the neoclassical theory of distribution was analytically as well as ideologically satisfying. The simultaneous determination of input and output prices through the operation of factor and product markets seemed to close the explanatory gap left by the classical economists. Seminal contributions by Philip Wicksteed, Knut Wicksell, J. B. Clark, Antoine Walras, Enrico Barone, and others generated a logically rigorous explanation of the sharing of value in a competitively organized economy, at least as stylized, and served to put a quietus on Marxian claims that labor is exploited in the shortfall between product and payment. 1
This major analytical achievement, with its profound implications for the organization of societies, was, however, attained only at considerable cost. Buying into the neoclassical theory of distribution seemed to require the abandonment of the central principle of Adam Smith’s Wealth of Nations (1776), which stressed the importance of the division of labor and its relationship to market size as a primary determinant of economic well-being.
The neoclassical emplacement of a static framework for analysis— [End Page 511] defined by fixity in the size of the resource base and in technology—left no room for direct linkage between policy action on the size of the economic nexus and the rate of growth. For any given market size, competitive organization of the economy serves to maximize value by assuring that all resources are directed to their most productive uses. But there is nothing in this idealized neoclassical model to suggest why or how market size, in itself, matters.
Neoclassical economists may have shied away from follow-on inquiry into Smith’s proposition because they thought that acceptance of Smith’s relationship would have wreaked havoc on their newly discovered theory of distribution. The advantages of specialization suggest increasing rather than constant or decreasing returns, and the observation that industries did not seem everywhere to become more and more concentrated suggests that abandonment of Smith’s theorem was, empirically as well as analytically, less damaging than abandonment of the constant returns postulate so critical to their whole enterprise.
Aside from the very brief excursus into external economies by Alfred Marshall ( 1961), Allyn Young’s oft-cited but little-understood 1928 paper, and Nicholas Kaldor’s (1972 and Nicholas Kaldor’s (1985) insistent criticism of neoclassical orthodoxy, increasing returns, as a topic for analytical inquiry, disappeared for more than three-quarters of a century. This situation changed somewhat dramatically in the 1980s with a resurgence of interest in increasing returns in several different, if broadly related, applications (endogenous growth theory, international trade theory, economic geography, unemployment theory, economics of ethics, and path dependence). As the title of our jointly edited volume, The Return to Increasing Returns (1994), was intended to suggest, there has been a major shift of interest toward this subject matter by modern economists.
Modern economists, however, do not exhibit the history-of-ideas focus that would lead them to reexamine the late-nineteenth- and early-twentieth-century neoclassical developments in the theory of distribution. They remain apparently unconcerned with either the continuing contradiction or the potential for reconciliation between Smith’s theorem and the conventional postulate of constant returns in general equilibrium theory. More specifically, few if any modern economists seem concerned with the implications of the reintroduction of increasing returns for either the theory of distribution or the theory of competitive equilibrium. We address these issues in this article. [End Page 512]
We shall first, in section 1, lay out several alternative conceptualizations of increasing returns as these have variously appeared in recent works and then concentrate our attention on what we have called “generalized increasing returns”— the phenomenon that seems to be most descriptive of Smith’s basic idea.
In section 2 we carefully examine the relationship between generalized increasing returns and the theory of distribution, especially with reference to the “adding up” problem that so troubled the neoclassical discoverers of marginal productivity and, in particular, to the application of Euler’s theorem. Here we suggest that generalized increasing returns, properly introduced and understood, need not generate results that...