Markets as Complex Distributed Networks: Implications for Efficiency and Inequality

In macroeconomic theory, a disproportionate amount of attention has been paid to models with 'global' or random interaction structures in which each agent interacts, or has an equal probability of interacting, with every other agent in the economy. By contrast, recent models have been developed in which the interaction structure is local but perfectly regular (e.g. nearest neighbor interactions on a two-dimensional lattice). Since local interaction structures are often implicitly due to some type of market failure, it is perhaps not surprising that the latter models may generate inefficient aggregate outcomes as well as persistent inequality. In this paper, it is argued that actual market economies tend to have network structures that lie somewhere between the two extremes. That is, most agents interact locally with only a few exchange partners (e.g. their local grocery store, their employer), whereas relatively few agents (e.g., chain stores, multinationals, governments) engage in 'global' trade. Starting from a perfectly 'localized' economy, it is shown that adding only a few global traders greatly enhances the efficiency of the network but not by enough to eliminate persistent inequality. This relation between efficiency and inequality is analyzed in the light of complexity theory.