The general equilibrium models of the period before the 1950s were based on individual maximization when explaining the behavior of economic agents, but they employed a quite different, more dynamic and non-optimizing, framework to explain the behavior of competitive markets. This is quite different from the recent literature where general equilibrium is simply the rational action of a single representative agent. The differences are not only mathematical but also represent fundamental methodological as well as political-economic differences. Understanding both why the earlier generation of economists resisted this characterization of competitive markets (despite the fact that it was mathematically more convenient) and how and why the later generation came to embrace it, is an important prerequisite for understanding one of the key difficulties with contemporary theory.
Agents and Markets: The Representative Agent in Mid- vs. Late-Twentieth Century Economics