Challenging Lucas: From Overlapping Generations to Infinite-Lived Agent Models

Abstract

The canonical history of macroeconomics, one of the rival schools of thought and the great economists, gives Robert Lucas a prominent role in shaping the recent developments in the area. According to it, his followers were initially split into two camps, the “real business cycle” theorists with models of efficient fluctuations, and the “new-Keynesians” with models in which fluctuations are costly, and the government has a role to play, due to departures from the competitive equilibrium (such as nominal rigidities and imperfect competition). Later on, a consensus view emerged (the so-called new neoclassical synthesis), based on the dynamic stochastic general equilibrium (DSGE) model, which combines elements of the models developed by economists of those two groups. However, this account misses critical developments, as already pointed out by Cherrier and Saidi (2015). As a reaction to Lucas’s 1972 policy ineffectiveness results, based on an overlapping generations (OLG) model, a group of macroeconomists realized that a competitive OLG model may have a continuum of equilibria and that this indeterminacy justified government intervention for competitive cycles that emerged even in deterministic models. We can identify here two distinct, but related, groups: one of the deterministic cycles of David Gale, David Cass, and Jean-Michel Grandmont, and another of the stochastic models and sunspots of Karl Shell, Roger Guesnerie, Roger Farmer and Costas Azariadis (Lucas’s PhD student). Here, the OLG was the workhorse model. Following from these works, a number of authors, including Michael Woodford, argued that similar results could occur in models with infinitely lived agents when there are various kinds of market imperfections. With such generalization, some of these macroeconomists saw that once these imperfections are introduced, nothing important for business cycle modeling was lost and they could therefore leave the OLG model aside as a model of business fluctuations, to the dismay of authors such as Grandmont, Robert Solow and Frank Hahn. In this paper, we scrutinize the differences between the deterministic cycles and sunspot groups and explore the many efforts of building a dynamic competitive business cycle model that implies a role for the government to play. We then assess the transformation process that took place in the late 1980s when several macroeconomists switched from OLG to infinitelived agents models with imperfections that eventually became central to the DSGE literature. With this we hope to shed more light on the origins of new neoclassical synthesis.