Complexity, causality, and expectations: Why macroeconomics needs a program of empirical microfoundations

The article engages in the well-known debate about the need to ground macroeconomics in a program of microfoundations. In this debate, new classical and new Keynesian macroeconomists argue for an eliminative program of microfoundations (which reduces macroeconomics completely to general equilibrium theory), agent-based macroeconomists for a program of “genuine” microfoundations (which models the macroeconomy as a complex system), and macroeconometricians for the possibility of conducting causal inference independently of any program of microfoundations. The article argues for a program of empirical microfoundations: for a program that has only recently begun to gather momentum and uses surveys to study the formation of individual expectations (their type-level causes, the degree to which these causes are relevant, and the potentially changing rules that govern their relevance). The article argues, more specifically, that the macroeconomy is a complex system, that the ultimate purpose of the study of macroeconomics is causal inference (or policy analysis), and that in the face of a complex macroeconomy, causal inference won’t be successful unless it is underpinned by the program of empirical microfoundations. The article also argues that this program should replace the eliminative program of neoclassical and neo-Keynesian macroeconomists, supplement the program of “genuine” microfoundations of agent-based macroeconomists and support causal inference in macroeconometrics.