In economic theory and econometrics, the term heterogeneity refers to differences across the units being studied. For example, a macroeconomic model in which consumers are assumed to differ from one another is said to have heterogeneous agents, which is contrasted with the case of a representative agent model in which all consumers are assumed to be identical.
In econometrics, statistical inferences may be erroneous if, in addition to the observed variables under study, there exist other relevant variables that are unobserved, but correlated with the observed variables.
Methods for obtaining valid statistical inferences in the presence of unobserved heterogeneity include the instrumental variables method; multilevel models, including fixed effects and random effects models; and the Heckman correction for selection bias.
Economic models are often simplified by assuming that all agents (decision makers) are identical; this is often called the representative agent assumption. However, some questions in economic theory cannot be accurately addressed without considering differences across agents, requiring a heterogeneous agent model.
How to solve a heterogeneous agent model depends on the assumptions that are made about the expectations of the agents in the model. Broadly speaking, models with heterogeneous agents fall into the category of agent-based computational economics (ACE) if the agents have adaptive expectations, or into the category of (DSGE) if the agents have rational expectations. DSGE models with heterogeneneous agents are especially difficult to solve, and have only recently become a widespread topic of research; most early DSGE research instead focused on representative agent models.