Matching in Perfect and Imperfect Worlds

We study frictionless matching models in large production economies. We provide necessary and sufficient conditions for segregation and for positive assortative matching. These conditions focus on the relationship between what we call the segregation payoff- and the feasible set for a pair of types. Our approach is useful for clarifying differences in the behavior of models in the literature. it also provides a basis for understanding the effects of changes in technology or in the severity of market imperfections on equilibrium welfare and matching patterns