Matching Models of the Labor and Marriage Markets

This dissertation studies equilibrium matching patterns in the marriage and labor markets when agents have heterogeneous preferences over partner\u27s traits. Economic analysis of matching has previously focused largely on equilibria obtained when agents have homogeneous preferences. While this assumption has made analysis simple, it is highly restrictive. It seems obvious that different industrial sectors prefer different groups of workers (since some are more productive in one sector, while others are in other sectors). It is equally obvious that individuals do not agree completely on desirable marriage partners. The goal of this research is to study the impact that heterogeneous preferences have on the relative performance of agents in long-term partnerships. In the first chapter, I characterize the marriage market equilibrium, when agents have heterogeneous tastes for the characteristics of their potential partners, and when they cannot invest to change their own traits in any way. The main finding is that, when it is difficult for agents to meet potential partners, individuals with general appeal (i.e., those that are fairly desirable along several dimensions) find better matches than individuals with specialized appeal (i.e., those that are undesirable along some dimensions but very desirable along others). As frictions decrease, those with specialized appeal do better than those with general appeal. If one were to interpret the extent of the market as the frequency with which one expects to meet potential partners, this result supports Adam Smith\u27s insight that the returns to specialization are limited by the extent of the market. In the second chapter, I explicitly allow agents to invest and change the characteristics they are born with, and look at whether agents choose to specialize as frictions decrease. I show that agents diversify (invest to become fairly desirable along many dimensions) when frictions are large; and specialize (invest to become very desirable along one dimension, at the cost of neglecting other traits) as frictions decrease. In the last chapter I look at the core of a static marriage market game. I prove that the core is non-empty, and is unique to the extent that any given agent gets the same pay-off in every core allocation. This pay-off is the same as what they would receive in the limit of the dynamic games considered in the first chapter, as frictions disappear