| Many economic systems can be described as two-sided matching markets, with labor and marriage markets being the leading examples. Although the theoretic work can be dated back to 1960's, little is known about estimating the structural two-sided matching models without transfer. This monograph contributes to the empirical literature of matching by providing several new results in identification, estimation, computation and testable implications, under different behavioral and sampling assumptions.;The first chapter focuses on the cases when the same set of players is involved in each matching market. The second chapter concerns the identifiability of the utility functions over partner's characteristics using aggregate data. I argue that the deferred-acceptance algorithm can be represented as a special demand/supply system. Such result leads to a novel IV-type identification argument, a 2-step minimum distance type estimator, and a fast algorithm to compute the implied joint distribution of characteristics of married couples without simulation. In the third chapter I consider the inference problem without imposing equilibrium selection.;I show that a set of moment inequalities can be derived from the no-blocking-pair condition without solving the game. In the last chapter I use the data of Current Population Survey to estimate the utility functions over spouse's education level. I find that men care more about the similarity in spouse's education than women do. Meanwhile, women's education level is becoming more attractive to men compared to 20 years ago. Moreover, it is well known that marriage market is one of the key considerations of education choice, but such endogeneity problem in empirical two-sided matching models is usually assumed away. Therefore, I develop a structural pre-marital educational choice model that accounts for marriage market prospect. I use the estimated preference to compute the "return to schooling within marriage", and estimate the cost distributions of education. By shifting the cost distribution of women via the lump-sum tuition subsidy, I find that the college attendance rate increases for both women and men. Moreover, the new policy creates more college-educated couples, while the correlation of the joint distribution of education of married couples remains unchanged. |
