Abstract

This paper contributes to closing the gap between collective models and matching models of the marriage market. It builds on previous work by Galichon, Kominers, and Weber (2018) and clarify the connection between these two literatures. In particular, I characterize the class of collective models that can be embedded in a matching framework with Imperfectly Transferable Utility (ITU). Estimation of these models depends crucially on the computation of the so-called "distance function", which, in most cases, has no analytic form. I provide a fast computation method that works in any case.? I also discuss applications.