From signaling to interviews in random matching markets

Maxwell Allman, Itai Ashlagi, Amin Saberi, Sophie H. Yu

Abstract: In many two-sided labor markets, interviews are conducted before matches are formed. An increase in the number of interviews in the market for medical residencies raised the demand for signaling mechanisms, in which applicants can send a limited number of signals to communicate interest. We study the role of signaling mechanisms in reducing the number of interviews in centralized random matching markets with post-interview shocks. For the market to clear we focus on interim stability, which extends the notion of stability to ensure that agents do not regret not interviewing with each other. A matching is almost interim stable if it is interim stable after removing a vanishingly small fraction of agents. We first study signaling mechanisms in random matching markets when agents on the short side, long side, or both sides signal their top~$d$ preferred partners. Interviews graphs are formed by including all pairs where at least one party has signaled the other. We show that when $d = \omega(1)$, short-side signaling leads to almost interim stable matchings. Long-side signaling is only effective when the market is almost balanced. Conversely, when the interview shocks are negligible and $d = o(\log n)$, both-side signaling fails to achieve almost interim stability. For larger $d \ge \Omega(\log^2 n)$, short-side signaling achieves perfect interim stability, while long-side signaling fails in imbalanced markets. We build on our findings to propose a signaling mechanism for multi-tiered random markets. Our analysis identifies conditions under which signaling mechanisms are incentive compatible. A technical contribution is the analysis of a message-passing algorithm that efficiently determines interim stability and matching outcomes by leveraging local neighborhood structures.