Publication:
Revealed preference tests of stability in aggregate matching with single-peaked preferences (with Andrés Carvajal), Economic Theory (2026), [PDF][https://www.doi.org/10.1007/s00199-026-01711-6]
Abstract: This paper studies the rationalizability of two-sided aggregate matchings with non-transferrable utility by single-peaked preferences, when agents' preferences are unobserved. This builds on Echenique, Lee, Shum, and Yenmez (2013)’s framework, which considers unrestricted preference domains. In contrast, I restrict preferences to be single-peaked. This is motivated by the fact that many observable characteristics of agents, such as age and income, are naturally ordered. A matching is rationalizable by single-peaked preferences if there exists a single-peaked preference profile such that the observed matching is stable. I find a characterization for such matchings and extend the result to a multidimensional type setting.
Working Papers:
Matching, Unanticipated Experiences, Divorce, Flirting, Rematching, Etc. (with Burkhard Schipper), R&R at Journal of Economic Theory [PDF]
Abstract: We study dynamic decentralized two-sided matching where players’ preferences evolve due to unanticipated experiences. Stability requires no pairwise common belief in blocking, but unanticipated experiences can destabilize matchings. We show the existence of self-confirming outcomes that are stable and do not lead to unanticipated experiences. We propose a decentralized matching process that prioritizes mutual optimal blocking pairs with probability 1−ε and picks any other optimal blocking pair with ε, representing market friction. Frictions are necessary for convergence to self-confirming stable outcomes even without unawareness. We extend our results by allowing communication and show convergence to flirt-proof self-confirming outcomes.
Satisficing Matching (with Andrés Carvajal and Ester Camiña), Submitted [Link]
We bring the notion of “satisficing” to matching theory and develop a weaker notion of stability. Satisficing behavior is an alternative to maximizing behavior, where instead of always going for the best alternative, ``satisficers'' settle on something that is “good enough”. We say that a matching is satisficing if every agent is matched to an achievable partner. (An agent is achievable to another if they are matched in some stable matching.) We show that satisficing matchings are Pareto efficient, and we also explore the structure and various properties of the set of satisficing matchings.
Rationalizability of Truth-telling in Stable Mechanisms (with Andrés Carvajal and Ester Camiña), Submitted [Link]
Roth (1982) shows that for any stable matching mechanism, truthful preference revelation fails to constitute a Nash equilibrium of the induced direct revelation game at some preference profile. We establish that, in the same game, truthful revelation is nevertheless a rationalizable strategy for every agent. The argument exhibits, for each side of the market, a Nash equilibrium of the direct revelation game in which that side reports truthfully while the other side declares preferences smitten with its match under the mechanism's selected stable matching; truthful revelation therefore survives the iterated elimination of never-bestresponses. The result complements existing positive incentive results obtained by restricting the strategic environment or by considering large markets, and identifies a sense-weaker than Nash equilibrium but consistent with common knowledge of rationality-in which truthful revelation is robust under any stable mechanism.
Alternative Characterizations of Revealed Preference Tests of Aggregate Matching
We provide alternative characterizations of rationalizable aggregate matching, based on Echenique, Lee, Shum, and Yenmez (2013) and Demuynck and Salman (2022). Under non-transferrable utility, a matching is rationalizable if and only if every sub-market of the partition of the matching market does not have more couple types than individual types. Under transferable utility, a matching is rationalizable if and only if every sub-market of the partition of the matching market has more individual types than couple types. Such alternative conditions make the revealed preference tests more convenient for empirical research on matching. We also discuss other empirical considerations such as the power of the test and potential issues with data binning.
Revealed Stability (with Andrés Carvajal)
We study the testable implications of stable matchings across different subsets of agents in matching markets. We develop conditions that are analogous to SARP and WARP in consumer theory and conditions that are analogous to persistence under contraction and expansion in revealed preference theory of game theory (as in Sprumont (2000)). We obtain a characterization of rationalizable matching correspondence, which also shed light on the structure of stable matchings.
Work in Progress:
Efficient Matchings under Unawareness (with Burkhard Schipper)