The Losses from Integration in Matching Markets can be Large

TL;DR

This paper demonstrates that integrating two-sided matching markets can sometimes lead to significant losses for agents, with worst-case scenarios showing a 37.5% decrease in average spouse rank, despite expected gains.

Contribution

It reveals that worst-case outcomes of market integration can be substantially negative, challenging the assumption that integration always benefits participants.

Findings

Worst-case losses can reach 37.5% decrease in spouse rank

Integration can produce negative outcomes despite expected gains

Large potential losses highlight risks in market integration

Abstract

Although the integration of two-sided matching markets using stable mechanisms generates expected gains from integration, I show that there are worst-case scenarios in which these are negative. The losses from integration can be large enough that the average rank of an agent's spouse decreases by 37.5% of the length of their preference list in any stable matching mechanism.