Strategic Waiting in Centralized Matching: Daycare Assignment

In centralized assignment mechanisms, unassigned participants are often placed on waitlists to improve their chances in future rounds. However, I demonstrate that this practice may bring unintended welfare consequences on the participants, using data from the Japanese daycare system. The prioritization of waitlisted applicants introduces a dynamic incentive for applicants to manipulate their priority by strategically choosing to be waitlisted to secure positions at more selective daycares. I show that 30.0% of applicants do not list safety options when they initially apply and that those who benefit from additional priority are 8.1 percentage points more likely to avoid listing safety options than those who do not. Given the prevalence of such strategic waiting, I estimate a structural model of daycare choice that extends Agarwal and Somaini (2018) to a two-period model allowing for reapplication. By simulating a scenario without waitlist priority, I find that the current priority functions as a redistributive mechanism: through its abolition, early starters (age 0) experience a 34.1% decrease in welfare and a 1.7 percentage point increase in the likelihood of being waitlisted, whereas late starters (age 1) experience a 29.0% increase in welfare and a 6.9 percentage point decrease in the likelihood of being waitlisted.