Lisa BOTBOL will defend her thesis on Wednesday, July 10 at 2pm, Auditorium 3, Building TSE
"Three Essays on Matching Markets"
Supervisors: Thierry MAGNAC and Ana GAZMURI
To attend the conference, please contact the secretariat Christelle Fotso Tatchum
Memberships are:
- Thierry MAGNAC : Professor in Economics, TSE - University of Toulouse Capitole Supervisor
- Ana GAZMURI : Assistant Professor in Economics, TSE - University of Toulouse Capitole Co-Supervisor
- Guillaume CHAPELLE : Assistant Professor, University of CY Cergy Paris Rapporteur
- Julien GRENET : Senior Researcher, CNRS/PSE Rapporteur
- Laurent GOBILLON : Senior Researcher, CNRS/PSE Examinateur
Abstract
This thesis explores the design of matching markets, where resources are allocated by public bodies without prices as a market-clearing mechanism, focusing on school admissions and social housing assignments. I investigate how the design of these mechanisms impacts socioeconomic inequalities and household welfare. To this effect, I study the decision-making of applicants and authorities in the allocation mechanism. I use reduced form and structural econometrics applied to data on applications and assignments. The first two chapters of the thesis also develop counterfactual strategies to replicate the allocations under a modified mechanism, which allows for direct comparison of different designs. The third chapter sheds light on potential frictions to the allocation to inform decisions on their design.
The first essay, co-authored with my advisor Ana Gazmuri, investigates the effects of a 2016 Chilean reform that centralized primary school applications and introduced priorities for low-socioeconomic status (SES) students so as to reduce education inequalities. Using a difference-in-differences framework, we quantify the reform's impact on access to good quality schools across SES. The findings indicate that the reform did not achieve its goal. To understand this outcome, we estimate a demand model based on students' rank-ordered application lists. The analysis reveals significant heterogeneity in preferences across SES groups, partly explaining why counterfactual analysis shows that increased affirmative action levels do not significantly reduce segregation. The study also highlights that students often rank too few schools, limiting the reform's effectiveness by leaving many unassigned and ultimately enrolled in underdemanded schools. Counterfactual simulations suggest that if students ranked more schools, the reform could better achieve its objectives.
The second essay delves into the allocation of social housing in France, using a dynamic model and a comprehensive dataset of French social housing applications. This paper aims to compare applicant welfare under different social housing allocation rules. In order to simulate applicants’ behavior under modified rules, the empirical strategy focuses on modelling and estimating applicant preferences by separately identifying them from expectations of future offers. Estimates of applicants’ utility surplus suggest that the current system favors households with French nationality while disadvantaging precarious households such as single mothers. Counterfactual analysis shows that better targeting low-income households would significantly improve welfare, when mechanisms solely based on applicant waiting time, like first-come-first-serve, are found to be welfare-reducing.
The third essay also examines the French social housing context, focusing on whether social housing is allocated opportunistically around municipal elections. By analyzing data on social housing assignments and leveraging the timing of municipal elections, the study investigates whether the demographic composition of assignees changes in response to electoral cycles. The empirical strategy takes advantage of the dual administration of social housing in France, where some units are managed nationally and remain unaffected by local elections. The results indicate little to no evidence of opportunistic behavior, suggesting that social housing allocations are influenced more by the political leaning of the municipality's mayor rather than by a political cycle. This finding challenges the notion of widespread electoral manipulation when decentralizing the allocation of local public goods.
The three essays collectively highlight the critical role of allocation mechanisms in shaping outcomes in matching markets. Conclusions shed light on desirable design features for mechanisms to improve equality and welfare. This insight offers recommendations for policymakers aiming to reform matching markets.
