Abstract
Starting from the path-breaking contribution by Gale and Shapley (1962), therehas been an increasing interest in the development of algorithms designed to solve
non-transferable utility matching problems. Matching theory and market design
have contributed to organize several real-world markets. A typical example are
educational markets, which offer fundamental applications, such as school choice,
college admissions, course allocation and teacher assignment.
This thesis studies the problem of reassigning tenured teachers to teaching positions in a centralized labor market. In particular, we address some market design
issues arising in the current Italian teacher assignment system.
Chapter 2 (with Battal Doğan) shows that the current mechanism in Italy suffers
from systematic violations of teachers’ priority rights, resulting from the way ties
in teacher preferences are resolved. To address this issue, we propose a novel
mechanism that is individually rational, non-wasteful, strategy-proof, eliminates
justified envy, and Pareto improves over the benchmark deferred acceptance with
simple tie-breaking.
Chapter 3 studies the design of commitment constraints in the Italian teacher
assignment. This type of constraints is used as a retention policy instrument
in order to retain teachers in their current positions, and therefore to promote
teaching continuity. We show that the current system gives rise to a series of shortcomings due to a fundamental restriction of the strategy space, which prevents
teachers from fully signalling their preferences. In particular, teachers are forced
to participate in a multi-item lottery over schools if they want to request a single
school without commitment to stay. To address these issues we propose a market
design based on the matching with contracts framework, where we restore the
ii
complete strategy space, without artificially introducing additional contractual
possibilities. Our proposed mechanism removes the shortcomings while satisfying
desirable properties, such as strategy-proofness, non-wastefulness, elimination of
justified envy, respecting improvements, and a weaker form of Pareto efficiency.
Chapter 4 evaluates the effects of revealing partial information about available
vacancies on the distribution of teachers across schools. Motivated by the current
information disclosure policy in the Italian teacher labor market, we provide some
insights on how to design informational interventions in centralized assignment
systems. To conduct the analysis, we construct a novel dataset by collecting public
announcements on teacher retirements, which are a relevant source of information about the next-year expected vacancies. Using a generalized difference-indifferences strategy, we find evidence of within-province distributional effects
Date of Award | 1 Oct 2024 |
---|---|
Original language | English |
Awarding Institution |
|
Supervisor | Simon M Burgess (Supervisor) & Hans Henrik Sievertsen (Supervisor) |
Keywords
- Education
- Market Design
- Teacher Assignment