What is it about?
When is it possible to reassign students to schools to improve diversity at schools while improving student welfare? We construct a method to assign students to schools that improves diversity at schools while assigning students to either their initial schools or more-preferred ones. Our method produces efficient allocations and cannot be gamed by students. Our results can be applied to other matching markets. For example, teachers need to be assigned to schools, doctors to hospitals, and workers to firms. We provide a class of distributional objectives, such as diversity, for which our method can be applied. The class is based on a permissive notion of discrete concavity that we introduce.
Featured Image
Photo by Alexis Brown on Unsplash
Why is it important?
Diversity is deemed highly desirable in many environments such as labor markets and school choice. However, there is a lacuna about how to achieve diversity while improving the welfare of the participating agents. We provide a ready-to-use method that can be adopted by policymakers in school districts, worker and student exchange markets, and medical residency programs. Our method can be easily used to address various distributional objectives including those related to diversity. Moreover, the new concavity notion of functions on discrete domains may be of independent interest, as it is reminiscent of, but different from, standard concepts in discrete convex analysis.
Perspectives
It was a pleasure to work on this paper with coauthors whom I have known for decades. We hope that our work will be helpful in achieving diversity in different contexts. We would be more than happy to help!
M. Bumin Yenmez
Washington University in Saint Louis
I am happy that my capable coauthors and I have the opportunity to contribute to developing effective methods to enhance diversity in various sectors. It's my aspiration that our methods will transition from theory to practical implementation.
Isa Hafalir
University of Technology Sydney
The best part of the project is for me to learn from wonderful coauthors. As a reader of the paper you cannot experience that part of the project :-), but I'm cautiously optimistic that you enjoy the paper nevertheless! The paper provides a class of methods that could be applied for matching authority who have difficult tradeoffs between many goals, including efficiency and distributional objectives. Comments are greatly appreciated, both from researchers and practitioners.
Fuhito Kojima
Read the Original
This page is a summary of: Efficient Market Design with Distributional Objectives, July 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3580507.3597716.
You can read the full text:
Resources
Contributors
The following have contributed to this page
