Tuesday, November 27, 2012

Scott Kominers on designing matching markets for diversity

Scott Kominers, who finished his Ph.D. in 2011, is on the market this year after a very productive two year postdoc at Chicago.

He has written a lot of papers, many on the frontiers of matching and market design. The one he has designated as his primary  job market paper, joint with Tayfun Sonmez, is Designing for Diversity in Matching.

The idea is that when a school has many places, some intended to facilitate the enrollment of different kinds of students, then a deferred acceptance algorithm can be implemented in a nuanced way, in which different positions may express different preferences. It turns out that this can make a big difference in how slots are filled.

"Abstract: To encourage diversity, schools often "reserve" some slots for students of specific types. Students only care about their school assignments and contractual terms like tuition, and hence are indifferent among slots within a school. Because these indifferences can be resolved in multiple ways, they present an opportunity for novel market design.

"We introduce a two-sided, many-to-one matching with contracts model in which agents with unit demand match to branches, which may have multiple slots available to accept contracts. Each slot has its own linear priority order over contracts; a branch chooses contracts by filling its slots sequentially. We demonstrate that in these matching markets with slot-specific priorities, branches' choice functions may not satisfy the substitutability conditions typically crucial for matching with contracts. Despite this complication, we are able to show that stable outcomes exist in this framework and can be found by a cumulative offer mechanism that is strategy-proof and respects unambiguous improvements in priority. Our results provide insight into the design of transparent affirmative action matching mechanisms, and show the value of a seemingly ad hoc administrative decision in the United States Military Academy's branch-of-choice program."

Scott's work is well worth following, and he's on the market, so you could hire him this year.

1 comment:

Hongkai Zhang said...

We recently discussed this paper at the theory reading group at MIT - Yusuke suggested this paper and presented it. I still feel like I attended market design course with Scott not long ago. Great paper! Go Scott~