Policy and Practice Impacts on School Choice
My research on assignment of students to schools, first in the context of centralized student admissions in Balinski and Sönmez (1999), and subsequently in the context of public school choice in Abdulkadiroğlu and Sönmez (2003) started bearing its fruits. Not only these two papers initiated an exciting literature, major school districts including Boston Public Schools, Denver Public Schools, and New York City Public High Schools have already adopted versions of student-optimal stable mechanism (SOSM) advocated in these two papers.
- Here is a 9/12/2003 dated Boston Globe story on page B1 titled “School assignment flaws detailed Two economists study problem, offer relief ” on Abdulkadiroğlu and Sönmez (2003).
Interaction with Boston Public Schools (BPS):
Following the Boston Globe story, I contacted Boston Public Schools (BPS) to alert them that their assignment mechanism is highly vulnerable to “gaming” and that there are plausible alternatives. While skeptical at the beginning, BPS agreed to an empirical study of their assignment mechanism that I conducted with Atila Abdulkadiroğlu, Parag Pathak and Alvin Roth. After two years of intense discussions we were able to convince them and on 07/20/2005 BPS gave up the “Boston mechanism” and adopted Gale-Shapley student-optimal stable mechanism (SOSM) starting with School Year 2006-2007. See the 05/11/2005 dated presentation of the BPS Strategic Planning Team recommending this reform.
2013 Update: Since 1999, BPS policy had been giving walk-zone applicants priority in half of the seats at each school. In Dur, Kominers, Pathak and Sönmez (2013), we have shown that the intended walk-zone priority was, to a large extent, eliminated due to the way school priorities are implemented. Following a testimony by me and Parag Pathak before the School Committee in February 2013, BPS completely eliminated walk-zone priority to bring the practice in line with stated policy. See the 03/13/2013 dated memo by Superintendent Carol Johnson recommending elimination of the walk-zone priority.
The Collapse of NYC School Assignment Mechanism and the Subsequent Reform:
While Abdulkadiroğlu and Sönmez (2003) was in press, the high school assignment mechanism in New York City “collapsed” and in May 2003 the Director of Strategic Planning at the New York City Dept. of Education (NYC DOE) contacted Alvin Roth for advice on a centralized assignment mechanism. Following his recommendation NYC DOE has adopted a version of the SOSM, advocated earlier by Abdulkadiroğlu and Sönmez (2003), for allocation of ninth graders to public high schools. Efficiency and incentives considerations have played a central role in this decision (see 11/3/2003 dated letter to the editor titled High School Admissions in New York Times). In May 2004 NYC DOE announced that the transition resulted in a significant improvement of the system.
Balinski and Sönmez (1999) vs. Abdulkadiroğlu and Sönmez (2003):
The major difference between Balinski and Sönmez (1999) and Abdulkadiroğlu and Sönmez (2003) is the following: While student priorities at schools are directly determined by exams in Balinski and Sönmez (1999), they are more flexible and can depend on factors such as proximity to school in Abdulkadiroğlu and Sönmez (2003). Therefore while fully respecting priorities is needed for the former model, there is more flexibility for the latter. This is important because fully respecting priorities is not always compatible with full efficiency. Hence, while SOSM is the only mechanism advocated in Balinski and Sönmez (1999), Abdulkadiroğlu and Sönmez (2003) advocates SOSM along with the top trading cycles mechanism (TTC).
While SOSM has so far done better and is already adopted by major school districts, TTC also generated a lot of interest. For example, prior to adoption of the SOSM, BPS Student Assignment Task Force recommended the adoption of TTC in their Task Force Report. More recently, my partners in BPS school choice reform along with Clayton Featherstone and Muriel Niederle have been advocating TTC to San Francisco United School District (SFUSD). (They refer TTC as assignment with transfer mechanism.) Thanks to them, SFUSD school board unanimously voted for the top trading cycles mechanism on March 9th 2010, further justifying the flexibility we offered in Abdulkadiroğlu and Sönmez (2003).
2012 Update: With the assistance of the Institute for Innovation in Public School Choice, New Orleans Recovery School District adopted TTC in 2012. New Orleans became the first school district to adopt TTC!
Adoption of SOSM Beyond the US:
The U.S. is not the only country where SOSM advocated in Balinski and Sönmez (1999) and Abdulkadiroğlu and Sönmez (2003) found practical applications: The Boston mechanism criticized by Abdulkadiroğlu and Sönmez (2003) is banned in England since 2007 on the basis that it is “unfair” (see Pathak and Sönmez 2008) and it forces families to play an admissions game with their future. Since then more than 60 local authorities in England have given up versions of Boston mechanism, and adopted SOSM. As of 2012, all local authorities in England rely on SOSM for admissions to public schools (see Pathak and Sönmez 2013). In addition to England, Hungary has adopted SOSM for secondary school admissions in 2000 (see Biro 2008) and Ghana has adopted the same mechanism for senior high school admissions in 2005 (see Ajayi 2011), among others.
School Choice and 2012 Economics Nobel Prize:
Section 5.2 of the following Scientific Background on the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2012 emphasizes the role of my theoretical and practical work on school choice on 2012 Nobel prize.