University of Rochester

Choice Assignments

Two Rochester-trained economists design a better way to match high schoolers in some of the nation’s biggest districts with their preferred public school. By David McKay Wilson
Tayfun Sönmez ’95 (PhD)
IMPACT: “I always wished to have a real impact in people’s lives,” says Sönmez, an economist who is joining the faculty at Boston College this fall.

Choosing the right public high school in some American cities can be as daunting as the most rigorous standardized test.

In New York City, 30,000 of almost 100,000 students two years ago didn’t get into any of their five top choices in the admissions process. They ended up assigned to schools that their parents didn’t want them to attend.

In Boston, parents often resorted to strategies of choosing safe schools that weren’t their top choices so they wouldn’t get shut out of schools they considered as just O.K.

Enter Tayfun Sönmez ’95 (PhD) and Atila Abdulkadiroglu ’00 (PhD), two Turkish-born economists trained at Rochester, who are emerging as leaders in a branch of economics that uses mathematical models to parcel out prized resources in a fair and effective manner.

The two, who have collaborated on several projects, have designed more effective systems to allocate places in urban schools.

New York City schools first used Abdulkadiroglu’s model for selecting high schools in 2003. Last year, only 3,000 of 100,000 didn’t get their preferred schools.

“We designed it from scratch, and using rigorous techniques, we were able to solve this problem for all those kids,” says Abdulkadiroglu, now on the economics faculty at Columbia University. “That makes me happy.”

For David Bloomfield, vice president of New York Citywide Council High Schools, the new admissions process provided a shot in the arm for the nation’s largest school district, which has been working hard to reform its high schools.

“The new system really worked,” says Bloomfield. “There’s a growing confidence among parents that it will help kids and will be fair.”

Atila Abdulkadiroglu ’00 (PhD)
CHOICES: “It was a choice program that wasn’t giving any choice.” Abdulkadiroglu says of Boston’s former system of placing students.

The Boston School Committee was expected to adopt their proposal to reform that city’s assignment process last summer, eliminating the need for parents to spend hours plotting strategy and doing away with a perceived penalty students faced for ranking schools in their true order of preference.

Solving such “real-world” problems has long been an interest of Sönmez and Abdulkadiroglu. Since their first collaboration in the late 1990s, the two have explored ways that economic models can be used to analyze complicated questions involving the allocation of kidneys for transplant, how best to assign college dorm rooms, and how to improve the school assignment process.

Abdulkadiroglu, who worked with Sönmez on the Boston public schools project, is an associate professor of economics at Columbia, where he teaches microeconomics. A research fellow with the Alfred P. Sloan Foundation for the past two years, he recently received a five-year, $400,000 grant from the National Science Foundation to support his research.

Sönmez, a research fellow in 2004–05 at the Harvard Business School, in September will join the faculty at Boston College, where he will teach economics and continue his research. That includes managing a kidney exchange system he has established for the New England Organ Bank, which will link donors to patients with kidney disease. Last May, Sönmez received the Medal of Science from the Turkish Academy of Sciences for his work on a kidney exchange and school choice.

A Tool for Organ Transplants

Like policymakers at urban schools, medical professionals at the New England Organ Bank needed a better system to allocate a scarce resource, in this case, working kidneys, either from a cadaver or a living donor.

Concerned with addressing the issue of patients with kidney disease who were waiting too long for kidney transplants, they contacted Tayfun Sönmez after learning about his work for the Boston school system.

In 2003, there were about 60,000 people on the waiting list for kidney transplants in the United States. About 8,700 received transplants from cadavers while 6,500 received kidneys from living donors, who can remain healthy with just one kidney.

But 3,500 patients died waiting for a transplant, according to a study by Sönmez and economists Alvin Roth and Utku Unver.

While it’s illegal to buy or sell organs, medical ethicists have approved exchanges between donor-patient pairs. Let’s say two donors can’t give a kidney to their intended recipients because their blood types are incompatible, but their blood types match up with another patient whose volunteer donor is incompatible. The donor relationships are rearranged so that each donor gives to a compatible patient in surgeries that occur simultaneously.

As of December 2004, only five such two-way exchanges had occurred in New England’s 14 transplant centers, mainly due to the lack of a database of incompatible patient-donor pairs.

Sönmez, who is also working with transplant officials in Ohio, has set up a system in New England in which incompatible donors can register in hopes that as other donors sign up finding matches will be easier. Transplant experts predict that 2,000 more kidney transplants a year could be performed if such an exchange system was established across the country.

“Until recently, it was all done in a haphazard way,” Sönmez says. “We bring tools that will optimize the exchanges. By organizing direct exchange between incompatible pairs, you’ll save as many people as possible.”

—David McKay Wilson

Although they shared similar academic histories and interests, the two didn’t know each other until their paths crossed at Rochester. Both grew up in Turkey, graduated from Ankara Science High School, one of the nation’s preeminent secondary schools, which had been established with a grant from the Ford Foundation in the 1960s. They earned bachelor’s degrees in electrical engineering at Bilkent University in Ankara, then crossed the Atlantic to come to the University. Separated by five years, they each studied economics, earning their doctoral degrees.

When Sönmez began teaching at the University of Michigan after graduation, he called his compatriot for consultation on a thorny mathematical problem. It led to their first academic collaboration.

“The first article we wrote together was pure mathematical theory,” says Abdulkadiroglu. “I remember telling him, ‘This is interesting, but I want to work on real problems.’ Solving problems on paper wasn’t enough.”

For their second paper, they looked at the problem of allocating campus housing to college students. When Sönmez was presenting a paper at Carnegie Mellon University, a scholar at the seminar suggested they look at the problems associated with school-choice systems in urban school districts.

This was a real-world problem with broad ramifications. Since the uproar over school desegregation in the 1970s and 1980s, educators and policymakers in cities like Boston, Denver, Minneapolis, Seattle, and Tampa have developed systems to assign students to schools.

After their study on school choice was reported by the Boston Globe, the economists were invited to discuss their findings with Boston school officials, who subsequently asked them to look at alternatives to a current system that was adopted in 1987 when the U.S. Court of Appeals freed the school system from a federal judge’s busing order. The court allowed the district to adopt a choice-based assignment plan for more than 60,000 students attending almost 140 schools.

The system requires Boston parents to select a new school in kindergarten, first grade, sixth grade, and ninth grade. Priorities are given at each school to students who have a sibling there or live within walking distance.

Under that system, schools considered all students who ranked their school as the top choice, and allocated the seats first to students who had a priority. Then schools with room selected those who ranked the school number two, and so on. If students received none of their choices, they would be assigned to the school closest to home.

Parents learned that the best Boston schools filled up quickly, so it made little sense to rank the best schools as their top choice. School officials even encouraged parents to choose less popular schools. A study by Sönmez and Abdulkadiroglu found that many parents who didn’t select a safety school ended up with worse assignments than they otherwise would have had.

“It was a choice program that wasn’t giving any choice,” says Abdulkadiroglu. “It was very risky to choose your top choice. You had to be careful and choose whatever you thought you could get. With the new system, the pressure on the shoulders of the parents should be relieved.”

In May, the superintendent of Boston’s schools recommended a system proposed by the economists. The new system looks at everyone’s first choice, temporarily assigns them to it, then places students at the best school they listed for which they had a high enough priority. The economists predict that families will be more inclined to list all schools they want, not just the one they think they are likely to gain admittance.

“They have been extremely helpful,” says Jonathan Palumbo, a spokesman for the Boston school system. “When you’re talking about an issue as complicated as the algorithm for student assignment, people’s eyes can easily glaze over. They’ve been able to translate it all into English and provide us with comparisons to what other cities are doing.”

Giving people better choices, Sönmez says, has always been the goal.

“What has happened over these past few years was better than my wildest dreams. I always wished to have real impact in people’s lives.”

David McKay Wilson is a New York–based freelance writer.