Panel Paper: Evaluating the Effects of Universal Place-Based Scholarships on Student Outcomes: The Buffalo Say Yes to Education Program

Saturday, November 5, 2016 : 8:50 AM
Columbia 3 (Washington Hilton)

*Names in bold indicate Presenter

Robert Bifulco1, Ross Rubenstein2 and Hosung Sohn1, (1)Syracuse University, (2)Georgia State University


Place-based scholarship programs award grants for college tuition based on residence in a specific school district or city, rather than on merit or need. In this paper we evaluate the effects of a universal place-based scholarship program, the Say Yes to Education program in Buffalo, NY, on student outcomes in high school, including test scores, course-taking and graduation, and on college matriculation.  The Buffalo program began in 2012 and includes a tiered scholarship component, requiring twelve years of attendance in the Buffalo Public Schools (BPS) for full tuition, and partial tuition scholarships for students spending less than twelve years in the BPS. Scholarships are available at both public in-state universities and almost 100 private institutions ranging from Ivy League universities to small liberal arts colleges. The tiered eligibility component is similar to other notable place-based scholarship programs, such as the Kalamazoo Promise, while the list of participating higher education institutions is broader than other similar programs. Moreover, unlike most other place-based scholarship programs, the Say Yes program includes a wide array of additional services, including summer and after-school programs, health and legal clinics, school social workers and program coordinators, and assistance with college financial aid applications.

The analyses use detailed longitudinal student-level data from the Buffalo Public Schools and from the National Student Clearinghouse to compare students in different cohorts passing through the same school.  We include a range of control variables and school fixed effects to ensure that estimates of the effect of Say Yes are based on comparison of cohorts exposed to Say Yes to earlier cohorts in the same school prior to Say Yes. Because the difference in treatment between cohorts is the entire Say Yes program, the effect estimates from this model reflect not only the impact of the Say Yes scholarship offer but also any effect of the additional services provided by the Say Yes program. The key assumption required to interpret the effect estimates as the causal impact of the Say Yes program is that differences between nearby cohorts in the same grades and schools are essentially random. 

The analyses offer evidence on how broad-based, transparent financial aid programs can affect student outcomes in elementary and secondary school, college matriculation choices after high school, and trade-offs inherent in universal vs. targeted aid programs.