Association for Public Policy Analysis & Management

In the early 1990s Georgia’s HOPE Scholarship program jump-started a national trend toward state-funded merit-based financial programs.  Until 2011 eligibility for free tuition, fees and a book allowance at any public university in Georgia was based solely on high school and college grades.  Students were required to earn a 3.0 or better high school GPA to earn the scholarship, and maintain the same GPA in college to keep the scholarship.  The program was financed exclusively with proceeds from the state lottery.

Following the Great Recession Georgia faced stagnant lottery revenue, increasing numbers of eligible students and rising tuition costs, leading to projected deficits in HOPE Scholarship funding.  To address these potential funding shortfalls, Georgia lawmakers in 2011 dramatically increased eligibility requirements for full scholarships and reduced HOPE Scholarship benefits.  A new program, Zell Miller Scholarships, was introduced offering full tuition for students with a minimum 3.7 high school GPA and combined 1200 SAT score.  Students must maintain a 3.3 GPA in college to keep the scholarship.  The previous HOPE Scholarship program was retained for students achieving a 3.0 GPA but benefits were reduced to partial tuition without fees or a book allowance.  Students already in the university system when the changes were announced were not grandfathered in, and eligibility was based on both high school GPA and SAT scores, along with college GPA.  Thus, some students lost partial scholarships while enrolled in college based on their high school grades and test scores.

In this paper we exploit this exogenous change in scholarship eligibility to examine the effects of loss of financial aid on student persistence in college, credit accumulation and likelihood of graduation.  Using student-by-semester data from the Georgia Board of Regents for the 2009-10 cohort, we follow the same students through five years spanning the two scholarship regimes and employ a regression-discontinuity design to compare outcomes for students just above and just below the new eligibility thresholds.

This design has several advantages over previous research using R-D methods to examine scholarship loss.  First, the eligibility change was unexpected and retroactive; students near the cutoff were unable to change their behavior to maintain the scholarship, thereby minimizing potential selection issues.  Second, because of the unexpected implementation we will not face potential bunching of students just above the eligibility threshold as found in other studies employing similar methods. Third, we have rich data on financial aid eligibility and receipt that allow us to examine heterogeneous effects by income and other student characteristics, and to explore substitution of other financial aid and loans in response to reductions in scholarship aid.

The results of our analyses will shed light on the extent to which unexpected financial aid loss affects students who are already achieving some degree of success in college.  Additionally, we will be able to examine how Georgia’s new eligibility criteria affect the profile of students receiving full and partial merit-based scholarships.