Incomplete take-up of federal aid remains a puzzle. This paper provides the first cross-state evidence on universal FAFSA policies, using a synthetic difference-in-differences design and administrative data from the Office of Federal Student Aid. Strict policies increase FAFSA submission rates by 7.8 percentage points, federal aid receipt by 1.8 percentage points, and conditional-on-filing college enrollment by 2.9 percentage points, with effects concentrated among low-income students. Students also apply to the FAFSA approximately two weeks earlier. A MVPF calculation yields a range of 2.2 to 6.8, placing universal FAFSA policies in the upper range of MVPFs reported for college access programs.

At-scale behavioral nudges on take-up, voter mobilization, and benefit enrollment have repeatedly delivered null effects. We document one that works, and for whom it works. In a 9.91-million-person block-randomized experiment, a U.S. Department of Education FAFSA renewal email increased FAFSA renewal by 0.98 percentage points (a 2.6% increase, roughly 95,000 additional submissions) and college re-enrollment by 0.31 percentage points (roughly 30,000 additional enrollments) at near-zero marginal cost. The response is sharply selective: approximately 77% of the email-induced renewal effect is for students who had been enrolled full-time, falling to 28% for non-full-time enrollees and to approximately zero for students who had not been enrolled. Unlike other recent at-scale nudge experiments, this government-run campaign moved both paperwork and enrollment, but the enrollment response is concentrated among students already attached to college.

In 2011, Pell Grants faced a $18 billion shortfall. Congress responded with the Consolidated Appropriations Act, 2012 which implemented four major changes to Pell Grant eligibility. This paper explores two of these policies and how they affected individual behavior. These two policies both led to small reductions in Pell Grant distributions in absolute terms, but on two different margins of the Pell distribution. The first policy eliminated Pell Grants for individuals on the margin of receiving the minimum amount of Pell Grants. The second policy reduced Pell Grants for individuals on the margin of receiving the maximum amount of Pell Grants. In the first policy, a loss of about $305 in Pell Grants led to a decrease of 9.3 percentage points in full-time enrollment rates. Individuals at this margin borrowed almost five times the amount they lost in Pell Grants. In the second policy, individuals who lost about $151 in Pell Grants did not change their enrollment intensity or borrowing behavior.

This paper documents inequalities in higher education revenue sources by college selectivity from 1987-2015. We find the greatest inequalities exist in federal funding (excluding Pell Grants) and PIE (private gifts, investment returns, and endowment returns) across college selectivity. Both of these funding streams are disproportionately captured by Ivy-Plus universities and colleges at the upper end of the selectivity distribution. We find that the share of federal funding received by Ivy-Plus colleges closely matches inequalities found in the top-end of the income distribution, ranging from 14%-15%. We also examine expenditures by college selectivity and as expected find that federal funding aligns very closely with research expenditure. Interestingly, Highly Selective Public universities tend to be greater net investors in research relative to Ivy-Plus universities. We explore this finding further, by asking whether the marginal federal dollar is more productive at an Ivy-Plus university or a Highly Selective Public university.