The administration of college financial aid is a key venue through which colleges can affect the likelihood that students will make it to graduation. We investigate the effects of an understudied yet consequential federal student aid policy: Return of Title IV Funds (R2T4). Under R2T4, students "earn" Federal Student Aid over a term or payment period as they remain enrolled--however, aid may be disbursed before or after it is earned. If a student withdraws during a term or payment period, the institution completes a R2T4 calculation, which could require the student or institution to pay back unearned funds to the government. By increasing the financial and administrative burdens for students who want to re-enroll in college, R2T4 and similar policies have often been linked to concerns around transcript withholding, stranded credits, and other issues that may negatively impact the ability of students to complete their degree or acquire stable employment. In this paper, we conduct a regression discontinuity design leveraging a policy threshold to reveal the causal impact of R2T4 on student outcomes. In short, students who withdraw strictly after the 60 percent point of the academic term are determined to have earned all of their federal financial aid, whereas those who withdraw just before earn only a prorated share. In practice, we find that students are largely unaware of the policy threshold and do not sort around this threshold, yet those who exit just before the 60 percent mark are required to return, on average, roughly $1800 to the federal government. Using data for the entire population of students from Fall 2019, we show that there are non-statistically significant changes in enrollment rates for students required to return federal aid. However, we do find that students who owe aid decrease their aid uptake in the next period of enrollment by 11 percent relative to those who do not require repayments and provide evidence that this may be driven by transfers. Moreover, we find that Pell recipients required to return funds are less likely to re-enroll in school than students who barely earn all of their aid. These findings show that the R2T4 policy may be leading some students to switch institutions and lower their borrowing, both of which may negatively impact degree completion or increase time to degree. In ongoing work, we plan to measure the impact on these degree attainment outcomes and further investigate mechanisms.