An Analysis of the Determinants & Costs of Federal Student Loan Default

Abstract

The increasing prevalence of federal student loan default is problematic for both students and the government. In this paper, I aim to better understand the determinants and costs of student loan default in two distinct ways. First, I conduct a logistic regression analysis using data collected from the Baccalaureate and Beyond Longitudinal Study (2008-­‐2012) to determine the effects of various individual and institutional characteristics on the probability of default. In doing this, I find that a borrower’s race, post-­‐graduation earnings level, cumulative debt level, and undergraduate major have the most significant impact on the likelihood of default, and that these effects are not mitigated by the rising popularity of Income-­‐ Based Repayment plans. Second, I estimate the government’s expected return on federal student loans given certain institutional characteristics in order to quantify the cost of loan default. Using data collected from the U.S. College Scorecard (2009-­‐ 2013), I find that the government’s expected return is nearly doubled when loans are withheld from higher risk institutions, such as 2-­‐year colleges and colleges with unacceptably high cohort default rates.