Abstract
This dissertation presents empirical evidence on the effects of postsecondary student debt relief. In the United States, federal student loans are the primary mechanism for relaxing credit constraints in a higher education market characterized by vast heterogeneity in institutional quality and student outcomes. The vast majority of this debt is supplied by the federal government with no traditional underwriting. Regardless of default risk, borrowers face similar loan limits and interest rates. As student loans have grown to be the second largest source of household debt, evidence has pointed to particular difficulties in debt management. Well-documented frictions in the loan repayment process have, in part, contributed to student loan delinquency rates that are among the highest of all consumer debt categories.
The three papers presented here examine whether debt alleviation measures shape individual economic behavior. This question is motivated in part by growing concern among policymakers that student debt burdens may suppress economic activity through debt overhang and constrained household spending. Various forms of relief have been proposed and implemented, including payment moratoria and broad-based debt cancellation. A central aim of this research is to provide empirical evidence informing the discussion of tradeoffs across these policy instruments. What types of borrowers stand to benefit from relief? Under what conditions may beneficiaries alter their behavior? As student loans enter a period with some of the highest-ever transition rates into delinquency, an overarching goal of this research agenda is to understand the economic choices of borrowers and identify areas of improvement for the higher education finance system.
The first chapter, Credit Scars: Credit Reporting Relief for Defaulted Student Loan Borrowers, estimates the causal effect of removing student loan default records from credit reports. Using a stacked difference-in-differences design that exploits the automatic removal of default information under the seven-year Fair Credit Reporting Act threshold, I find that record removal raises credit scores by fifteen points and induces significant new borrowing in mortgage, auto, and credit card markets. These effects are driven primarily by entry on the extensive margin, indicating that derogatory marks function as binding barriers to credit long after the underlying financial distress has resolved. These findings suggest substantial social returns to policies that facilitate credit rehabilitation.
The second chapter, Labor, Loans and Leisure: The Impact of the Student Loan Payment Pause, co-authored with Sarah Turner, examines how the suspension of required federal student loan payments—from March 2020 through September 2023—affected labor supply. Under the permanent income hypothesis, a temporary liquidity transfer should leave work effort unchanged; however, for borrowers facing binding credit constraints, additional cash-on-hand may facilitate reductions in hours worked. Using a matched difference-in-differences design with panel data from the Survey of Income and Program Participation, we find that the pause reduced average weekly hours worked by 3.3 hours—an eleven percent decline—among borrowers who had not completed a bachelor’s degree. Among college- and graduate-degree holders with federal student loans, there is no detectable change in employment or hours. Consumer finance data corroborate this heterogeneity: households with federal student debt and no bachelor’s degree report being liquidity constrained at roughly twice the rate of degree-holding borrowers, suggesting the differential labor supply response reflects differences in financial flexibility rather than preferences alone.
The third chapter, Voting on a Promise: Did the Proposed Student Loan Forgiveness Impact Electoral Outcomes?, co-authored with Arnav Dharmagadda and Jaden Shawyer, studies whether the Biden administration’s August 2022 broad-based debt cancellation program influenced voting in the November midterm elections. Announced less than three months before the election and attracting over twenty-five million applications before legal challenges halted implementation, the program provides an unusual opportunity to study voter responsiveness to a salient but unrealized transfer. Using geographic variation in the share of the voting-age population that submitted applications—constructed from administrative Department of Education data—alongside county-level electoral returns, we find that a one-standard-deviation increase in the county application rate corresponds to approximately a one-percentage point increase in the Democratic House vote share, an association that does not predict Democratic voting in prior midterms. Individual-level survey data corroborate the county-level estimates. These results indicate that the electoral returns to debt relief accrued before any benefits were realized, underscoring the role of program salience and credibility in translating redistributive policy into political support.
