(Job Market Paper)
Using individual level credit information, I estimate the impact of access to ride-sharing on student debt repayment and take-up. I find that following the introduction of ride-sharing services in a city, individuals decrease their student debt balance and probability of default. These results are primarily driven by former students, who are 0.4pp more likely to finish repaying their student loans and are 0.8pp less likely to default on their student debt in the three years after ride-sharing arrives. This effect is absent for current students. For potential students, I find that access to ride-sharing increases the likelihood of getting a first student loan by 0.5pp. This suggests that there is a willingness to attend higher education that is not met given the structure of the current labor market, and that having access to the gig economy is allowing individuals who otherwise would have chosen not to enroll to do so. Taken together, results suggest that access to flexible work improves student loan repayment rates, while simultaneously fostering enrollment.
(with Taylor Begley, Radhakrishnan Gopalan and Naser Hamdi)
We use individual-level data to quantify the effect of getting a mortgage on non-mortgage credit outcomes. We use a regression discontinuity design and find that individuals that transition to homeownership increase their credit card and auto balances by $8,300 and $14,800, suggesting a debt spillover effect from home ownership. This increase in debt is equivalent to 13% of the average mortgage loan, and we provide evidence that it is mainly driven by a change in credit demand. We find that this increase in debt is driven by individuals with higher financial experience, while their overall ability to service their debt remains unchanged. In contrast, low-experience individuals do not increase their debt, but are relatively more likely to experience a deterioration in their financial health. Taken together, these results highlight the role financial experience plays in managing the debt burden associated to a new home.
(with Naser Hamdi, Ankit Kalda and David Sovich)
We use administrative payroll data to examine how firms adjust employment in response to aggregate shocks. We use the COVID-19 pandemic as a laboratory. Exploiting within firm-state variation, we find that firms are more likely to lay off low-income and high-tenure employees before other classes of workers. This pattern disappears a few weeks into the pandemic, after which layoffs are more uniformly distributed. This pecking order of layoffs is most pronounced in low-skilled industries (e.g., retail trade) and in firms with high turnover costs. Our results are consistent with theories which document that reputational costs and search frictions play an important role in determining firms' responses to aggregate shocks. To further evaluate these frictions in our setting we examine whether expansion of unemployment benefits under the CARES Act increased the likelihood of layoffs. Consistent with both frictions, we find that even within the same income bucket, firms are more likely to lay off employees in states with more generous benefits.