The average federal student loan balance at graduation has more than doubled from $18,230 in 2007 to $37,850 in 2024. Housing costs comprise over a third of the cost of four-year college attendance and in the last decade, student housing costs have begun rising more rapidly than tuition and fees. This paper examines the causal role of student housing costs in the growth of student loan debt. I exploit housing cost shocks generated by significant expansions to on-campus housing capacity. On-campus housing expansions make students more likely to live on-campus and represent a shock to housing costs equal to the difference between the on-campus and off-campus cost of living. The timing of housing expansions due to construction is intuitively subject to numerous random factors that are exogenous to student debt outcomes. I identify large, permanent jumps in institutions’ on-campus room capacity reported in IPEDS. Using institution-cohort level student debt outcomes from the College Scorecard, I perform an event study around capacity expansions, comparing median student debt amounts at expanding institutions relative to never-expanding and not-yet-expanded institutions. This analysis also evaluates the potential role for affordable, on-campus housing to alleviate student debt burdens.
Classical microeconomic theory posits that a worker should value an hour of time according to her hourly wage. However, this standard theory neglects commonly understood facts about the modern labor market, like the increased availability of paid time off, work from home arrangements, and salaried work enjoyed by high income workers. This project investigates the empirical relationship between a time shock and subsequent earnings and how this relationship varies across the income spectrum. I use weather-related school cancellations to evaluate how different workers respond to an unanticipated time shock. Because school closures are not recorded centrally by any state or the federal government, I use three data sources to predict school cancellation: winter weather, reports of reduced work hours by teachers in the CPS, and school closure announcements on social media and local news, collected by Chat GPT. My preliminary results use CPS data to show that imputed snow days are associated with a reduction of reported work hours among parents of school-aged children, and that this effect varies by socioeconomic status (though the specific relationships vary based on the imputation strategy used).