Abstract:

Using de-identified bank data, we study the impact of taking out loans on household financial and labor outcomes. We use a difference-in-differences design to test neoclassical versus behavioral models of borrowing in response to shocks. We find early evidence of a short-term increase in consumption, driven by spending on luxuries, consistent with models with present-biased agents. There is also suggestive evidence of changes to employment driving the decision to borrow.