| Popis: |
I study how the trends and cycles of unsecured consumer credit impact the aggregate economy. This is an essential topic because the ratios of both credit card debt and student loan debt to GDP have become quite large. In chapter 1, I study how the interaction between unemployment insurance (UI) and unsecured consumer credit impacts consumption and welfare over the business cycle. I measure the effects of this relationship using a quantitative equilibrium model of labor markets and credit markets calibrated to depict the employment risk, credit and bankruptcy behavior in the US economy. I find that the extension in the duration of UI benefits during the Great Recession prevented over a 29 percentage point further drop in unsecured consumer credit and a 2 percentage point further drop in aggregate consumption. I show that improvements in the terms of credit accounted for over 60% of the gains in aggregate consumption from extensions in the duration of benefits. In chapter 2, I study the implications of allowing student loans to be discharged in bankruptcy. I develop a model of education, student loans, credit card debt and bankruptcy. I explicitly model the institution that student loans cannot be discharged in bankruptcy in the United States. I find that allowing student debt to be discharged in bankruptcy results in more college education, but a significant rise in consumer bankruptcies. However, the policy more than pays for itself in the sense that the increase in tax revenue offsets the losses from more bankruptcy. Moreover, the policy leads to significant welfare gains where low-productivity and moderately wealthy households experience the largest gains. In chapter 3, I develop a new solution method that is faster and more accurate than existing methods in solving problems with non-convexities. This method relies on creating numerous endogenous grids, one for each convex set of the initial assets consistent with optimal decisions. Each each endogenous grid has a proxy solution, and a global solution is chosen as the optimal proxy solution. Because models of consumer credit and default generate non-convexities, this method is particularly useful in solving the models of chapters 1 and 2. |
