Zhang, Shuoxun (2011-06). Essays on Banks' and Consumers' Behavior in the Presence of Government as the Credit Insurer of Last Resort. Doctoral Dissertation.
Thesis
My dissertation investigates the behavior of consumers and banks in the presence of government as the credit insurer of last resort. Consumers have an option to file for bankruptcy under law when there are unexpected adverse shocks, while banks, especially large banks, are supported by the government during financial crisis because of systemic risk. I explore the heterogeneous behavior among consumers and banks with adverse shocks. In the first chapter of my dissertation, my inquiry focuses on the heterogeneous behavior of households in filing for bankruptcy. In the literature, there are two theories in explaining personal bankruptcy: adverse event theory and strategic timing theory. Fay, Hurst and White(FHW) 2002(AER) include both financial benefit and adverse event variables in explaining the bankruptcy decision, and they find only financial benefit from filing is significant in explaining whether to file or not. Our argument is that adverse events may not work directly on bankruptcy decisions, however, they operate by running a higher amount of debt. Thus FHW's setting may not be appropriate. Instead, adverse event consumers' debt occurs after adverse events, while strategic timing consumers' debt decision and bankruptcy decision are jointly determined, which means their debt or financial benefit is endogenous; thus we propose that the endogeneity test of financial benefit is a way to distinguish the two types of consumers. Assuming only one type exists in the sample, we find support for adverse event theory. Extending the analysis to allow for both adverse events and strategic timing consumers shows existence of both types of filers, and strategic timing filers are more sensitive to financial benefit. Additionally, lower access to debt markets and lower income significantly increase the chance of strategic behavior. The second part of my dissertation is to study the effectiveness of the Troubled Asset Relief Program(TARP) on banks' loan to asset ratio. One of the fundamental objectives of the Troubled Asset Relief Program (TARP) is to stimulate bank loan growth. I use panel data to study the dynamic effect of TARP investments on banks' loan to asset ratio (LTA). I find that TARP stimulate recipients' LTA growth as a whole, and the effect is significant only for medium banks(asset between 1 billion and 10 billion), with an annual decrease of 14 percentage points in LTA with the LTA in treatment quarter as benchmark. In terms of a dollar amount, 7.71 dollar more loans are generated for every TARP dollar invested in medium banks, compared with the average level of the quarters before TARP. There is no significant effect on small banks or big banks. Using graphs and different regression models, I argue that the dynamic setting, rather than the cross-sectional comparison, is more appropriate.
