Supervision, regulation, and financial instability: The political economy of banking during the Great Depression

This dissertation uses the historical experience of the Great Depression to explore the relationship between bank regulation, supervision, and financial stability and to shed light on current policy. It first shows that competing theories of regulation help account for differences in financial regulation across states in 1929. Consistent with the private interest view, laws prohibiting branch banking were most likely adopted in states where the unit-banking lobby was influential as they sought to extract economic rents by maintaining their local monopoly power. Capital requirements (a barrier to entry) were positively associated with the strength of a state's insurance industry as this pressure group lobbied to limit competition from banks, which could also provide insurance brokerage and securities underwriting services. Consistent with the public interest view, laws permitting statewide branching were more likely or reserve requirements were higher in states that had experienced a higher percentage of bank failures in the 1920s or where state deposit insurance schemes had failed. Next, the analysis tests whether differences in bank regulation and state banking departments influenced financial stability during the Great Depression. Direct measures of the quality of bank supervision are first constructed using new survey data from state banking departments, and instrumental variables estimation is employed to control for the potential endogeneity of regulation. Even after accounting for state differences in the structure of the banking industry and the severity of the Depression, states that prohibited branching, had higher reserve requirements, failed to insulate their supervisors from industry influence, or restricted supervisory powers to liquidate banks experienced higher bank failure rates during the Great Depression. The analysis then employs a previously underutilized data set on U.S. county bank failures to test the robustness of earlier results and control for both county and state effects. With respect to regulation and supervision, the results from the full sample of U.S. counties are largely consistent with the state-level results. When a matched pair analysis of border counties is used to remove unobserved heterogeneity across counties, only the supervisory characteristics account for county-level variation in failure rates. However, further tests reveal that the effects of regulation likely still matter.