We employ a unique hand-collected dataset and a novel methodology to examine systemic risk before and after the largest U.S. banking crisis of the 20th century. Our systemic risk measure captures both the credit risk of an individual bank as well as a bank’s position in the network. We construct linkages between all U.S. commercial banks in 1929 and 1934 so that we can measure how predisposed the entire network was to risk, where risk was concentrated, and how the failure of more than 9,000 banks during the Great Depression altered risk in the network. We find that the pyramid structure of the commercial banking system (i.e., the network’s topology) created more inherent fragility, but systemic risk was nevertheless fairly dispersed throughout banks in 1929, with the top 20 banks contributing roughly 18% of total systemic risk. The massive banking crisis that occurred between 1930–33 raised systemic risk per bank by 33% and increased the riskiness of the very largest banks in the system. We use Bayesian methods to demonstrate that when network measures, such as eigenvector centrality and a bank’s systemic risk contribution, are combined with balance sheet data capturing ex ante bank default risk, they strongly predict bank survivorship in 1934.
Sanjiv R. Das, Kris James Mitchener, and Angela Vossmeyer