A new historical database distinguishes bank runs that exposed weak fundamentals from runs that strong banks survived without failure.
Research question
What causes bank runs, and when do they lead to bank failure and local economic contraction?
Method & data
Large language models are applied to historical newspapers to build a database of U.S. bank runs from 1863 to 1934, linked to bank fundamentals and local outcomes.
Main finding
The abstract reports that runs occurred at both weak and strong banks, but failure and larger local declines were concentrated among banks with poor fundamentals.
Why it matters
The evidence tempers accounts in which runs alone create severe discontinuous outcomes regardless of bank condition.
What to verify
The public abstract does not state the validation error of the newspaper classification or how the historical setting maps to modern deposit insurance.
Show public abstract used for this summary
We study the causes and consequences of bank runs. By applying large language models to historical newspapers, we create a comprehensive database of bank runs in U.S. history with information on 3,984 runs on individual banks from 1863 to 1934. Our novel data allow us to establish that runs are considerably more likely in weak banks but also occur in strong banks, especially in response to negative news about the real economy or the broader banking system. However, runs typically only result in failure for banks with poor fundamentals. Strong banks survive runs through various mechanisms, including signaling strength, interbank cooperation, and temporary suspension. At the local level, runs on banks with poor fundamentals translate into substantially larger declines in deposits, lending, and manufacturing activity than runs on strong banks. Our findings imply that poor fundamentals are central to explaining both when runs occur and when they have severe economic effects, tempering the view that small shocks can generate discontinuous jumps to bad equilibria through self-fulfilling run dynamics.