When a Run Becomes an Economic Shock

This chapter measures what remains after the immediate withdrawal wave: the damage to a surviving bank, the contraction of local deposits and lending, and the decline in manufacturing activity. The important comparison is not simply run versus no run. It is run without failure versus run with failure, and run at a strong bank versus run at a weak one. That comparison shows why survival is meaningful without being costless, and why permanent institutional exit is often the point at which a banking disturbance becomes a broader economic shock.

Survival can leave a long funding scar

A bank does not need to fail for a run to change its balance sheet. In the historical national-bank sample, banks that experienced a run without failure saw deposits and loans fall by roughly 7 percent of their pre-run assets. The reduction persisted for more than four years. The bank remained open, but it was operating with a smaller funding base and a smaller loan book than before the episode.

The severity differed by the form of the episode. Runs in which the bank never suspended were associated with a deposit decline of about 5 percent, and deposits recovered within five years. Runs followed by a temporary suspension were associated with a decline of about 15 percent that remained visible for the same broad horizon. Suspension can preserve the institution while imposing a lasting confidence and funding cost.

The difference between strong and weak banks is even more revealing. Surviving runs at banks with strong fundamentals were associated with essentially no lasting decline in deposits or loans. Surviving runs at banks with weak fundamentals were followed by persistent declines of about 7 percent in both. The run therefore does not carry one universal economic effect. Its consequences depend on the bank’s ability to replace lost funding and continue lending after the immediate pressure has passed.

The study also examines the smaller group of non-fundamental runs—episodes attributed in newspaper accounts to false rumors or confusion and not to bank-specific or wider economic distress. Even these more plausibly exogenous liquidity shocks were associated with roughly 5 percent declines in deposits and loans at the affected bank. Deposits recovered within about two years, while loans remained roughly 5 percent lower three years after the run.

That pattern is important for two reasons. First, it rejects the claim that a run is harmless whenever the bank survives. Funding disruption can persist after the doors remain open. Second, it preserves the distinction between moderate bank-level damage and severe system-level damage. A smaller loan book at one bank is not the same as a city losing a large share of its credit supply or manufacturing activity. The next comparison concerns that difference in scale.

Reading persistence without confusing the denominator

The reported declines are measured relative to the bank’s assets before the run. That denominator matters. A 7 percent decline in deposits or loans is not a claim that 7 percent of all deposits in the historical banking system disappeared, nor is it a direct estimate of losses to every depositor.