Bank Run

A bank run occurs when a large number of depositors attempt to withdraw their funds at the same time, typically because they fear the institution may become insolvent. Because banks operate on a fractional reserve basis and keep only a small portion of deposits as cash, a sufficiently large, simultaneous demand for withdrawals can exhaust available liquidity and force the bank to sell assets at a loss or fail outright. The run is the moment a bank's defining maturity mismatch — long-dated assets funded by on-demand liabilities — stops being an abstraction.

The self-fulfilling dynamic

Bank runs are often self-reinforcing. A rumor, a credit downgrade, or visible queues can convince otherwise-secure depositors that they must withdraw first to avoid being last in line. Even a fundamentally solvent bank can be toppled if enough customers act on this incentive at once, because its assets (loans, bonds) cannot be liquidated quickly enough to meet immediate cash demands, and forced sales into a falling market convert a liquidity problem into a genuine solvency problem. Economists model this as a coordination game with two stable outcomes: if everyone believes the bank is fine, it is; if everyone believes it will fail, it does. When the panic spreads across multiple institutions it becomes a systemic banking crisis.

Modern speed

Historically a run meant physical lines at a teller window, which imposed a natural speed limit measured in days. Digital banking has removed it: deposits can now leave an institution in minutes through online transfers, and social media synchronizes depositor fear far faster than rumor ever did. The 2023 failures of several US regional banks demonstrated the new physics — tens of billions of dollars in withdrawal demands materialized within hours, faster than any historical precedent. Deposit insurance and central-bank liquidity facilities (the lender of last resort) exist to dampen runs by removing the incentive to be first in line, but insurance coverage limits mean uninsured balances remain exposed, and businesses holding operating cash above those limits face exactly the classic run incentive.

Runs beyond banks — and the Bitcoin angle

The run dynamic applies to any custodian whose liabilities exceed its liquid assets, and the Bitcoin world has produced its own textbook cases: exchanges and lending platforms that rehypothecated customer coins collapsed in 2022 under withdrawal waves that were functionally identical to bank runs, just without the insurance backstop. "Not your keys, not your coins" is the run-proofing principle stated plainly. Coins held in self-custody cannot be run on, because there is no pooled reserve and no queue — a UTXO in your cold storage is not a claim on anyone's balance sheet. That is also why periodic mass withdrawals from exchanges, sometimes organized as "proof-of-keys" events, function as voluntary stress tests: a custodian that cannot honor them was fractionally reserved all along.

For the structural reason runs are possible in the first place, see fractional reserve banking. The lesson for a sovereign holder is not that banks are villains — it is that a claim and a bearer asset are different instruments, and you should always know which one you are holding.

Reading the warning signs

Runs look sudden but rarely are; the preconditions accumulate in public view. Classic tells include a deposit base concentrated in one industry or a few large accounts, heavy reliance on uninsured balances, long-duration assets purchased when rates were low, and management communications that suddenly emphasize "ample liquidity" — a phrase depositors historically hear as its opposite. For custodial Bitcoin services, the equivalent red flags are yield offers with no visible revenue source, withdrawal friction that appears during volatility, and proof-of-reserves disclosures that show assets but omit liabilities. None of these guarantees failure, but each one shortens the fuse. The practical posture is unglamorous: keep insured balances under coverage limits, keep operating float small, withdraw what you don't actively need — and keep long-term savings in instruments that have no queue to be first in.