Where Is the Money in the Bank? Mostly Not There

A bank takes deposits that can be withdrawn at short notice and makes loans that cannot be called in, which is the central function and the central vulnerability. That mismatch between the maturity of what it owes and what it owns is what a bank exists to manage, since depositors want access and borrowers want time, and the bank stands between them. It works because depositors do not all want their money at once, so a fraction held as reserves covers ordinary withdrawals while the rest is lent. It stops working if enough depositors want their money simultaneously, since the loans cannot be converted to cash quickly, which is a run and which can destroy a solvent bank.

The term covers several things and the distinctions matter:

• •Physical cash held in branches and vaults, which is a small part of the total
• •Balances held at the central bank, which are the main form and are electronic
• •Required reserves, a proportion of deposits that regulation obliges a bank to hold
• •Excess reserves held beyond that requirement, which grew enormously after 2008
• •Liquid assets that can be sold quickly, which regulation now treats separately
• •Capital, which is something different entirely and is frequently confused with reserves

The textbook account of how lending creates money has been corrected by central banks themselves. The traditional description has banks receiving deposits and lending out a multiple of them, constrained by the reserve requirement, with money created by successive rounds of lending and redepositing. Central banks including the Bank of England have published explicit statements that this is not how it works. A bank making a loan creates a deposit in the borrower's account by writing it into existence, which is the act that creates new money, and it then obtains whatever reserves it needs afterwards. The binding constraints are the bank's capital, its assessment of whether the borrower will repay, and the demand for loans, rather than a quantity of reserves sitting on hand.

Confusing capital with reserves is the commonest error in discussions of banking and the distinction is clear once stated. Reserves are an asset, being money the bank holds. Capital is not an asset at all but the difference between what the bank owns and what it owes, which is the share of its assets funded by its owners rather than by depositors and lenders. Capital absorbs losses, since a loan that is not repaid reduces the owners' stake before it reaches anybody else, and a bank with more capital can lose more before it becomes insolvent. Reserves address liquidity, meaning whether the bank can meet withdrawals today. A bank can be solvent and illiquid, which is a run, or liquid and insolvent, which is worse and lasts longer.

The vulnerability is structural, so the remedies are institutional rather than a matter of banks holding more cash. Deposit insurance guarantees balances up to a stated limit, which removes the incentive to run for the great majority of depositors and is the single most effective measure, introduced in the United States in 1933 after widespread failures. A central bank acting as lender of last resort supplies funds against good collateral to a bank facing a run, on the principle stated by Walter Bagehot in 1873 of lending freely at a penalty rate against sound security. Capital requirements ensure a bank can absorb losses. Liquidity requirements ensure it holds enough assets convertible quickly. The events of 2023 demonstrated that runs can now occur within hours rather than days, which has prompted reconsideration of several of these.

A bank owes money repayable on demand and owns loans that cannot be called in, which is what it exists to manage and what makes it fragile. Reserves are mostly electronic balances at the central bank rather than cash. Lending creates deposits rather than lending out existing ones, which central banks have stated explicitly, and deposit insurance rather than reserves is what stops runs.