Money and Modern Banking without Bank Runs

Abstract

Deposit insurance has been adopted at an accelerating rate by countries around the world, but evidence indicates that it causes increased risk from moral hazard and may be adopted for rent-seeking reasons. A primary justification for deposit insurance is the theoretical literature on bank runs. Following Diamond and Dybvig (1983), bank runs take the form of withdrawals of real demand deposits that deplete a fixed reserve of goods in the banking system. This describes historical runs in which there were large withdrawals of currency from the banking system. However, in a modern banking system, large withdrawals take the form of electronic payments within a clearinghouse system of banks by wholesale depositors for either redepositing funds or making purchases. These transfers shift balances among banks, with no analog of a depletion of a scarce reserve from the banking system. To study the threat of bank runs due to multiple equilibria in a modern economy, I examine a model of nominal demand deposits repayable in money within a clearinghouse. Bank runs do not occur for two reasons. First, any money withdrawn from one bank is used to redeposit at another bank or to buy goods. The money is transferred to another bank and there is no depletion of money from the banking system. The latter bank always prefers to lend back to the former bank since it is fundamentally solvent. The first bank never fails, even if excess early withdrawals occur, implying the fear of bank defaults never occurs in equilibrium. Second, depositors do not run the bank for fear that consumption goods will be depleted from the economy. A bank run to purchase goods would drive prices up. Nominal deposits imply the bank's real liabilities do not increase. The bank is fundamentally solvent and can borrow back the excess funds it paid out. Flexible prices in the goods market ration would consumption to those who run the bank and would dissuade a run, so a run never occurs in equilibrium. This finding suggests that deposit insurance, which prevents bank runs in models of real deposits as in Diamond-Dybvig, may not be needed to prevent bank runs in a modern economy.