Liquidity Crisis - A Model of Liquidity Crisis

A Model of Liquidity Crisis

One of the earliest and most influential models of liquidity crisis and bank runs was given by Diamond and Dybvig in 1983. The Diamond–Dybvig model demonstrates how financial intermediation by banks, performed by accepting assets that are inherently illiquid and offering liabilities which are much more liquid (offer a smoother pattern of returns), can make banks vulnerable to a bank run. Emphasizing the role played by demand deposit contracts in providing liquidity and better risk sharing among people, they argue that such a demand deposit contract has a potential undesirable equilibrium where all depositors panic and withdraw their deposits immediately. This gives rise to self-fulfilling panics among depositors, as we observe withdrawals by even those depositors who would have actually preferred to leave their deposits in, if they were not concerned about the bank failing. This can lead to failure of even ‘healthy’ banks and eventually an economy-wide contraction of liquidity, resulting in a full blown financial crisis.

Diamond and Dybvig demonstrate that when banks provide pure demand deposit contracts, we can actually have multiple equilibria. If confidence is maintained, such contracts can actually improve on the competitive market outcome and provide better risk sharing. In such an equilibrium, a depositor will only withdraw when it is appropriate for him to do so under optimal risk–sharing. However if agents panic, their incentives are distorted and in such an equilibrium, all depositors withdraw their deposits. Since liquidated assets are sold at a loss, therefore in this scenario, a bank will liquidate all its assets, even if not all depositors withdraw.

Note that the underlying reason for withdrawals by depositors in the Diamond–Dybvig model is a shift in expectations. Alternatively, a bank run may occur because bank’s assets, which are liquid but risky, no longer cover the nominally fixed liability (demand deposits), and depositors therefore withdraw quickly to minimize their potential losses.

The model also provides a suitable framework for analysis of devices that can be used to contain and even prevent a liquidity crisis (elaborated below).