We develop a model of liquidity shortages that incorporates a general equilibrium feature of liquidity: when banks hold more liquidity, other agents in the economy hold less of it and will supply less in times of crisis. We show that the private holdings of liquidity at banks are inefficient, with the direction of the bias being determined by the characteristics of the suppliers of liquidity to banks. Minimum liquidity requirements for banks may reduce welfare; in such cases interest rate policies that stimulate the ex-post supply of liquidity can restore efficiency. Overall, our results show that optimal liquidity policies critically depend on a financial institution’s (marginal) source of liquidity and will hence differ across institutions of different types.

We examine how public liquidity should be distributed to firms when immediate production entails externalities, such as by spreading a virus. Direct provision of liquidity can address externalities, but traditional distribution of liquidity (through banks) has informational advantages. We show that which mode is preferred is determined by the variance (but not the level) of firm characteristics in the economy. Traditional provision is always part of the optimal policy when liquidity modes can be combined, and involves promising low interest rates for when the pandemic is over in order to incentivize temporary production shutdowns at firms.