CBDC and banks: disintermediating fast and slow

Summary

Focus

Advances in payment technologies have led central banks to consider issuing to the public central bank money in digital form, commonly referred to as digital cash or retail central bank digital currency (CBDC). The potential impact of CBDC on the banking system has been hotly debated, with two phenomena receiving particular attention: "slow disintermediation", by which CBDC competes with bank deposits in normal times, leading to more expensive funding and a shrinking of the sector, and "fast disintermediation", by which CBDC provides an especially convenient asset to hold in times of banking stress, enhancing the scope for bank runs.

Contribution

We examine the impact of a central bank digital currency, combining survey evidence from German households with a macroeconomic model featuring endogenous systemic bank runs. We document novel evidence regarding households' projected use of a hypothetical digital euro – in normal times and in periods of banking stress. We build a model of the economy featuring CBDC and endogenous bank runs, which is matched to empirical data and our survey findings. Using the quantitative model, we explore the implications of CBDC for welfare, the banking sector and policy design choices.

Findings

Our survey indicates that German households are open to CBDC in normal times and that they may replace some of their bank deposit holdings with a digital euro. However, the demand for CBDC also raises financial stability concerns as households seem more likely to withdraw funds from banks during times of stress if a CBDC is available. Our quantitative model is designed to capture these phenomena and allow for policy prescription. We find that a CBDC introduced with a suitable holding limit increases financial stability and welfare. The limit allows some of the benefits of a digital euro as money, while choking off its effect on run risk.

Abstract

We examine the impact of a retail central bank digital currency, combining survey evidence from German households with a macroeconomic model featuring endogenous systemic bank runs. The survey reveals non-trivial demand for retail CBDC as a substitute for bank deposits in normal times ("slow disintermediation") and increased withdrawal risks during financial distress ("fast disintermediation"). Informed by the survey, the model indicates that introducing a retail CBDC might reduce financial stability because CBDC offers storage-at-scale - making it attractive to run to. We estimate an optimal holding limit which chokes off fast disintermediation and enhances financial stability by shrinking a fragile banking system.

JEL classification: E42, E44, E51, E52, G21

Keywords: Central bank digital currencies, financial crises, disintermediation, bank runs, banking system, money