How Might Central Bank Digital Currency Affect Banks?  

Central banks around the world either have issued or are considering the issuance of a central bank digital currency (CBDC). While most modern transactions take place digitally, a CBDC differs from existing digital money such as bank deposits because a CBDC is a direct liability of the central bank, rather than of a commercial bank. Several policy issues and risks surround the creation of a CBDC. In our paper, we consider one of the most important: Does a CBDC draw deposits away from banks and does this disintermediation restrict lending? As noted in a report from the Federal Reserve, disintermediation is an issue because “a widely available CBDC […] could reduce the aggregate amount of deposits in the banking system, which could in turn increase bank funding expenses, and reduce credit availability or raise credit costs for households and businesses.” In a 2021 article in Business Reporter, Greg Baer, the President and CEO of the Bank Policy Institute, writes that “given that the average loan-to-deposit ratio for banks is generally around 1:1, every dollar that migrates from commercial bank deposits to CBDC is one less dollar of lending.” However, deposits are not the only source of funding for banks, as many banks, especially large ones, can fund deposit shortfalls using wholesale funding. More fundamentally, the assumption that bank lending must be funded by deposits rests upon the basic question of whether loan origination is necessarily dependent on deposit creation. 

We start our investigation of the effects of a CBDC on bank lending with the following observation: Deposit creation and loan origination need not be linked to each other, although this decoupling occurs only under two stringent conditions. First, loan rates are set in a market that ensures the participation of both borrowers and lenders. In other words, banks find these loans profitable. Second, loans remain profitable even if banks lose cheap deposits if banks can replace those deposits with wholesale funding at the same cost. These two conditions imply that CBDC will not affect bank lending because the two sides of banks’ balance sheets are independent. While banks do not operate under these conditions, this baseline allows us to isolate the various features of the banking system that allow a CBDC to affect lending. 

The bulk of our analysis involves estimating the parameters of a model of banks, depositors, borrowers, and a central bank, and then examining the behavior of loans in the model when we counterfactually add a CBDC. As the model is disciplined tightly by data, these counterfactual experiments offer quantitative rather than just qualitative conclusions. In our model, banks compete with one another by setting loan and deposit rates, making loans, and taking deposits. Bank default is the main realistic feature of the model that links lending to deposits. Since bank wholesale funding is not insured, default risk and funding costs rise when banks use more wholesale funding. Thus, any shrinkage in cheap deposits cannot be covered fully by this alternative funding source. Nor can deposit shrinkage be covered by equity issuance. Banks only tap this market infrequently because of information frictions and issuance costs. 

We model the deposit and lending markets by using what is called a characteristic-based demand approach, in which the demand for products such as deposits or loans depends on features such as convenience, maturity, and the relevant interest rate. In other words, we get data estimates of the value of specific features of financial products. We then introduce a CBDC as a new product. This approach is conducive to studying different policy proposals concerning a CBDC because we have already estimated the value of the characteristics that can be attributed to a CBDC. 

In the model with all parameters estimated from our data, we introduce a CBDC as a direct central bank liability that households can hold, with the interest rate determined by the central bank. Facing competition from the CBDC, banks can increase deposit rates, replace deposits with wholesale funding, or cut lending. The exact quantitative margins that banks use to accommodate the introduction of CBDC depend on the links between deposit taking and loan origination. 

First, we consider a CBDC that pays no interest and provides the same transaction services as bank checking accounts. We find that the introduction of a CBDC does impact bank deposits with an overall 10% drop in the level of bank deposits. However, only a third of the impact on deposits is passed through to lending. The attenuated impact on bank lending happens because banks can replace deposits with wholesale funding. 

However, this substitution is not completely harmless. As banks use more uninsured funding, bank default risk rises. We estimate that a one percentage point increase in the market share of a CBDC would increase a bank’s default probability by 1.1%, thus raising banks’ funding costs. In addition, as banks replace interest-insensitive deposits with wholesale funding, banks’ exposure to interest rate risk also rises. For example, an unexpected rise in short-term rates boosts funding costs more than it would if banks were more reliant on rate-insensitive deposits. As such, bank capital suffers. 

We also consider an alternative policy proposal that allows a CBDC to pay interest. On the one hand, this proposal might improve general consumer welfare, especially for those that lack access to interest-bearing bank deposits. Moreover, competition from a CBDC can force banks to pay higher deposit rates, which benefits depositors. On the other hand, an interest-bearing CBDC might be too disruptive because it would divert a significant amount of deposits from the banking sector. Indeed, we find that banks would lose around 30% of their deposits if a CBDC paid the federal funds rate, despite the competitive response from  

banks to raise deposit rates. An interest-bearing CBDC also generates substitution patterns different from a non-interest-bearing one. Banks lose more savings deposits than transaction deposits if the CBDC pays interest, while the opposite is true when the CBDC does not pay interest. 

The impact of a CBDC affects small banks more sharply. Big banks rely less on deposits and have better access to non-deposit financing. As result, they are better equipped to adapt to a financial system with a CBDC than small banks. We find the impact of CBDC on bank lending is three times greater for small banks than for big banks even though CBDC has comparable effects on deposits across both groups. Because small firms disproportionately rely on small banks for credit, this result suggests that CBDC could have distributional implications across the firm size distribution. 

We consider possible policy responses to alleviate the adverse impacts of a CBDC on the banking system. Because a CBDC makes banks use more interest-sensitive liabilities, it makes banks more sensitive to interest rate hikes, as the rates they pay on short-term, wholesale funding respond immediately, while the rates they earn on long-term assets (loans) do not. One possible, but highly speculative central bank policy that would undo this effect would be to use funds raised from a CBDC to make term loans to banks. This policy would allow banks to access more long-term maturity liabilities, thus reducing their interest rate risk. However, this policy would also cause the central bank to bear default risk and might not be feasible as such. 

Another issue is bank profitability, which might fall as banks face competition from a CBDC. One potential way to alleviate the impact of CBDC on bank profitability is to use an “intermediated account-based” CBDC. Under this approach, private banks would offer accounts or digital wallets to facilitate the management of CBDC holdings and earn a fee.  

This approach also has the advantage of addressing money laundering and theft concerns for an anonymous “token-based” CBDC because banks have the infrastructure to verify the account holders’ identities. Banks earn a fee from providing these valuable services, which reduces the impact of CBDC on bank profitability. 

Toni M. Whited is the Dale Dykema Professor of Business Administration at the Ross School of Business at the University of Michigan and a Research Associate at the National Bureau of Economic Research.  

Yufeng Wu is an Associate Professor of Finance at the Gies School of Business at the University of Illinois. 

Kairong Xiao is an Associate Professor of Finance at Columbia Business School.  

This post is adapted from their paper, “Central Bank Digital Currency Banks,” available on SSRN.  

The views expressed in this post are those of the authors and do not represent the views of the Global Financial Markets Center or Duke Law.