“The Federal Reserve does not intend to proceed with issuance of a CBDC without clear support from the executive branch and from Congress, ideally in the form of a specific authorizing law.”
Board of Governors of the Federal Reserve System, Money and Payments: The U.S. Dollar in the Age of Digital Transformation, January 2022.
Note: A PDF version of this post is available here.
Last month, the Federal Reserve issued a long-awaited discussion paper on the possibility of introducing a central bank digital currency (CBDC) for retail use. The Fed paper calls for comments on the benefits and risk of introducing a U.S. CBDC, as well as on its optimal design. In this post, we respond to each of the 22 questions posed in the discussion paper. For the most part, these responses are based on our previous analyses of CBDC (here and here).
Before addressing the Fed’s specific questions, we would like to highlight our doubt that the benefits of a U.S. CBDC will exceed the risks. In our view, other, less risky, means are available to achieve all the key benefits that CBDC advocates anticipate. Moreover, we are not aware of sustainable design features that would reduce the risks of financial instability that many analysts agree will accompany the introduction of a digital U.S. dollar.
However, this overall judgment regarding a CBDC’s benefits and risks is sensitive to two considerations that appear in the Fed’s analysis either explicitly or implicitly. First, CBDC may be a less risky alternative to stablecoins, should regulation of the latter prove politically infeasible (see our earlier post). Second, if other highly trustworthy financial jurisdictions (with convertible currencies, credible property rights protections, and free cross-border flow of capital) offer their own CBDC, the case for a U.S. CBDC—as a device to sustain widespread use of the dollar—would become stronger.
Against this background, we applaud the Fed’s conservative approach. Most important, the U.S. authorities are not rushing to act. Instead, they are thinking carefully about the design elements, are actively engaged in public outreach, and have committed not to proceed without first securing broad public support (see the opening citation).
In addition, we share the Fed’s preliminary analysis that the following four features would best serve domestic needs, namely that a CBDC be: (1) privacy-protected; (2) intermediated; (3) transferable; and (4) identity-verified. These features are necessary to promote trust in the CBDC, to make it an efficient and accessible means of payment, and to minimize illicit use.
We would add two other strongly preferred features. First, to avoid a premium on CBDC (compared to other central bank liabilities), there should be no cap on individual holdings. Runs into CBDC pose a notable risk to the financial system. But, while limiting individual holdings could help mitigate this risk, it could lead to circumstances when the CBDC commands a premium, re-directing runs into close substitutes (such as Treasury bills). Second, the Federal Reserve should expect to pay interest on CBDC. Again, while paying interest will boost the risk of runs, we believe it politically unsustainable for the Federal Reserve to pay interest on wholesale digital liabilities (bank reserves, including those held by large foreign banks) without paying interest on retail digital liabilities (CBDC held by individuals and businesses).
Now, we turn to addressing the Fed’s questions. Since our responses rely on the same general principles, there is some repetition.
BENEFITS, RISKS AND POLICY CONSIDERATIONS
1. What additional potential benefits, policy considerations, or risks of a CBDC may exist that have not been raised in the paper?
Omitted benefit:
Negative nominal policy interest rate. In the absence of cash, an uncapped CBDC can bear a substantially negative interest rate. A deeply negative interest rate could be a useful macroeconomic policy stabilization tool in the event of a deep recession in which deflation becomes a threat.
Omitted risks:
Credit guidance and policy independence. An attractive CBDC (especially one that is uncapped and bears interest) is likely to prompt a significant and sustained shift out of commercial bank deposits. To counter this disintermediation and to ensure that commercial banks can still meet private credit needs, the Federal Reserve will need to recycle these funds through a system of collateralized loans (that is, some form of repurchase agreements). With its balance sheet vastly enlarged, the central bank’s collateral framework (including the haircuts on various commercial bank assets) will unavoidably influence the allocation of private credit. At the extreme, the central bank will be tempted to become a state bank supplying credit directly. Over time, this enlarged role in the allocation of private credit will be inconsistent with the independence needed for effective monetary policy.
Currency substitution from less trustworthy jurisdictions. As the digital currency of a highly trusted central bank in a stable political and financial jurisdiction, uncapped, interest-bearing U.S. CBDC likely will attract large capital inflows from abroad, especially during periods of global stress. The consequences could be catastrophic for the financial and monetary systems of emerging market and developing economies.
2. Could some or all of the potential benefits of a CBDC be better achieved in a different way?
We believe it is possible to reap the bulk of the benefits of a CBDC without the attendant risks. Below we identify alternative mechanisms for five widely claimed benefits:
More efficient payments. New retail payments frameworks with greater accessibility already are facilitating faster and less costly transactions, both within and across the borders of major jurisdictions. For example, the euro area has the TIPS system, with a processing time of 10 seconds at a cost of €0.002 per transaction. In the Nordic countries, the P27 project aims to provide a single, cross-border payments platform. When it launches in 2023, the FedNow Service should bring comparable benefits within the United States and eventually serve as a platform to which payments services abroad can link.
Improving payments access for the unbanked and underbanked. Experience suggests that improved access relies on two elements: subsidies for no-frills accounts and a low-cost mechanism for individual identity verification. As the Fed report highlights, the private Bank On initiative is promoting low-cost, low-risk consumer checking accounts. The U.S. government can support this effort directly through subsidies and indirectly by providing a zero-fee biometric identification tool. In the absence of subsidies and more efficient identity verification, we do not see how CBDC would add meaningfully to access.
Distributing government benefits. The federal and state governments already use the commercial banking system for distributing benefits and for receiving tax payments. To prepare for a potentially broader (and more rapid) use of these channels in the future, the government could invest now in the infrastructure needed for the universal and timely distribution of benefits.
Monetary policy at the effective lower bound. The Federal Reserve currently employs forward guidance and quantitative easing to improve financial conditions and support economic activity when its policy rate hits the effective lower bound. In the absence of paper currency, CBDC could permit a substantially negative nominal interest rate as an additional monetary policy stabilization tool in a deep recession that threatens deflation. However, deeply negative rates could do unintended damage to the financial system. In lieu of that, and as the pandemic experience shows, fiscal expansion can be effective in such circumstances.
Substituting for stablecoins. Should stablecoins pose an increased threat to financial stability or expand links with the real economy, uncapped, interest-bearing CBDC could substitute for them. However, a far better approach for preventing stablecoins from becoming a source of systemic risk is to implement the proposals of the President’s Working Group on Financial Markets. That is, the U.S. government should require stablecoin issuers to be insured depositories. Moreover, the advent of efficient retail payments frameworks (such as FedNow) that rely on commercial banks should help limit future links between the broader economy and stablecoins.
3. Could a CBDC affect financial inclusion? Would the net effect be positive or negative for inclusion?
In our view, on its own, a CBDC would not have a material impact on financial inclusion. As noted in our response to question #2, the most effective means for improving access for the unbanked and underbanked is to provide subsidized no-frills bank accounts combined with zero-cost publicly provided identity verification mechanisms.
4. How might a U.S. CBDC affect the Federal Reserve’s ability to effectively implement monetary policy in the pursuit of its maximum-employment and price-stability goals?
In normal times, the presence of CBDC should have little impact on the effectiveness of monetary policy. The reason is that the Fed can operate a policy interest rate channel system for the repo rate, establishing a standing repo facility (that sets the interest rate ceiling) and a standing reverse repo facility (that sets the interest rate floor) while implementing a daily, unlimited auction of reserve liabilities to banks at the market rate. The result would be a bank-demand determined policy rate within the channel set by the administered repo and reverse-repo rates. Since the channel need not be above zero, in the absence of paper currency, it also could be used to set a negative policy rate.
However, unless the Fed effectively recycles funds as they shift from commercial bank deposits into CBDC, disintermediation could undermine the bank lending channel of monetary policy transmission. This problem could be especially severe in a period of financial stress that triggers disorderly disintermediation. To the extent that the Federal Reserve relies on haircuts and a collateral framework for recycling CBDC to commercial banks, the framework will need to be consistent with the central bank’s monetary policy goals.
5. How could a CBDC affect financial stability? Would the net effect be positive or negative for stability?
An uncapped, interest-bearing CBDC would increase run risk in the private financial system. Put differently, the odds of disorderly disintermediation will rise in periods of stress, diminishing the supply of credit to healthy borrowers, and amplifying procyclicality. In addition, CBDC issued by a credible central bank like the Federal Reserve is likely to attract large capital inflows, further exacerbating the impact of global financial strains.
6. Could a CBDC adversely affect the financial sector? How might a CBDC affect the financial sector differently from stablecoins or other nonbank money?
A CBDC has no liquidity or credit risk. Consequently, it has a stable value in more states of the world than any other medium of exchange. In contrast, trust in nonbank money can plunge in periods of strain. As a result, uncapped, interest-bearing CBDC would be more attractive than stablecoins or other nonbank money. Putting all this together, uncapped, interest-bearing CBDC would increase run risk in the private financial system (as in the response to question #5).
7. What tools could be considered to mitigate any adverse impact of CBDC on the financial sector? Would some of these tools diminish the potential benefits of a CBDC?
We know of no means to limit the adverse impact on financial stability of uncapped, interest-bearing CBDC.
Capping holdings of CBDC would limit its use as a means of payment that substitutes for cash. More important, capping the amount of CBDC in periods of strain could limit runs into CBDC, but would not halt runs. Any scarcity of CBDC would result in a premium for CBDC relative to other central bank liabilities (such as currency in circulation and bank reserves) and to insured deposit balances. That premium would encourage runs into other safe, liquid instruments that are close substitutes for CBDC, such as Treasury bills and paper currency (if the Fed continues to supply it).
Not paying interest on CBDC would diminish its attractiveness relative to safe, liquid instruments such as Treasury bills. However, a key element of the Fed’s monetary policy framework is the payment of interest on banks’ reserve balances (including those of large foreign banks). So long as the Federal Reserve pays interest on reserves to banks, a policy of not paying interest on CBDC to retail holders is unlikely to remain politically viable.
8. If cash usage declines, is it important to preserve the general public’s access to a form of central bank money that can be used for payments?
No. Today, nearly all retail payments in the United States settle using commercial bank money. People trust commercial bank money—that is, it is accepted at par. The mix of credible prudential regulation, deposit insurance and access to the lender of last resort renders these private liabilities information-insensitive. So long as such information-insensitive commercial bank money is widely available, there is no need for direct access to central bank money to settle retail payments.
9. How might domestic and cross-border digital payments evolve in the absence of a U.S. CBDC?
Payments systems display strong network externalities that typically favor supplier concentration, and can lead to natural monopolies. Efficient, reliable payments mechanisms also constitute critical public goods, consistent with government provision of the payments infrastructure.
However, these two features of payments systems do not imply a need for CBDC. Indeed, all faster payments systems, whether domestic (such as the TIPS system operating in the euro area or the prospective FedNow mechanism in the United States) or cross-border (such as the P27 framework in the Nordic countries) involve network externalities, public infrastructure support, or both. None requires CBDC. Moreover, the technical challenges facing such payments—including interoperability or the use of multiple currencies—exist regardless of the presence of CBDC.
10. How should decisions by other large economy nations to issue CBDCs influence the decision whether the United States should do so?
We presume that the United States wishes to preserve the dollar’s reserve currency status, as well as its prominence in global trade and finance. Consequently, if highly trustworthy jurisdictions abroad (with unfettered cross-border capital flows and credible protection of property rights) offer CBDCs, increased competition for the dollar could shift the balance of benefits and risks in the direction of favoring a U.S. CBDC. For this reason, it is wise for U.S. officials to be prepared should the introduction of CBDC become desirable.
However, given the associated risks for the financial system, it is doubtful that somewhat greater competition for the dollar would be sufficient to warrant the introduction of a U.S. CBDC. Moreover, various uses of the dollar (as an invoicing currency, as the denomination of financial instruments, and as an official reserve asset) are highly persistent and may be insensitive to the presence of other CBDCs. Consequently, in our view, U.S. officials need not rush to introduce a CBDC regardless of the actions of others.
11. Are there additional ways to manage potential risks associated with CBDC that were not raised in this paper?
The most important risk associated with a CBDC is the potential for increased financial instability. Domestically the most likely source of such instability is disintermediation (the shift from commercial bank deposits to CBDC). Additionally, a U.S. CBDC is likely to attract inflows from abroad. So long as there are no caps and interest is paid—giving CBDC the features of central bank reserves—we are not aware of any means to manage these financial stability risks.
Moreover, to the extent that CBDC proves popular, the central bank will face additional risks arising from the need to recycle large inflows back to the commercial banking system (which will need the funding if it is to continue supplying private credit). Even if the Fed with a CBDC-expanded balance sheet is not tempted to become a state bank (directly allocating credit to the private sector), its collateral framework will be a strong influence on private lending behavior. This expanded role for the central bank will pose risks to its independence and, thereby, to the effectiveness of monetary policy.
12. How could a CBDC provide privacy to consumers without providing complete anonymity and facilitating illicit financial activity?
In line with the Federal Reserve’s report, we assume that any U.S. CBDC will be an intermediated instrument held in identity-verified accounts, rather than in an anonymous, tokenized form. Put differently, it will not be anonymous in the manner of currency or a bearer security.
This approach is likely to be the most efficient for balancing the goals of protecting privacy—which is necessary to make CBDC attractive—and of hindering illicit activity. Currently, accounts at U.S. banks provide extensive privacy, but are subject to legal reporting requirements (for example, in the event of suspicious activity). In effect, an intermediated CBDC will exploit banks’ comparative advantage in applying state-of-the-art know-your-customer (KYC) and anti-money laundering (AML) practices. Similarly, as is the case with commercial bank accounts today, officials also should be able to seek judicial approval for reviewing suspicious CBDC account activity.
13. How could a CBDC be designed to foster operational and cyber resiliency? What operational or cyber risks might be unavoidable?
We assume that CBDC will be provided in intermediated accounts that hold only CBDC (that is, they will not mingle CBDC with riskier assets). Nevertheless, the introduction of CBDC can increase the operational and cyber risks that central banks face because CBDC transactions would settle on the central bank’s balance sheet. Moreover, hostile agents who currently cannot directly hold bank reserves will be able to transact in CBDC through their intermediated accounts.
Fortunately, the Fed can anticipate these risks. Today, trillions of dollars of transactions among intermediaries settle each day on Fedwire, an infrastructure provided by the Federal Reserve that is robust to cyberthreats. Moreover, the Fed is preparing for the increased operational and cyberrisks that will be associated with the 24/7/365 FedNow faster payments framework that will have millions of potential entry points for hostile agents. From a domestic perspective, then, CBDC does not appear to pose novel operational challenges beyond those that the central bank already faces.
It remains to be seen whether new and different challenges will arise if the Federal Reserve seeks to promote interoperability with the CBDCs of other jurisdictions.
14. Should a CBDC be legal tender?
In our view, what matters for encouraging the widespread use of a financial instrument is public trust. If trust were to erode, the designation of U.S. currency in circulation as legal tender probably would do little to sustain its use. However, since CBDC aims to be a digital substitute for currency in circulation, and because U.S. cash today is designated as legal tender, it makes sense to give U.S. CBDC equivalent legal status to cash.
CBDC DESIGN
15. Should a CBDC pay interest? If so, why and how? If not, why not?
The Federal Reserve should expect to pay interest on CBDC. So long as the Federal Reserve pays interest on commercial bank reserve balances, a policy of not paying interest on CBDC to retail holders is unlikely to be politically viable. Moreover, not paying interest on CBDC would diminish its attractiveness relative to other safe, liquid instruments. Finally, since CBDC will be held in intermediated accounts, the interest that the central bank pays on CBDC will be in part a mechanism to compensate the private intermediaries for performing the necessary services of creating and maintaining accounts, verifying identities, tracking assets, implementing transactions, and monitoring for suspicious activity. Put differently, the private provider will require a net interest margin to cover these costs.
In practice, we expect that the interest rate paid on CBDC will be an administered rate that is linked to, but modestly lower than, the interest rate on reserves.
16. Should the amount of CBDC held by a single end user be subject to quantity limits?
CBDC held by a single end user should not be subject to quantity limits. In periods of financial strain, binding caps on individual holdings would lead to a premium on CBDC relative to other central bank liabilities. Rather than preventing runs, a premium on CBDC will simply re-direct runs into close substitutes such as Treasury bills.
17. What types of firms should serve as intermediaries for CBDC? What should be the role and regulatory structure for these intermediaries?
For CBDC transactions to settle on the balance sheet of the Federal Reserve, each CBDC intermediary will need an account at a Federal Reserve Bank. Commercial banks already hold their reserves in such accounts. For nonbanks that wish to intermediate CBDC, the Fed may need to create a different (but functionally equivalent) liability that can be held in a Fed account.
To obtain an account, banks today must be chartered and satisfy prudential standards. Going forward, the Fed should consider allowing nonbanks to serve as CBDC intermediaries provided they hold only CBDC (that is, they do not mingle other assets with CBDC) and that they qualify for a special CBDC license. While the license for such a “narrow” nonbank would not impose liquidity requirements, it should come with capital requirements sufficient to address operational risks.
18. Should a CBDC have “offline” capabilities? If so, how might that be achieved?
A robust, safe and efficient offline capability would add to the attractiveness of a U.S. CBDC. When the internet is not available, mobile telephony could provide an alternative communications tool. To facilitate everyday payments, small transactions (say, using Bluetooth links between counterparty electronic devices) also should be feasible offline.
19. Should a CBDC be designed to maximize ease of use and acceptance at the point of sale?
A key goal of CBDC should be to support fast and cheap payments. The FedNow payments framework that is scheduled to launch in 2023 does not require CBDC. However, if the Fed introduces a U.S. CBDC, paying with it could be made an option in FedNow.
20. How could a CBDC be designed to achieve transferability across multiple payment platforms? Would new technology or technical standards be needed?
Today, credit card networks already link multiple payment platforms. Domestically, these include point-of-sale devices that accept cards, smartphone wallets, and online payment systems. Links to comparable systems abroad extend these credit card networks vastly further.
With these examples in mind, we expect that domestic payment system providers will wish to allow the option of settling in CBDC, provided that the framework for doing so is efficient and resilient. Internationally, central banks will need to develop cooperative arrangements to achieve interoperability and implement safe, timely, and low-cost settlement systems.
21. How might future technological innovations affect design and policy choices related to CBDC?
For a CBDC to retain public trust, it is critical to maintain safety (including privacy) and resilience. In addition, transactions costs must remain low. Accordingly, we expect that CBDC issuers will be in a perpetual arms race with hostile actors to ensure cybersecurity and protect privacy.
22. Are there additional design principles that should be considered? Are there tradeoffs around any of the identified design principles, especially in trying to achieve the potential benefits of a CBDC?
We identified six key design features of a U.S. CBDC at the outset. In addition to the four features noted in the Federal Reserve report (privacy-protected, intermediated, transferable, and identity-verified), we added quantity-unlimited and interest-bearing.
These additional features involve clear tradeoffs. Avoiding a binding quantity cap is necessary to prevent a CBDC scarcity premium, while CBDC interest payments likely will be politically necessary so long as the Federal Reserve pays interest on bank reserves. However, an interest-bearing CBDC without a quantity cap risks greater disintermediation of domestic banks and currency substitution from less trustworthy foreign jurisdictions, as well as the expansion of the central bank balance sheet to the point where policymakers will directly or indirectly allocate the bulk of credit in the economy (eventually posing a risk to the independence of monetary policy).